From 44a8898e259afd43c7cb77bc026ef62ed5971c4f Mon Sep 17 00:00:00 2001 From: Yorke Rhodes III Date: Tue, 25 Aug 2026 16:27:57 -0400 Subject: [PATCH 1/2] Fall back to a monospace face, not Arial The body face is the monospace - that is the editorial choice the design system describes - but the fallback chain was Arial, Helvetica, sans-serif. Two consequences. If the webfont ever failed to load, the whole site rendered in a proportional face. And glyph by glyph: next/font subsets Space Mono to latin, latin-ext and vietnamese, so any character outside those ranges was drawn by the first fallback that had it. Chrome reports the arrow in a back link as Arial (1 glyph) beside Space Mono (31) - a proportional arrow inside monospace text, in the header's Contact button, every back link, the play triangles and the close crosses. That is the same root cause as the nav caret, which was fixed one glyph at a time. Naming a monospace fallback fixes all of them at once: the same probe now reports Cascadia Mono for the arrow. Body copy and headings are untouched. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> --- src/app/globals.css | 12 +++++++++++- src/app/layout.tsx | 13 +++++++++++++ 2 files changed, 24 insertions(+), 1 deletion(-) diff --git a/src/app/globals.css b/src/app/globals.css index aafa37ece..da1e6c7f6 100644 --- a/src/app/globals.css +++ b/src/app/globals.css @@ -91,7 +91,17 @@ body { background: var(--background); color: var(--foreground); - font-family: var(--font-sans), Arial, Helvetica, sans-serif; + /* The fallback chain is monospace on purpose. The body face *is* the + monospace (see docs/DESIGN-SYSTEM.md); falling back to Arial would have + rendered the whole site in a proportional face if the webfont ever failed. + It also mattered glyph by glyph: next/font subsets Space Mono to latin, + latin-ext and vietnamese, so a character outside those ranges — the arrows + in "Contact →" and "← Publications", the play triangle, the close cross — + was drawn from the first fallback that had it, which was Arial. A + proportional arrow sitting inside monospace text is the mismatch that made + the header look inconsistent. */ + font-family: var(--font-sans), ui-monospace, "Cascadia Mono", "Segoe UI Mono", + "DejaVu Sans Mono", Menlo, Consolas, monospace; -webkit-font-smoothing: antialiased; transition: background-color 0.2s ease, color 0.2s ease; } diff --git a/src/app/layout.tsx b/src/app/layout.tsx index a26e1bb5d..27bc7c145 100644 --- a/src/app/layout.tsx +++ b/src/app/layout.tsx @@ -20,6 +20,19 @@ const spaceMono = Space_Mono({ subsets: ["latin"], weight: ["400", "700"], style: ["normal", "italic"], + // next/font otherwise builds its metric-adjusted "Space Mono Fallback" face + // from Arial, which then supplies any glyph the latin subset leaves out — + // arrows, the play triangle, the close cross — in a proportional face inside + // monospace text. Naming a monospace fallback keeps those glyphs in kind. + fallback: [ + "ui-monospace", + "Cascadia Mono", + "Segoe UI Mono", + "DejaVu Sans Mono", + "Menlo", + "Consolas", + "monospace", + ], }); export const metadata: Metadata = { From 00e34d94698868f250023d630acb884b8646a18c Mon Sep 17 00:00:00 2001 From: Yorke Rhodes III Date: Tue, 25 Aug 2026 16:28:00 -0400 Subject: [PATCH 2/2] chore: regenerate the static snapshot Picks up the monospace fallback. next/font renames its generated font assets when the fallback metrics change, so 12 woff2 files are replaced; all 833 font references resolve. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> --- static-site/404.html | 2 +- static-site/404/index.html | 2 +- static-site/__next.__PAGE__.txt | 24 +-- static-site/__next._full.txt | 58 +++--- static-site/__next._head.txt | 6 +- static-site/__next._index.txt | 18 +- static-site/__next._tree.txt | 10 +- .../_next/static/chunks/0-5whcwv1z55i.js | 1 - .../_next/static/chunks/0jub-lza-_qsd.js | 1 + .../_next/static/chunks/1-kh6uz_49jxr.js | 1 - .../_next/static/chunks/1l1xxdmwck4_f.js | 1 + .../{1w0nyqw5bqniv.css => 1sag07fioh6gb.css} | 4 +- ...=> 17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2} | Bin ...=> 1ae2575eb5be4118.p.29un6u1l3mwyp.woff2} | Bin ...2 => 28e60ca39c9ae554.44izq0ia6orkf.woff2} | Bin ...2 => 2c4e8b1b42841af7.2p_e9n24plgne.woff2} | Bin ...2 => 417912ede2e82152.1umk1654cmfxm.woff2} | Bin ...2 => 4ba802ed8e67eac5.0fuxtgzazslqx.woff2} | Bin ...2 => 51f2519120a6b711.038kgsu-hd-vh.woff2} | Bin ...2 => 7663dec790de7af8.2b96g58xb1jk9.woff2} | Bin ...=> 8d05cfa5faa8406c.p.1s123piy9_v1m.woff2} | Bin ...2 => b8f2b92a9960dd69.0we3rzhp_h228.woff2} | Bin ...2 => d7a0600e467cf0bd.451yigaaz_4mg.woff2} | Bin ...=> dc0d9adbac686440.p.3cik_s2si-ft-.woff2} | Bin static-site/_not-found/__next._full.txt | 30 +-- static-site/_not-found/__next._head.txt | 6 +- static-site/_not-found/__next._index.txt | 18 +- static-site/_not-found/__next._not-found.txt | 4 +- static-site/_not-found/__next._tree.txt | 2 +- static-site/_not-found/index.html | 2 +- static-site/_not-found/index.txt | 30 +-- static-site/about/__next._full.txt | 38 ++-- static-site/about/__next._head.txt | 6 +- static-site/about/__next._index.txt | 18 +- static-site/about/__next._tree.txt | 10 +- static-site/about/__next.about.txt | 4 +- static-site/about/index.html | 2 +- static-site/about/index.txt | 38 ++-- static-site/contact/__next._full.txt | 38 ++-- static-site/contact/__next._head.txt | 6 +- static-site/contact/__next._index.txt | 18 +- static-site/contact/__next._tree.txt | 10 +- static-site/contact/__next.contact.txt | 4 +- static-site/contact/index.html | 2 +- static-site/contact/index.txt | 38 ++-- static-site/demos/__next._full.txt | 44 ++--- static-site/demos/__next._head.txt | 6 +- static-site/demos/__next._index.txt | 18 +- static-site/demos/__next._tree.txt | 10 +- static-site/demos/__next.demos.txt | 4 +- static-site/demos/index.html | 2 +- static-site/demos/index.txt | 44 ++--- static-site/font-lab/__next._full.txt | 38 ++-- static-site/font-lab/__next._head.txt | 6 +- static-site/font-lab/__next._index.txt | 18 +- static-site/font-lab/__next._tree.txt | 10 +- static-site/font-lab/__next.font-lab.txt | 4 +- static-site/font-lab/index.html | 2 +- static-site/font-lab/index.txt | 38 ++-- static-site/index.html | 2 +- static-site/index.txt | 58 +++--- static-site/media/__next._full.txt | 42 ++-- static-site/media/__next._head.txt | 6 +- static-site/media/__next._index.txt | 18 +- static-site/media/__next._tree.txt | 10 +- static-site/media/__next.media.txt | 4 +- static-site/media/index.html | 2 +- static-site/media/index.txt | 42 ++-- .../newsletter/2026-07/__next._full.txt | 44 ++--- .../newsletter/2026-07/__next._head.txt | 6 +- .../newsletter/2026-07/__next._index.txt | 18 +- .../newsletter/2026-07/__next._tree.txt | 10 +- .../newsletter/2026-07/__next.newsletter.txt | 4 +- static-site/newsletter/2026-07/index.html | 2 +- static-site/newsletter/2026-07/index.txt | 44 ++--- static-site/newsletter/__next._full.txt | 42 ++-- static-site/newsletter/__next._head.txt | 6 +- static-site/newsletter/__next._index.txt | 18 +- static-site/newsletter/__next._tree.txt | 10 +- static-site/newsletter/__next.newsletter.txt | 4 +- static-site/newsletter/index.html | 2 +- static-site/newsletter/index.txt | 42 ++-- static-site/portfolio/__next._full.txt | 46 ++--- static-site/portfolio/__next._head.txt | 6 +- static-site/portfolio/__next._index.txt | 18 +- static-site/portfolio/__next._tree.txt | 10 +- static-site/portfolio/__next.portfolio.txt | 4 +- static-site/portfolio/index.html | 2 +- static-site/portfolio/index.txt | 46 ++--- .../print/after-the-corridor/__next._full.txt | 38 ++-- .../print/after-the-corridor/__next._head.txt | 6 +- .../after-the-corridor/__next._index.txt | 18 +- .../print/after-the-corridor/__next._tree.txt | 10 +- .../print/after-the-corridor/__next.print.txt | 4 +- .../print/after-the-corridor/index.html | 4 +- .../print/after-the-corridor/index.txt | 38 ++-- .../__next._full.txt | 186 ++++++++++++++++++ .../__next._head.txt | 6 + .../__next._index.txt | 15 ++ .../__next._tree.txt | 7 + .../__next.print.txt | 4 + .../__next.print/$d$slug.txt | 5 + .../__next.print/$d$slug/__PAGE__.txt | 155 +++++++++++++++ .../agentic-language-development/index.html | 89 +++++++++ .../agentic-language-development/index.txt | 186 ++++++++++++++++++ .../ai-carbon-footprint/__next._full.txt | 38 ++-- .../ai-carbon-footprint/__next._head.txt | 6 +- .../ai-carbon-footprint/__next._index.txt | 18 +- .../ai-carbon-footprint/__next._tree.txt | 10 +- .../ai-carbon-footprint/__next.print.txt | 4 +- .../print/ai-carbon-footprint/index.html | 4 +- .../print/ai-carbon-footprint/index.txt | 38 ++-- .../print/ai-models-research/__next._full.txt | 38 ++-- .../print/ai-models-research/__next._head.txt | 6 +- .../ai-models-research/__next._index.txt | 18 +- .../print/ai-models-research/__next._tree.txt | 10 +- .../print/ai-models-research/__next.print.txt | 4 +- .../print/ai-models-research/index.html | 4 +- .../print/ai-models-research/index.txt | 38 ++-- .../ai-research-assistant/__next._full.txt | 38 ++-- .../ai-research-assistant/__next._head.txt | 6 +- .../ai-research-assistant/__next._index.txt | 18 +- .../ai-research-assistant/__next._tree.txt | 10 +- .../ai-research-assistant/__next.print.txt | 4 +- .../print/ai-research-assistant/index.html | 4 +- .../print/ai-research-assistant/index.txt | 38 ++-- static-site/print/cerai/__next._full.txt | 38 ++-- static-site/print/cerai/__next._head.txt | 6 +- static-site/print/cerai/__next._index.txt | 18 +- static-site/print/cerai/__next._tree.txt | 10 +- static-site/print/cerai/__next.print.txt | 4 +- static-site/print/cerai/index.html | 4 +- static-site/print/cerai/index.txt | 38 ++-- .../__next._full.txt | 38 ++-- .../__next._head.txt | 6 +- .../__next._index.txt | 18 +- .../__next._tree.txt | 10 +- .../__next.print.txt | 4 +- .../digital-provenance-passport/index.html | 4 +- .../digital-provenance-passport/index.txt | 38 ++-- .../diplomatic-simulator/__next._full.txt | 38 ++-- .../diplomatic-simulator/__next._head.txt | 6 +- .../diplomatic-simulator/__next._index.txt | 18 +- .../diplomatic-simulator/__next._tree.txt | 10 +- .../diplomatic-simulator/__next.print.txt | 4 +- .../print/diplomatic-simulator/index.html | 4 +- .../print/diplomatic-simulator/index.txt | 38 ++-- static-site/print/ercf/__next._full.txt | 38 ++-- static-site/print/ercf/__next._head.txt | 6 +- static-site/print/ercf/__next._index.txt | 18 +- static-site/print/ercf/__next._tree.txt | 10 +- static-site/print/ercf/__next.print.txt | 4 +- static-site/print/ercf/index.html | 4 +- static-site/print/ercf/index.txt | 38 ++-- static-site/print/erus/__next._full.txt | 38 ++-- static-site/print/erus/__next._head.txt | 6 +- static-site/print/erus/__next._index.txt | 18 +- static-site/print/erus/__next._tree.txt | 10 +- static-site/print/erus/__next.print.txt | 4 +- static-site/print/erus/index.html | 4 +- static-site/print/erus/index.txt | 38 ++-- .../evacuation-inform-index/__next._full.txt | 38 ++-- .../evacuation-inform-index/__next._head.txt | 6 +- .../evacuation-inform-index/__next._index.txt | 18 +- .../evacuation-inform-index/__next._tree.txt | 10 +- .../evacuation-inform-index/__next.print.txt | 4 +- .../print/evacuation-inform-index/index.html | 4 +- .../print/evacuation-inform-index/index.txt | 38 ++-- .../evacuation-simulation/__next._full.txt | 38 ++-- .../evacuation-simulation/__next._head.txt | 6 +- .../evacuation-simulation/__next._index.txt | 18 +- .../evacuation-simulation/__next._tree.txt | 10 +- .../evacuation-simulation/__next.print.txt | 4 +- .../print/evacuation-simulation/index.html | 4 +- .../print/evacuation-simulation/index.txt | 38 ++-- .../__next._full.txt | 38 ++-- .../__next._head.txt | 6 +- .../__next._index.txt | 18 +- .../__next._tree.txt | 10 +- .../__next.print.txt | 4 +- .../index.html | 4 +- .../index.txt | 38 ++-- static-site/print/haste/__next._full.txt | 38 ++-- static-site/print/haste/__next._head.txt | 6 +- static-site/print/haste/__next._index.txt | 18 +- static-site/print/haste/__next._tree.txt | 10 +- static-site/print/haste/__next.print.txt | 4 +- static-site/print/haste/index.html | 4 +- static-site/print/haste/index.txt | 38 ++-- .../mariupol-severity-model/__next._full.txt | 38 ++-- .../mariupol-severity-model/__next._head.txt | 6 +- .../mariupol-severity-model/__next._index.txt | 18 +- .../mariupol-severity-model/__next._tree.txt | 10 +- .../mariupol-severity-model/__next.print.txt | 4 +- .../print/mariupol-severity-model/index.html | 4 +- .../print/mariupol-severity-model/index.txt | 38 ++-- .../print/provenance-search/__next._full.txt | 38 ++-- .../print/provenance-search/__next._head.txt | 6 +- .../print/provenance-search/__next._index.txt | 18 +- .../print/provenance-search/__next._tree.txt | 10 +- .../print/provenance-search/__next.print.txt | 4 +- .../print/provenance-search/index.html | 4 +- static-site/print/provenance-search/index.txt | 38 ++-- static-site/print/vango/__next._full.txt | 38 ++-- static-site/print/vango/__next._head.txt | 6 +- static-site/print/vango/__next._index.txt | 18 +- static-site/print/vango/__next._tree.txt | 10 +- static-site/print/vango/__next.print.txt | 4 +- static-site/print/vango/index.html | 4 +- static-site/print/vango/index.txt | 38 ++-- static-site/print/war-games/__next._full.txt | 38 ++-- static-site/print/war-games/__next._head.txt | 6 +- static-site/print/war-games/__next._index.txt | 18 +- static-site/print/war-games/__next._tree.txt | 10 +- static-site/print/war-games/__next.print.txt | 4 +- static-site/print/war-games/index.html | 4 +- static-site/print/war-games/index.txt | 38 ++-- .../print/what-is-ethical-ai/__next._full.txt | 38 ++-- .../print/what-is-ethical-ai/__next._head.txt | 6 +- .../what-is-ethical-ai/__next._index.txt | 18 +- .../print/what-is-ethical-ai/__next._tree.txt | 10 +- .../print/what-is-ethical-ai/__next.print.txt | 4 +- .../print/what-is-ethical-ai/index.html | 4 +- .../print/what-is-ethical-ai/index.txt | 38 ++-- static-site/publications/__next._full.txt | 46 ++--- static-site/publications/__next._head.txt | 6 +- static-site/publications/__next._index.txt | 18 +- static-site/publications/__next._tree.txt | 10 +- .../publications/__next.publications.txt | 4 +- .../after-the-corridor/__next._full.txt | 44 ++--- .../after-the-corridor/__next._head.txt | 6 +- .../after-the-corridor/__next._index.txt | 18 +- .../after-the-corridor/__next._tree.txt | 10 +- .../__next.publications.txt | 4 +- .../after-the-corridor/index.html | 2 +- .../publications/after-the-corridor/index.txt | 44 ++--- .../__next._full.txt | 127 ++++++++++++ .../__next._head.txt | 6 + .../__next._index.txt | 15 ++ .../__next._tree.txt | 8 + .../__next.publications.txt | 5 + .../agentic-language-development.txt | 5 + .../agentic-language-development/__PAGE__.txt | 97 +++++++++ .../agentic-language-development/index.html | 1 + .../agentic-language-development/index.txt | 127 ++++++++++++ .../pages/manifest.json | 32 +++ .../pages/p01.webp | Bin 0 -> 81840 bytes .../pages/p02.webp | Bin 0 -> 103462 bytes .../pages/p03.webp | Bin 0 -> 123006 bytes .../pages/p04.webp | Bin 0 -> 116568 bytes .../pages/p05.webp | Bin 0 -> 157916 bytes .../pages/p06.webp | Bin 0 -> 134240 bytes .../pages/p07.webp | Bin 0 -> 130476 bytes .../pages/p08.webp | Bin 0 -> 109850 bytes .../pages/p09.webp | Bin 0 -> 127774 bytes .../pages/p10.webp | Bin 0 -> 140338 bytes .../pages/p11.webp | Bin 0 -> 109448 bytes .../pages/p12.webp | Bin 0 -> 122006 bytes .../pages/p13.webp | Bin 0 -> 156226 bytes .../pages/p14.webp | Bin 0 -> 151396 bytes .../pages/p15.webp | Bin 0 -> 152510 bytes .../pages/p16.webp | Bin 0 -> 125442 bytes .../pages/p17.webp | Bin 0 -> 139692 bytes .../pages/p18.webp | Bin 0 -> 115288 bytes .../pages/p19.webp | Bin 0 -> 131088 bytes .../pages/p20.webp | Bin 0 -> 76658 bytes .../pages/p21.webp | Bin 0 -> 74032 bytes .../pages/p22.webp | Bin 0 -> 151734 bytes .../pages/p23.webp | Bin 0 -> 165624 bytes .../pages/p24.webp | Bin 0 -> 161790 bytes .../pages/p25.webp | Bin 0 -> 30466 bytes .../agentic-language-development/report.pdf | Bin 0 -> 614046 bytes .../ai-carbon-footprint/__next._full.txt | 44 ++--- .../ai-carbon-footprint/__next._head.txt | 6 +- .../ai-carbon-footprint/__next._index.txt | 18 +- .../ai-carbon-footprint/__next._tree.txt | 10 +- .../__next.publications.txt | 4 +- .../ai-carbon-footprint/index.html | 2 +- .../ai-carbon-footprint/index.txt | 44 ++--- .../ai-models-research/__next._full.txt | 44 ++--- .../ai-models-research/__next._head.txt | 6 +- .../ai-models-research/__next._index.txt | 18 +- .../ai-models-research/__next._tree.txt | 10 +- .../__next.publications.txt | 4 +- .../ai-models-research/index.html | 2 +- .../publications/ai-models-research/index.txt | 44 ++--- .../ai-research-assistant/__next._full.txt | 44 ++--- .../ai-research-assistant/__next._head.txt | 6 +- .../ai-research-assistant/__next._index.txt | 18 +- .../ai-research-assistant/__next._tree.txt | 10 +- .../__next.publications.txt | 4 +- .../ai-research-assistant/index.html | 4 +- .../ai-research-assistant/index.txt | 44 ++--- .../publications/cerai/__next._full.txt | 44 ++--- .../publications/cerai/__next._head.txt | 6 +- .../publications/cerai/__next._index.txt | 18 +- .../publications/cerai/__next._tree.txt | 10 +- .../cerai/__next.publications.txt | 4 +- static-site/publications/cerai/index.html | 2 +- static-site/publications/cerai/index.txt | 44 ++--- .../__next._full.txt | 44 ++--- .../__next._head.txt | 6 +- .../__next._index.txt | 18 +- .../__next._tree.txt | 10 +- .../__next.publications.txt | 4 +- .../digital-provenance-passport/index.html | 2 +- .../digital-provenance-passport/index.txt | 44 ++--- .../diplomatic-simulator/__next._full.txt | 44 ++--- .../diplomatic-simulator/__next._head.txt | 6 +- .../diplomatic-simulator/__next._index.txt | 18 +- .../diplomatic-simulator/__next._tree.txt | 10 +- .../__next.publications.txt | 4 +- .../diplomatic-simulator/index.html | 2 +- .../diplomatic-simulator/index.txt | 44 ++--- .../publications/ercf/__next._full.txt | 44 ++--- .../publications/ercf/__next._head.txt | 6 +- .../publications/ercf/__next._index.txt | 18 +- .../publications/ercf/__next._tree.txt | 10 +- .../publications/ercf/__next.publications.txt | 4 +- static-site/publications/ercf/index.html | 2 +- static-site/publications/ercf/index.txt | 44 ++--- .../publications/erus/__next._full.txt | 44 ++--- .../publications/erus/__next._head.txt | 6 +- .../publications/erus/__next._index.txt | 18 +- .../publications/erus/__next._tree.txt | 10 +- .../publications/erus/__next.publications.txt | 4 +- static-site/publications/erus/index.html | 2 +- static-site/publications/erus/index.txt | 44 ++--- .../evacuation-inform-index/__next._full.txt | 44 ++--- .../evacuation-inform-index/__next._head.txt | 6 +- .../evacuation-inform-index/__next._index.txt | 18 +- .../evacuation-inform-index/__next._tree.txt | 10 +- .../__next.publications.txt | 4 +- .../evacuation-inform-index/index.html | 2 +- .../evacuation-inform-index/index.txt | 44 ++--- .../evacuation-simulation/__next._full.txt | 44 ++--- .../evacuation-simulation/__next._head.txt | 6 +- .../evacuation-simulation/__next._index.txt | 18 +- .../evacuation-simulation/__next._tree.txt | 10 +- .../__next.publications.txt | 4 +- .../evacuation-simulation/index.html | 2 +- .../evacuation-simulation/index.txt | 44 ++--- .../__next._full.txt | 44 ++--- .../__next._head.txt | 6 +- .../__next._index.txt | 18 +- .../__next._tree.txt | 10 +- .../__next.publications.txt | 4 +- .../index.html | 2 +- .../index.txt | 44 ++--- .../publications/haste/__next._full.txt | 44 ++--- .../publications/haste/__next._head.txt | 6 +- .../publications/haste/__next._index.txt | 18 +- .../publications/haste/__next._tree.txt | 10 +- .../haste/__next.publications.txt | 4 +- static-site/publications/haste/index.html | 2 +- static-site/publications/haste/index.txt | 44 ++--- static-site/publications/index.html | 2 +- static-site/publications/index.txt | 46 ++--- .../mariupol-severity-model/__next._full.txt | 44 ++--- .../mariupol-severity-model/__next._head.txt | 6 +- .../mariupol-severity-model/__next._index.txt | 18 +- .../mariupol-severity-model/__next._tree.txt | 10 +- .../__next.publications.txt | 4 +- .../mariupol-severity-model/index.html | 2 +- .../mariupol-severity-model/index.txt | 44 ++--- .../provenance-search/__next._full.txt | 44 ++--- .../provenance-search/__next._head.txt | 6 +- .../provenance-search/__next._index.txt | 18 +- .../provenance-search/__next._tree.txt | 10 +- .../provenance-search/__next.publications.txt | 4 +- .../publications/provenance-search/index.html | 2 +- .../publications/provenance-search/index.txt | 44 ++--- .../publications/vango/__next._full.txt | 44 ++--- .../publications/vango/__next._head.txt | 6 +- .../publications/vango/__next._index.txt | 18 +- .../publications/vango/__next._tree.txt | 10 +- .../vango/__next.publications.txt | 4 +- static-site/publications/vango/index.html | 2 +- static-site/publications/vango/index.txt | 44 ++--- .../publications/war-games/__next._full.txt | 44 ++--- .../publications/war-games/__next._head.txt | 6 +- .../publications/war-games/__next._index.txt | 18 +- .../publications/war-games/__next._tree.txt | 10 +- .../war-games/__next.publications.txt | 4 +- static-site/publications/war-games/index.html | 2 +- static-site/publications/war-games/index.txt | 44 ++--- .../what-is-ethical-ai/__next._full.txt | 44 ++--- .../what-is-ethical-ai/__next._head.txt | 6 +- .../what-is-ethical-ai/__next._index.txt | 18 +- .../what-is-ethical-ai/__next._tree.txt | 10 +- .../__next.publications.txt | 4 +- .../what-is-ethical-ai/index.html | 2 +- .../publications/what-is-ethical-ai/index.txt | 44 ++--- static-site/team/__next._full.txt | 48 ++--- static-site/team/__next._head.txt | 6 +- static-site/team/__next._index.txt | 18 +- static-site/team/__next._tree.txt | 10 +- static-site/team/__next.team.txt | 4 +- .../team/adeline-daab/__next._full.txt | 44 ++--- .../team/adeline-daab/__next._head.txt | 6 +- .../team/adeline-daab/__next._index.txt | 18 +- .../team/adeline-daab/__next._tree.txt | 10 +- static-site/team/adeline-daab/__next.team.txt | 4 +- static-site/team/adeline-daab/index.html | 4 +- static-site/team/adeline-daab/index.txt | 44 ++--- .../team/alana-robertson/__next._full.txt | 44 ++--- .../team/alana-robertson/__next._head.txt | 6 +- .../team/alana-robertson/__next._index.txt | 18 +- .../team/alana-robertson/__next._tree.txt | 10 +- .../team/alana-robertson/__next.team.txt | 4 +- static-site/team/alana-robertson/index.html | 4 +- static-site/team/alana-robertson/index.txt | 44 ++--- static-site/team/alex-du/__next._full.txt | 44 ++--- static-site/team/alex-du/__next._head.txt | 6 +- static-site/team/alex-du/__next._index.txt | 18 +- static-site/team/alex-du/__next._tree.txt | 10 +- static-site/team/alex-du/__next.team.txt | 4 +- static-site/team/alex-du/index.html | 2 +- static-site/team/alex-du/index.txt | 44 ++--- .../team/alexa-shamie/__next._full.txt | 44 ++--- .../team/alexa-shamie/__next._head.txt | 6 +- .../team/alexa-shamie/__next._index.txt | 18 +- .../team/alexa-shamie/__next._tree.txt | 10 +- static-site/team/alexa-shamie/__next.team.txt | 4 +- static-site/team/alexa-shamie/index.html | 2 +- static-site/team/alexa-shamie/index.txt | 44 ++--- .../team/amanda-lindsey/__next._full.txt | 44 ++--- .../team/amanda-lindsey/__next._head.txt | 6 +- .../team/amanda-lindsey/__next._index.txt | 18 +- .../team/amanda-lindsey/__next._tree.txt | 10 +- .../team/amanda-lindsey/__next.team.txt | 4 +- static-site/team/amanda-lindsey/index.html | 2 +- static-site/team/amanda-lindsey/index.txt | 44 ++--- static-site/team/carlos-ruiz/__next._full.txt | 44 ++--- static-site/team/carlos-ruiz/__next._head.txt | 6 +- .../team/carlos-ruiz/__next._index.txt | 18 +- static-site/team/carlos-ruiz/__next._tree.txt | 10 +- static-site/team/carlos-ruiz/__next.team.txt | 4 +- static-site/team/carlos-ruiz/index.html | 2 +- static-site/team/carlos-ruiz/index.txt | 44 ++--- .../team/carolina-moron/__next._full.txt | 44 ++--- .../team/carolina-moron/__next._head.txt | 6 +- .../team/carolina-moron/__next._index.txt | 18 +- .../team/carolina-moron/__next._tree.txt | 10 +- .../team/carolina-moron/__next.team.txt | 4 +- static-site/team/carolina-moron/index.html | 4 +- static-site/team/carolina-moron/index.txt | 44 ++--- .../team/christine-lumen/__next._full.txt | 44 ++--- .../team/christine-lumen/__next._head.txt | 6 +- .../team/christine-lumen/__next._index.txt | 18 +- .../team/christine-lumen/__next._tree.txt | 10 +- .../team/christine-lumen/__next.team.txt | 4 +- static-site/team/christine-lumen/index.html | 4 +- static-site/team/christine-lumen/index.txt | 44 ++--- .../team/elizabeth-matthews/__next._full.txt | 44 ++--- .../team/elizabeth-matthews/__next._head.txt | 6 +- .../team/elizabeth-matthews/__next._index.txt | 18 +- .../team/elizabeth-matthews/__next._tree.txt | 10 +- .../team/elizabeth-matthews/__next.team.txt | 4 +- .../team/elizabeth-matthews/index.html | 2 +- static-site/team/elizabeth-matthews/index.txt | 44 ++--- .../team/emily-harrington/__next._full.txt | 44 ++--- .../team/emily-harrington/__next._head.txt | 6 +- .../team/emily-harrington/__next._index.txt | 18 +- .../team/emily-harrington/__next._tree.txt | 10 +- .../team/emily-harrington/__next.team.txt | 4 +- static-site/team/emily-harrington/index.html | 2 +- static-site/team/emily-harrington/index.txt | 44 ++--- .../team/grace-driscoll/__next._full.txt | 44 ++--- .../team/grace-driscoll/__next._head.txt | 6 +- .../team/grace-driscoll/__next._index.txt | 18 +- .../team/grace-driscoll/__next._tree.txt | 10 +- .../team/grace-driscoll/__next.team.txt | 4 +- static-site/team/grace-driscoll/index.html | 2 +- static-site/team/grace-driscoll/index.txt | 44 ++--- static-site/team/hannah-zhao/__next._full.txt | 44 ++--- static-site/team/hannah-zhao/__next._head.txt | 6 +- .../team/hannah-zhao/__next._index.txt | 18 +- static-site/team/hannah-zhao/__next._tree.txt | 10 +- static-site/team/hannah-zhao/__next.team.txt | 4 +- static-site/team/hannah-zhao/index.html | 2 +- static-site/team/hannah-zhao/index.txt | 44 ++--- static-site/team/index.html | 2 +- static-site/team/index.txt | 48 ++--- .../team/india-clarke/__next._full.txt | 44 ++--- .../team/india-clarke/__next._head.txt | 6 +- .../team/india-clarke/__next._index.txt | 18 +- .../team/india-clarke/__next._tree.txt | 10 +- static-site/team/india-clarke/__next.team.txt | 4 +- static-site/team/india-clarke/index.html | 4 +- static-site/team/india-clarke/index.txt | 44 ++--- .../team/jennifer-hofmann/__next._full.txt | 44 ++--- .../team/jennifer-hofmann/__next._head.txt | 6 +- .../team/jennifer-hofmann/__next._index.txt | 18 +- .../team/jennifer-hofmann/__next._tree.txt | 10 +- .../team/jennifer-hofmann/__next.team.txt | 4 +- static-site/team/jennifer-hofmann/index.html | 2 +- static-site/team/jennifer-hofmann/index.txt | 44 ++--- static-site/team/kirsten-co/__next._full.txt | 44 ++--- static-site/team/kirsten-co/__next._head.txt | 6 +- static-site/team/kirsten-co/__next._index.txt | 18 +- static-site/team/kirsten-co/__next._tree.txt | 10 +- static-site/team/kirsten-co/__next.team.txt | 4 +- static-site/team/kirsten-co/index.html | 4 +- static-site/team/kirsten-co/index.txt | 44 ++--- .../team/melanie-mackew/__next._full.txt | 44 ++--- .../team/melanie-mackew/__next._head.txt | 6 +- .../team/melanie-mackew/__next._index.txt | 18 +- .../team/melanie-mackew/__next._tree.txt | 10 +- .../team/melanie-mackew/__next.team.txt | 4 +- static-site/team/melanie-mackew/index.html | 2 +- static-site/team/melanie-mackew/index.txt | 44 ++--- .../team/mohagani-townsend/__next._full.txt | 44 ++--- .../team/mohagani-townsend/__next._head.txt | 6 +- .../team/mohagani-townsend/__next._index.txt | 18 +- .../team/mohagani-townsend/__next._tree.txt | 10 +- .../team/mohagani-townsend/__next.team.txt | 4 +- static-site/team/mohagani-townsend/index.html | 2 +- static-site/team/mohagani-townsend/index.txt | 44 ++--- .../team/natasha-nagarajan/__next._full.txt | 44 ++--- .../team/natasha-nagarajan/__next._head.txt | 6 +- .../team/natasha-nagarajan/__next._index.txt | 18 +- .../team/natasha-nagarajan/__next._tree.txt | 10 +- .../team/natasha-nagarajan/__next.team.txt | 4 +- static-site/team/natasha-nagarajan/index.html | 2 +- static-site/team/natasha-nagarajan/index.txt | 44 ++--- static-site/team/pegi-bracaj/__next._full.txt | 44 ++--- static-site/team/pegi-bracaj/__next._head.txt | 6 +- .../team/pegi-bracaj/__next._index.txt | 18 +- static-site/team/pegi-bracaj/__next._tree.txt | 10 +- static-site/team/pegi-bracaj/__next.team.txt | 4 +- static-site/team/pegi-bracaj/index.html | 2 +- static-site/team/pegi-bracaj/index.txt | 44 ++--- .../team/renata-gladkikh/__next._full.txt | 44 ++--- .../team/renata-gladkikh/__next._head.txt | 6 +- .../team/renata-gladkikh/__next._index.txt | 18 +- .../team/renata-gladkikh/__next._tree.txt | 10 +- .../team/renata-gladkikh/__next.team.txt | 4 +- static-site/team/renata-gladkikh/index.html | 2 +- static-site/team/renata-gladkikh/index.txt | 44 ++--- .../team/smita-samanta/__next._full.txt | 44 ++--- .../team/smita-samanta/__next._head.txt | 6 +- .../team/smita-samanta/__next._index.txt | 18 +- .../team/smita-samanta/__next._tree.txt | 10 +- .../team/smita-samanta/__next.team.txt | 4 +- static-site/team/smita-samanta/index.html | 2 +- static-site/team/smita-samanta/index.txt | 44 ++--- .../team/susan-deminil/__next._full.txt | 44 ++--- .../team/susan-deminil/__next._head.txt | 6 +- .../team/susan-deminil/__next._index.txt | 18 +- .../team/susan-deminil/__next._tree.txt | 10 +- .../team/susan-deminil/__next.team.txt | 4 +- static-site/team/susan-deminil/index.html | 4 +- static-site/team/susan-deminil/index.txt | 44 ++--- .../team/sylvia-maier/__next._full.txt | 44 ++--- .../team/sylvia-maier/__next._head.txt | 6 +- .../team/sylvia-maier/__next._index.txt | 18 +- .../team/sylvia-maier/__next._tree.txt | 10 +- static-site/team/sylvia-maier/__next.team.txt | 4 +- static-site/team/sylvia-maier/index.html | 4 +- static-site/team/sylvia-maier/index.txt | 44 ++--- static-site/team/taylor-badt/__next._full.txt | 44 ++--- static-site/team/taylor-badt/__next._head.txt | 6 +- .../team/taylor-badt/__next._index.txt | 18 +- static-site/team/taylor-badt/__next._tree.txt | 10 +- static-site/team/taylor-badt/__next.team.txt | 4 +- static-site/team/taylor-badt/index.html | 2 +- static-site/team/taylor-badt/index.txt | 44 ++--- .../team/teresa-cantero/__next._full.txt | 44 ++--- .../team/teresa-cantero/__next._head.txt | 6 +- .../team/teresa-cantero/__next._index.txt | 18 +- .../team/teresa-cantero/__next._tree.txt | 10 +- .../team/teresa-cantero/__next.team.txt | 4 +- static-site/team/teresa-cantero/index.html | 4 +- static-site/team/teresa-cantero/index.txt | 44 ++--- static-site/team/vedant-jain/__next._full.txt | 44 ++--- static-site/team/vedant-jain/__next._head.txt | 6 +- .../team/vedant-jain/__next._index.txt | 18 +- static-site/team/vedant-jain/__next._tree.txt | 10 +- static-site/team/vedant-jain/__next.team.txt | 4 +- static-site/team/vedant-jain/index.html | 2 +- static-site/team/vedant-jain/index.txt | 44 ++--- static-site/team/yago-rocha/__next._full.txt | 44 ++--- static-site/team/yago-rocha/__next._head.txt | 6 +- static-site/team/yago-rocha/__next._index.txt | 18 +- static-site/team/yago-rocha/__next._tree.txt | 10 +- static-site/team/yago-rocha/__next.team.txt | 4 +- static-site/team/yago-rocha/index.html | 4 +- static-site/team/yago-rocha/index.txt | 44 ++--- .../team/yorke-rhodes/__next._full.txt | 44 ++--- .../team/yorke-rhodes/__next._head.txt | 6 +- .../team/yorke-rhodes/__next._index.txt | 18 +- .../team/yorke-rhodes/__next._tree.txt | 10 +- static-site/team/yorke-rhodes/__next.team.txt | 4 +- static-site/team/yorke-rhodes/index.html | 2 +- static-site/team/yorke-rhodes/index.txt | 44 ++--- 596 files changed, 5859 insertions(+), 4783 deletions(-) delete mode 100644 static-site/_next/static/chunks/0-5whcwv1z55i.js create mode 100644 static-site/_next/static/chunks/0jub-lza-_qsd.js delete mode 100644 static-site/_next/static/chunks/1-kh6uz_49jxr.js create mode 100644 static-site/_next/static/chunks/1l1xxdmwck4_f.js rename static-site/_next/static/chunks/{1w0nyqw5bqniv.css => 1sag07fioh6gb.css} (75%) rename static-site/_next/static/media/{17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2 => 17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2} (100%) rename static-site/_next/static/media/{1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2 => 1ae2575eb5be4118.p.29un6u1l3mwyp.woff2} (100%) rename static-site/_next/static/media/{28e60ca39c9ae554-s.44izq0ia6orkf.woff2 => 28e60ca39c9ae554.44izq0ia6orkf.woff2} (100%) rename static-site/_next/static/media/{2c4e8b1b42841af7-s.2p_e9n24plgne.woff2 => 2c4e8b1b42841af7.2p_e9n24plgne.woff2} (100%) rename static-site/_next/static/media/{417912ede2e82152-s.1umk1654cmfxm.woff2 => 417912ede2e82152.1umk1654cmfxm.woff2} (100%) rename static-site/_next/static/media/{4ba802ed8e67eac5-s.0fuxtgzazslqx.woff2 => 4ba802ed8e67eac5.0fuxtgzazslqx.woff2} (100%) rename static-site/_next/static/media/{51f2519120a6b711-s.038kgsu-hd-vh.woff2 => 51f2519120a6b711.038kgsu-hd-vh.woff2} (100%) rename static-site/_next/static/media/{7663dec790de7af8-s.2b96g58xb1jk9.woff2 => 7663dec790de7af8.2b96g58xb1jk9.woff2} (100%) rename static-site/_next/static/media/{8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2 => 8d05cfa5faa8406c.p.1s123piy9_v1m.woff2} (100%) rename static-site/_next/static/media/{b8f2b92a9960dd69-s.0we3rzhp_h228.woff2 => b8f2b92a9960dd69.0we3rzhp_h228.woff2} (100%) rename static-site/_next/static/media/{d7a0600e467cf0bd-s.451yigaaz_4mg.woff2 => d7a0600e467cf0bd.451yigaaz_4mg.woff2} (100%) rename static-site/_next/static/media/{dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2 => dc0d9adbac686440.p.3cik_s2si-ft-.woff2} (100%) create mode 100644 static-site/print/agentic-language-development/__next._full.txt create mode 100644 static-site/print/agentic-language-development/__next._head.txt create mode 100644 static-site/print/agentic-language-development/__next._index.txt create mode 100644 static-site/print/agentic-language-development/__next._tree.txt create mode 100644 static-site/print/agentic-language-development/__next.print.txt create mode 100644 static-site/print/agentic-language-development/__next.print/$d$slug.txt create mode 100644 static-site/print/agentic-language-development/__next.print/$d$slug/__PAGE__.txt create mode 100644 static-site/print/agentic-language-development/index.html create mode 100644 static-site/print/agentic-language-development/index.txt create mode 100644 static-site/publications/agentic-language-development/__next._full.txt create mode 100644 static-site/publications/agentic-language-development/__next._head.txt create mode 100644 static-site/publications/agentic-language-development/__next._index.txt create mode 100644 static-site/publications/agentic-language-development/__next._tree.txt create mode 100644 static-site/publications/agentic-language-development/__next.publications.txt create mode 100644 static-site/publications/agentic-language-development/__next.publications/agentic-language-development.txt create mode 100644 static-site/publications/agentic-language-development/__next.publications/agentic-language-development/__PAGE__.txt create mode 100644 static-site/publications/agentic-language-development/index.html create mode 100644 static-site/publications/agentic-language-development/index.txt create mode 100644 static-site/publications/agentic-language-development/pages/manifest.json create mode 100644 static-site/publications/agentic-language-development/pages/p01.webp create mode 100644 static-site/publications/agentic-language-development/pages/p02.webp create mode 100644 static-site/publications/agentic-language-development/pages/p03.webp create mode 100644 static-site/publications/agentic-language-development/pages/p04.webp create mode 100644 static-site/publications/agentic-language-development/pages/p05.webp create mode 100644 static-site/publications/agentic-language-development/pages/p06.webp create mode 100644 static-site/publications/agentic-language-development/pages/p07.webp create mode 100644 static-site/publications/agentic-language-development/pages/p08.webp create mode 100644 static-site/publications/agentic-language-development/pages/p09.webp create mode 100644 static-site/publications/agentic-language-development/pages/p10.webp create mode 100644 static-site/publications/agentic-language-development/pages/p11.webp create mode 100644 static-site/publications/agentic-language-development/pages/p12.webp create mode 100644 static-site/publications/agentic-language-development/pages/p13.webp create mode 100644 static-site/publications/agentic-language-development/pages/p14.webp create mode 100644 static-site/publications/agentic-language-development/pages/p15.webp create mode 100644 static-site/publications/agentic-language-development/pages/p16.webp create mode 100644 static-site/publications/agentic-language-development/pages/p17.webp create mode 100644 static-site/publications/agentic-language-development/pages/p18.webp create mode 100644 static-site/publications/agentic-language-development/pages/p19.webp create mode 100644 static-site/publications/agentic-language-development/pages/p20.webp create mode 100644 static-site/publications/agentic-language-development/pages/p21.webp create mode 100644 static-site/publications/agentic-language-development/pages/p22.webp create mode 100644 static-site/publications/agentic-language-development/pages/p23.webp create mode 100644 static-site/publications/agentic-language-development/pages/p24.webp create mode 100644 static-site/publications/agentic-language-development/pages/p25.webp create mode 100644 static-site/publications/agentic-language-development/report.pdf diff --git a/static-site/404.html b/static-site/404.html index 324c33ebb..ffa713f4e 100644 --- a/static-site/404.html +++ b/static-site/404.html @@ -1 +1 @@ -404: This page could not be found.NYU Ethical Tech CoLab · Emerging Tech, Human Condition

404

This page could not be found.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file +404: This page could not be found.NYU Ethical Tech CoLab · Emerging Tech, Human Condition

404

This page could not be found.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file diff --git a/static-site/404/index.html b/static-site/404/index.html index 324c33ebb..ffa713f4e 100644 --- a/static-site/404/index.html +++ b/static-site/404/index.html @@ -1 +1 @@ -404: This page could not be found.NYU Ethical Tech CoLab · Emerging Tech, Human Condition

404

This page could not be found.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file +404: This page could not be found.NYU Ethical Tech CoLab · Emerging Tech, Human Condition

404

This page could not be found.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file diff --git a/static-site/__next.__PAGE__.txt b/static-site/__next.__PAGE__.txt index a89db6622..de8f1beba 100644 --- a/static-site/__next.__PAGE__.txt +++ b/static-site/__next.__PAGE__.txt @@ -1,17 +1,17 @@ 1:"$Sreact.fragment" -2:I[85437,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"Image"] -3:I[25150,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"HeroField"] -4:I[82987,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"Reveal"] -5:I[53675,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"StatementCarousel"] -e:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"Link"] -f:I[82987,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"Stagger"] -10:I[82987,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"StaggerItem"] -11:I[35852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"ProjectDiagram"] -12:I[12243,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"Tilt3D"] -16:I[94376,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"Magnetic"] -17:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +2:I[85437,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"Image"] +3:I[25150,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"HeroField"] +4:I[82987,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"Reveal"] +5:I[53675,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"StatementCarousel"] +e:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"Link"] +f:I[82987,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"Stagger"] +10:I[82987,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"StaggerItem"] +11:I[35852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"ProjectDiagram"] +12:I[12243,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"Tilt3D"] +16:I[94376,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"Magnetic"] +17:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 18:"$Sreact.suspense" -0:{"rsc":["$","$1","c",{"children":[[["$","section",null,{"className":"relative overflow-hidden border-b border-border","children":[["$","$L2",null,{"src":"/website/nyu-subway.jpg","alt":"","aria-hidden":true,"fill":true,"priority":true,"sizes":"100vw","className":"pointer-events-none object-cover object-center opacity-25"}],["$","div",null,{"aria-hidden":true,"className":"pointer-events-none absolute inset-0 bg-background/70"}],["$","span",null,{"className":"aura"}],["$","div",null,{"aria-hidden":true,"className":"pointer-events-none absolute inset-0","style":{"background":"radial-gradient(70% 60% at 20% 0%, color-mix(in oklab, var(--glow) 26%, transparent), transparent 65%)"}}],["$","$L3",null,{}],["$","div",null,{"className":"relative mx-auto max-w-6xl px-6 py-24 text-center sm:py-28","children":[["$","$L4",null,{"children":["$","p",null,{"className":"text-base uppercase tracking-[0.25em] text-accent sm:text-lg","children":"NYU CGA × Microsoft"}]}],["$","$L4",null,{"delay":0.15,"children":["$","$L5",null,{"statements":[{"lead":"Ethical Tech CoLab","headingClass":"text-[clamp(4.25rem,13vw,11rem)] leading-[0.88]","figure":["$","p",null,{"className":"font-serif uppercase leading-[0.95] tracking-tight text-foreground","style":{"fontSize":"clamp(1.9rem, 4vw, 3rem)"},"children":["Exploring technology to improve",["$","br",null,{"className":"hidden sm:block"}]," the"," ",["$","span",null,{"className":"display-em","children":"human condition"}],"."]}],"line":["$","p",null,{"className":"mx-auto mt-7 max-w-2xl leading-relaxed text-foreground/85","children":["A research collaboration between NYU's"," ",["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"link-underline text-accent hover:opacity-80","children":"Center for Global Affairs"}]," ","and Microsoft — changing the conversation on how people are informed, and how emerging technology can be used for good."]}]},{"lead":"Four questions. ","em":"One frontier.","figure":"14 projects in the portfolio","line":"Evacuation, cultural heritage, traceability, diplomacy — each question carried through to a fielded prototype, in the open.","cta":"Explore the portfolio","href":"/portfolio"},{"lead":"Run the ","em":"research","tail":".","figure":"18 demos you can open","line":"Not screenshots: the prototypes themselves, running in the browser with their source alongside.","cta":"Open the live demos","href":"/demos"},{"lead":"The research, ","em":"written up","tail":".","figure":"25 in the catalogue","line":"Every research question the CoLab takes on is written up academically, including what did not hold.","cta":"Read the publications","href":"/publications"},{"lead":"The people ","em":"building","tail":" this.","figure":"23 researchers across 3 cohorts","line":"Graduate researchers at NYU's Center for Global Affairs, with advisors and resident fellows alongside.","cta":"Meet the team","href":"/team"}],"label":"Statement","className":"mt-5"}]}]]}]]}],[["$","section",null,{"className":"border-b border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16 text-center","children":[["$","$L4",null,{"children":["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Mission"}]}],["$","$L4",null,{"delay":0.05,"children":["$","p",null,{"className":"mx-auto mt-4 max-w-4xl fluid-h2 font-heading uppercase leading-[1.02]","children":["Achieving ",["$","span",null,{"className":"display-em","children":"full human potential"}]," ","through technology — changing how people are informed, and how tech is used for good."]}]}],["$","$L4",null,{"delay":0.1,"children":["$","p",null,{"className":"mx-auto mt-6 max-w-2xl text-lg leading-relaxed text-muted","children":"What are the circumstances of the human being, and how can technology improve them? We see technology as a tool — part of the solution."}]}]]}]}],["$","section",null,{"className":"mx-auto max-w-6xl px-6 py-24","children":[["$","$L4",null,{"children":["$","div",null,{"className":"flex items-end justify-between","children":["$L6","$L7"]}]}],"$L8"]}],"$L9","$La","$Lb"]],["$Lc"],"$Ld"]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} +0:{"rsc":["$","$1","c",{"children":[[["$","section",null,{"className":"relative overflow-hidden border-b border-border","children":[["$","$L2",null,{"src":"/website/nyu-subway.jpg","alt":"","aria-hidden":true,"fill":true,"priority":true,"sizes":"100vw","className":"pointer-events-none object-cover object-center opacity-25"}],["$","div",null,{"aria-hidden":true,"className":"pointer-events-none absolute inset-0 bg-background/70"}],["$","span",null,{"className":"aura"}],["$","div",null,{"aria-hidden":true,"className":"pointer-events-none absolute inset-0","style":{"background":"radial-gradient(70% 60% at 20% 0%, color-mix(in oklab, var(--glow) 26%, transparent), transparent 65%)"}}],["$","$L3",null,{}],["$","div",null,{"className":"relative mx-auto max-w-6xl px-6 py-24 text-center sm:py-28","children":[["$","$L4",null,{"children":["$","p",null,{"className":"text-base uppercase tracking-[0.25em] text-accent sm:text-lg","children":"NYU CGA × Microsoft"}]}],["$","$L4",null,{"delay":0.15,"children":["$","$L5",null,{"statements":[{"lead":"Ethical Tech CoLab","headingClass":"text-[clamp(4.25rem,13vw,11rem)] leading-[0.88]","figure":["$","p",null,{"className":"font-serif uppercase leading-[0.95] tracking-tight text-foreground","style":{"fontSize":"clamp(1.9rem, 4vw, 3rem)"},"children":["Exploring technology to improve",["$","br",null,{"className":"hidden sm:block"}]," the"," ",["$","span",null,{"className":"display-em","children":"human condition"}],"."]}],"line":["$","p",null,{"className":"mx-auto mt-7 max-w-2xl leading-relaxed text-foreground/85","children":["A research collaboration between NYU's"," ",["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"link-underline text-accent hover:opacity-80","children":"Center for Global Affairs"}]," ","and Microsoft — changing the conversation on how people are informed, and how emerging technology can be used for good."]}]},{"lead":"Four questions. ","em":"One frontier.","figure":"14 projects in the portfolio","line":"Evacuation, cultural heritage, traceability, diplomacy — each question carried through to a fielded prototype, in the open.","cta":"Explore the portfolio","href":"/portfolio"},{"lead":"Run the ","em":"research","tail":".","figure":"21 demos you can open","line":"Not screenshots: the prototypes themselves, running in the browser with their source alongside.","cta":"Open the live demos","href":"/demos"},{"lead":"The research, ","em":"written up","tail":".","figure":"28 in the catalogue","line":"Every research question the CoLab takes on is written up academically, including what did not hold.","cta":"Read the publications","href":"/publications"},{"lead":"The people ","em":"building","tail":" this.","figure":"23 researchers across 3 cohorts","line":"Graduate researchers at NYU's Center for Global Affairs, with advisors and resident fellows alongside.","cta":"Meet the team","href":"/team"}],"label":"Statement","className":"mt-5"}]}]]}]]}],[["$","section",null,{"className":"border-b border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16 text-center","children":[["$","$L4",null,{"children":["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Mission"}]}],["$","$L4",null,{"delay":0.05,"children":["$","p",null,{"className":"mx-auto mt-4 max-w-4xl fluid-h2 font-heading uppercase leading-[1.02]","children":["Achieving ",["$","span",null,{"className":"display-em","children":"full human potential"}]," ","through technology — changing how people are informed, and how tech is used for good."]}]}],["$","$L4",null,{"delay":0.1,"children":["$","p",null,{"className":"mx-auto mt-6 max-w-2xl text-lg leading-relaxed text-muted","children":"What are the circumstances of the human being, and how can technology improve them? We see technology as a tool — part of the solution."}]}]]}]}],["$","section",null,{"className":"mx-auto max-w-6xl px-6 py-24","children":[["$","$L4",null,{"children":["$","div",null,{"className":"flex items-end justify-between","children":["$L6","$L7"]}]}],"$L8"]}],"$L9","$La","$Lb"]],["$Lc"],"$Ld"]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} 6:["$","div",null,{"children":[["$","$Le",null,{"href":"/portfolio","className":"link-underline text-xs uppercase tracking-wider text-muted transition-colors hover:text-accent","children":"Current projects"}],["$","h2",null,{"className":"mt-3 fluid-h2 font-heading uppercase","children":"Building at the frontier."}]]}] 7:["$","$Le",null,{"href":"/portfolio","className":"link-underline hidden text-sm text-accent sm:block","children":["View all ","four"," →"]}] 8:["$","$Lf",null,{"className":"mt-12 grid gap-px overflow-hidden rounded-2xl border border-border bg-border sm:grid-cols-2","children":[[["$","$L10","Evacuation",{"children":["$","$Le",null,{"href":"/portfolio#evacuation","className":"group card-glow flex h-full flex-col gap-3 bg-background p-8 transition-colors hover:bg-surface","children":[["$","$L11",null,{"variant":"Evacuation","className":"diagram-live mb-3 aspect-[16/7] w-full overflow-hidden rounded-xl border border-border bg-surface/60"}],["$","span",null,{"className":"font-mono text-xs text-muted","children":["01"," / ","Evacuation"]}],["$","h3",null,{"className":"font-heading text-2xl uppercase tracking-wide sm:text-3xl group-hover:text-accent","children":"How can AI inform evacuation decisions?"}]]}]}],["$","$L10","Cultural heritage",{"children":["$","$Le",null,{"href":"/portfolio#cultural-heritage","className":"group card-glow flex h-full flex-col gap-3 bg-background p-8 transition-colors hover:bg-surface","children":[["$","$L11",null,{"variant":"Cultural heritage","className":"diagram-live mb-3 aspect-[16/7] w-full overflow-hidden rounded-xl border border-border bg-surface/60"}],["$","span",null,{"className":"font-mono text-xs text-muted","children":["02"," / ","Cultural heritage"]}],["$","h3",null,{"className":"font-heading text-2xl uppercase tracking-wide sm:text-3xl group-hover:text-accent","children":"How can technology support the ethical return of cultural artifacts?"}]]}]}],["$","$L10","Traceability",{"children":["$","$Le",null,{"href":"/portfolio#traceability","className":"group card-glow flex h-full flex-col gap-3 bg-background p-8 transition-colors hover:bg-surface","children":[["$","$L11",null,{"variant":"Traceability","className":"diagram-live mb-3 aspect-[16/7] w-full overflow-hidden rounded-xl border border-border bg-surface/60"}],["$","span",null,{"className":"font-mono text-xs text-muted","children":["03"," / ","Traceability"]}],["$","h3",null,{"className":"font-heading text-2xl uppercase tracking-wide sm:text-3xl group-hover:text-accent","children":"How can ethical claims in supply chains be made verifiable?"}]]}]}],["$","$L10","Diplomacy",{"children":["$","$Le",null,{"href":"/portfolio#diplomacy","className":"group card-glow flex h-full flex-col gap-3 bg-background p-8 transition-colors hover:bg-surface","children":[["$","$L11",null,{"variant":"Diplomacy","className":"diagram-live mb-3 aspect-[16/7] w-full overflow-hidden rounded-xl border border-border bg-surface/60"}],["$","span",null,{"className":"font-mono text-xs text-muted","children":["04"," / ","Diplomacy"]}],["$","h3",null,{"className":"font-heading text-2xl uppercase tracking-wide sm:text-3xl group-hover:text-accent","children":"How can AI help practitioners rehearse high-stakes diplomacy?"}]]}]}]],false]}] diff --git a/static-site/__next._full.txt b/static-site/__next._full.txt index 390ae5f57..2361f8185 100644 --- a/static-site/__next._full.txt +++ b/static-site/__next._full.txt @@ -1,40 +1,40 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -f:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +f:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["",""],"q":"","i":false,"f":[[["",{"children":["__PAGE__",{}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{},null,false,null]},null,false,null],"$Le",false]],"m":"$undefined","G":["$f",["$L10"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -11:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"Link"] -12:I[85437,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"Image"] -13:I[25150,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"HeroField"] -14:I[82987,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"Reveal"] -15:I[53675,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"StatementCarousel"] -1e:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -20:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +0:{"P":null,"c":["",""],"q":"","i":false,"f":[[["",{"children":["__PAGE__",{}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{},null,false,null]},null,false,null],"$Le",false]],"m":"$undefined","G":["$f",["$L10"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +11:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"Link"] +12:I[85437,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"Image"] +13:I[25150,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"HeroField"] +14:I[82987,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"Reveal"] +15:I[53675,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"StatementCarousel"] +1e:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +20:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 21:"$Sreact.suspense" 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L11",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L11",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L11",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L11",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] b:["$","div",null,{"className":"mt-12 flex flex-col gap-2 border-t border-border pt-6 text-xs text-muted sm:flex-row sm:items-center sm:justify-between","children":[["$","span",null,{"children":["© ",2026," NYU Ethical Tech CoLab"]}],["$","span",null,{"children":"Four cohorts · est. 2024-2026"}]]}] c:["$","div",null,{"className":"mt-6 space-y-3 border-t border-border pt-6 text-[11px] leading-relaxed text-muted/80","children":[["$","p","0",{"children":"The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution."}],["$","p","1",{"children":"Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners."}]]}] -d:["$","$1","c",{"children":[[["$","section",null,{"className":"relative overflow-hidden border-b border-border","children":[["$","$L12",null,{"src":"/website/nyu-subway.jpg","alt":"","aria-hidden":true,"fill":true,"priority":true,"sizes":"100vw","className":"pointer-events-none object-cover object-center opacity-25"}],["$","div",null,{"aria-hidden":true,"className":"pointer-events-none absolute inset-0 bg-background/70"}],["$","span",null,{"className":"aura"}],["$","div",null,{"aria-hidden":true,"className":"pointer-events-none absolute inset-0","style":{"background":"radial-gradient(70% 60% at 20% 0%, color-mix(in oklab, var(--glow) 26%, transparent), transparent 65%)"}}],["$","$L13",null,{}],["$","div",null,{"className":"relative mx-auto max-w-6xl px-6 py-24 text-center sm:py-28","children":[["$","$L14",null,{"children":["$","p",null,{"className":"text-base uppercase tracking-[0.25em] text-accent sm:text-lg","children":"NYU CGA × Microsoft"}]}],["$","$L14",null,{"delay":0.15,"children":["$","$L15",null,{"statements":[{"lead":"Ethical Tech CoLab","headingClass":"text-[clamp(4.25rem,13vw,11rem)] leading-[0.88]","figure":["$","p",null,{"className":"font-serif uppercase leading-[0.95] tracking-tight text-foreground","style":{"fontSize":"clamp(1.9rem, 4vw, 3rem)"},"children":["Exploring technology to improve",["$","br",null,{"className":"hidden sm:block"}]," the"," ",["$","span",null,{"className":"display-em","children":"human condition"}],"."]}],"line":["$","p",null,{"className":"mx-auto mt-7 max-w-2xl leading-relaxed text-foreground/85","children":["A research collaboration between NYU's"," ",["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"link-underline text-accent hover:opacity-80","children":"Center for Global Affairs"}]," ","and Microsoft — changing the conversation on how people are informed, and how emerging technology can be used for good."]}]},{"lead":"Four questions. ","em":"One frontier.","figure":"14 projects in the portfolio","line":"Evacuation, cultural heritage, traceability, diplomacy — each question carried through to a fielded prototype, in the open.","cta":"Explore the portfolio","href":"/portfolio"},{"lead":"Run the ","em":"research","tail":".","figure":"18 demos you can open","line":"Not screenshots: the prototypes themselves, running in the browser with their source alongside.","cta":"Open the live demos","href":"/demos"},{"lead":"The research, ","em":"written up","tail":".","figure":"25 in the catalogue","line":"Every research question the CoLab takes on is written up academically, including what did not hold.","cta":"Read the publications","href":"/publications"},{"lead":"The people ","em":"building","tail":" this.","figure":"23 researchers across 3 cohorts","line":"Graduate researchers at NYU's Center for Global Affairs, with advisors and resident fellows alongside.","cta":"Meet the team","href":"/team"}],"label":"Statement","className":"mt-5"}]}]]}]]}],[["$","section",null,{"className":"border-b border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16 text-center","children":[["$","$L14",null,{"children":["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Mission"}]}],["$","$L14",null,{"delay":0.05,"children":["$","p",null,{"className":"mx-auto mt-4 max-w-4xl fluid-h2 font-heading uppercase leading-[1.02]","children":["Achieving ",["$","span",null,{"className":"display-em","children":"full human potential"}]," ","through technology — changing how people are informed, and how tech is used for good."]}]}],["$","$L14",null,{"delay":0.1,"children":["$","p",null,{"className":"mx-auto mt-6 max-w-2xl text-lg leading-relaxed text-muted","children":"What are the circumstances of the human being, and how can technology improve them? We see technology as a tool — part of the solution."}]}]]}]}],["$","section",null,{"className":"mx-auto max-w-6xl px-6 py-24","children":[["$","$L14",null,{"children":["$","div",null,{"className":"flex items-end justify-between","children":["$L16","$L17"]}]}],"$L18"]}],"$L19","$L1a","$L1b"]],["$L1c"],"$L1d"]}] +d:["$","$1","c",{"children":[[["$","section",null,{"className":"relative overflow-hidden border-b border-border","children":[["$","$L12",null,{"src":"/website/nyu-subway.jpg","alt":"","aria-hidden":true,"fill":true,"priority":true,"sizes":"100vw","className":"pointer-events-none object-cover object-center opacity-25"}],["$","div",null,{"aria-hidden":true,"className":"pointer-events-none absolute inset-0 bg-background/70"}],["$","span",null,{"className":"aura"}],["$","div",null,{"aria-hidden":true,"className":"pointer-events-none absolute inset-0","style":{"background":"radial-gradient(70% 60% at 20% 0%, color-mix(in oklab, var(--glow) 26%, transparent), transparent 65%)"}}],["$","$L13",null,{}],["$","div",null,{"className":"relative mx-auto max-w-6xl px-6 py-24 text-center sm:py-28","children":[["$","$L14",null,{"children":["$","p",null,{"className":"text-base uppercase tracking-[0.25em] text-accent sm:text-lg","children":"NYU CGA × Microsoft"}]}],["$","$L14",null,{"delay":0.15,"children":["$","$L15",null,{"statements":[{"lead":"Ethical Tech CoLab","headingClass":"text-[clamp(4.25rem,13vw,11rem)] leading-[0.88]","figure":["$","p",null,{"className":"font-serif uppercase leading-[0.95] tracking-tight text-foreground","style":{"fontSize":"clamp(1.9rem, 4vw, 3rem)"},"children":["Exploring technology to improve",["$","br",null,{"className":"hidden sm:block"}]," the"," ",["$","span",null,{"className":"display-em","children":"human condition"}],"."]}],"line":["$","p",null,{"className":"mx-auto mt-7 max-w-2xl leading-relaxed text-foreground/85","children":["A research collaboration between NYU's"," ",["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"link-underline text-accent hover:opacity-80","children":"Center for Global Affairs"}]," ","and Microsoft — changing the conversation on how people are informed, and how emerging technology can be used for good."]}]},{"lead":"Four questions. ","em":"One frontier.","figure":"14 projects in the portfolio","line":"Evacuation, cultural heritage, traceability, diplomacy — each question carried through to a fielded prototype, in the open.","cta":"Explore the portfolio","href":"/portfolio"},{"lead":"Run the ","em":"research","tail":".","figure":"21 demos you can open","line":"Not screenshots: the prototypes themselves, running in the browser with their source alongside.","cta":"Open the live demos","href":"/demos"},{"lead":"The research, ","em":"written up","tail":".","figure":"28 in the catalogue","line":"Every research question the CoLab takes on is written up academically, including what did not hold.","cta":"Read the publications","href":"/publications"},{"lead":"The people ","em":"building","tail":" this.","figure":"23 researchers across 3 cohorts","line":"Graduate researchers at NYU's Center for Global Affairs, with advisors and resident fellows alongside.","cta":"Meet the team","href":"/team"}],"label":"Statement","className":"mt-5"}]}]]}]]}],[["$","section",null,{"className":"border-b border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16 text-center","children":[["$","$L14",null,{"children":["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Mission"}]}],["$","$L14",null,{"delay":0.05,"children":["$","p",null,{"className":"mx-auto mt-4 max-w-4xl fluid-h2 font-heading uppercase leading-[1.02]","children":["Achieving ",["$","span",null,{"className":"display-em","children":"full human potential"}]," ","through technology — changing how people are informed, and how tech is used for good."]}]}],["$","$L14",null,{"delay":0.1,"children":["$","p",null,{"className":"mx-auto mt-6 max-w-2xl text-lg leading-relaxed text-muted","children":"What are the circumstances of the human being, and how can technology improve them? We see technology as a tool — part of the solution."}]}]]}]}],["$","section",null,{"className":"mx-auto max-w-6xl px-6 py-24","children":[["$","$L14",null,{"children":["$","div",null,{"className":"flex items-end justify-between","children":["$L16","$L17"]}]}],"$L18"]}],"$L19","$L1a","$L1b"]],["$L1c"],"$L1d"]}] e:["$","$1","h",{"children":[null,["$","$L1e",null,{"children":"$L1f"}],["$","div",null,{"hidden":true,"children":["$","$L20",null,{"children":["$","$21",null,{"name":"Next.Metadata","children":"$L22"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -10:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] -23:I[82987,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"Stagger"] -24:I[82987,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"StaggerItem"] -25:I[35852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"ProjectDiagram"] -26:I[12243,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"Tilt3D"] -2a:I[94376,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"Magnetic"] -2b:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +10:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +23:I[82987,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"Stagger"] +24:I[82987,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"StaggerItem"] +25:I[35852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"ProjectDiagram"] +26:I[12243,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"Tilt3D"] +2a:I[94376,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"Magnetic"] +2b:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 16:["$","div",null,{"children":[["$","$L11",null,{"href":"/portfolio","className":"link-underline text-xs uppercase tracking-wider text-muted transition-colors hover:text-accent","children":"Current projects"}],["$","h2",null,{"className":"mt-3 fluid-h2 font-heading uppercase","children":"Building at the frontier."}]]}] 17:["$","$L11",null,{"href":"/portfolio","className":"link-underline hidden text-sm text-accent sm:block","children":["View all ","four"," →"]}] 18:["$","$L23",null,{"className":"mt-12 grid gap-px overflow-hidden rounded-2xl border border-border bg-border sm:grid-cols-2","children":[[["$","$L24","Evacuation",{"children":["$","$L11",null,{"href":"/portfolio#evacuation","className":"group card-glow flex h-full flex-col gap-3 bg-background p-8 transition-colors hover:bg-surface","children":[["$","$L25",null,{"variant":"Evacuation","className":"diagram-live mb-3 aspect-[16/7] w-full overflow-hidden rounded-xl border border-border bg-surface/60"}],["$","span",null,{"className":"font-mono text-xs text-muted","children":["01"," / ","Evacuation"]}],["$","h3",null,{"className":"font-heading text-2xl uppercase tracking-wide sm:text-3xl group-hover:text-accent","children":"How can AI inform evacuation decisions?"}]]}]}],["$","$L24","Cultural heritage",{"children":["$","$L11",null,{"href":"/portfolio#cultural-heritage","className":"group card-glow flex h-full flex-col gap-3 bg-background p-8 transition-colors hover:bg-surface","children":[["$","$L25",null,{"variant":"Cultural heritage","className":"diagram-live mb-3 aspect-[16/7] w-full overflow-hidden rounded-xl border border-border bg-surface/60"}],["$","span",null,{"className":"font-mono text-xs text-muted","children":["02"," / ","Cultural heritage"]}],["$","h3",null,{"className":"font-heading text-2xl uppercase tracking-wide sm:text-3xl group-hover:text-accent","children":"How can technology support the ethical return of cultural artifacts?"}]]}]}],["$","$L24","Traceability",{"children":["$","$L11",null,{"href":"/portfolio#traceability","className":"group card-glow flex h-full flex-col gap-3 bg-background p-8 transition-colors hover:bg-surface","children":[["$","$L25",null,{"variant":"Traceability","className":"diagram-live mb-3 aspect-[16/7] w-full overflow-hidden rounded-xl border border-border bg-surface/60"}],["$","span",null,{"className":"font-mono text-xs text-muted","children":["03"," / ","Traceability"]}],["$","h3",null,{"className":"font-heading text-2xl uppercase tracking-wide sm:text-3xl group-hover:text-accent","children":"How can ethical claims in supply chains be made verifiable?"}]]}]}],["$","$L24","Diplomacy",{"children":["$","$L11",null,{"href":"/portfolio#diplomacy","className":"group card-glow flex h-full flex-col gap-3 bg-background p-8 transition-colors hover:bg-surface","children":[["$","$L25",null,{"variant":"Diplomacy","className":"diagram-live mb-3 aspect-[16/7] w-full overflow-hidden rounded-xl border border-border bg-surface/60"}],["$","span",null,{"className":"font-mono text-xs text-muted","children":["04"," / ","Diplomacy"]}],["$","h3",null,{"className":"font-heading text-2xl uppercase tracking-wide sm:text-3xl group-hover:text-accent","children":"How can AI help practitioners rehearse high-stakes diplomacy?"}]]}]}]],false]}] @@ -47,6 +47,6 @@ e:["$","$1","h",{"children":[null,["$","$L1e",null,{"children":"$L1f"}],["$","di 28:["$","$L26","02",{"max":7,"children":["$","article",null,{"className":"flex h-full flex-col rounded-2xl border bg-card p-7 transition-colors border-border hover:border-foreground/25","children":[["$","div",null,{"className":"flex items-center justify-end","children":["$","span",null,{"className":"rounded-full border border-border px-2.5 py-0.5 text-xs text-muted","children":"Past"}]}],["$","h3",null,{"className":"mt-5 font-heading text-3xl uppercase leading-none tracking-[0.06em] sm:text-4xl","children":"Fall 2025"}],["$","p",null,{"className":"mt-2 text-sm font-medium text-accent","children":"Prototyping and partner pilots."}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Technical spikes (multi-agent harnesses, verifiable credentials, geospatial pipelines) tested against real partner needs."}],["$","ul",null,{"className":"mt-5 space-y-2 text-sm text-foreground/85","children":[["$","li","8 researchers",{"className":"flex gap-2.5","children":[["$","span",null,{"aria-hidden":true,"className":"mt-1 text-accent","children":"◦"}],["$","span",null,{"children":"8 researchers"}]]}],["$","li","Forced Labor Structural Risk Index",{"className":"flex gap-2.5","children":[["$","span",null,{"aria-hidden":true,"className":"mt-1 text-accent","children":"◦"}],["$","span",null,{"children":"Forced Labor Structural Risk Index"}]]}],["$","li","AI Research Question Assistant",{"className":"flex gap-2.5","children":[["$","span",null,{"aria-hidden":true,"className":"mt-1 text-accent","children":"◦"}],["$","span",null,{"children":"AI Research Question Assistant"}]]}]]}],["$","div",null,{"className":"mt-auto pt-6","children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Full archive coming soon"}],["$","div",null,{"className":"mt-3 flex flex-wrap gap-3","children":[["$","$L11",null,{"href":"/portfolio","className":"inline-flex items-center gap-2 rounded-full bg-accent px-4 py-2 text-sm font-semibold text-background transition-opacity hover:opacity-90","children":"Portfolio →"}],["$","$L11",null,{"href":"/team#alumni-fall-2025","className":"inline-flex items-center gap-2 rounded-full border border-border px-4 py-2 text-sm font-medium transition-colors hover:border-accent hover:text-accent","children":"Meet the cohort →"}]]}]]}]]}]}] 29:["$","$L26","01",{"max":7,"children":["$","article",null,{"className":"flex h-full flex-col rounded-2xl border bg-card p-7 transition-colors border-border hover:border-foreground/25","children":[["$","div",null,{"className":"flex items-center justify-end","children":["$","span",null,{"className":"rounded-full border border-border px-2.5 py-0.5 text-xs text-muted","children":"Past"}]}],["$","h3",null,{"className":"mt-5 font-heading text-3xl uppercase leading-none tracking-[0.06em] sm:text-4xl","children":"Spring 2025"}],["$","p",null,{"className":"mt-2 text-sm font-medium text-accent","children":"First applied research projects."}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"The lab's first applied research projects, spanning online safety, sustainability, AI's footprint, and generative storytelling."}],["$","ul",null,{"className":"mt-5 space-y-2 text-sm text-foreground/85","children":[["$","li","8 researchers",{"className":"flex gap-2.5","children":[["$","span",null,{"aria-hidden":true,"className":"mt-1 text-accent","children":"◦"}],["$","span",null,{"children":"8 researchers"}]]}],["$","li","Online Grooming Prevention",{"className":"flex gap-2.5","children":[["$","span",null,{"aria-hidden":true,"className":"mt-1 text-accent","children":"◦"}],["$","span",null,{"children":"Online Grooming Prevention"}]]}],["$","li","ESG Labels & Certificates Transparency",{"className":"flex gap-2.5","children":[["$","span",null,{"aria-hidden":true,"className":"mt-1 text-accent","children":"◦"}],["$","span",null,{"children":"ESG Labels & Certificates Transparency"}]]}],["$","li","AI's Carbon Footprint",{"className":"flex gap-2.5","children":[["$","span",null,{"aria-hidden":true,"className":"mt-1 text-accent","children":"◦"}],["$","span",null,{"children":"AI's Carbon Footprint"}]]}],["$","li","Generative AI for Good — Avatar Storytelling",{"className":"flex gap-2.5","children":[["$","span",null,{"aria-hidden":true,"className":"mt-1 text-accent","children":"◦"}],["$","span",null,{"children":"Generative AI for Good — Avatar Storytelling"}]]}]]}],["$","div",null,{"className":"mt-auto pt-6","children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Full archive coming soon"}],["$","div",null,{"className":"mt-3 flex flex-wrap gap-3","children":[["$","$L11",null,{"href":"/portfolio","className":"inline-flex items-center gap-2 rounded-full bg-accent px-4 py-2 text-sm font-semibold text-background transition-opacity hover:opacity-90","children":"Portfolio →"}],["$","$L11",null,{"href":"/team#alumni-spring-2025","className":"inline-flex items-center gap-2 rounded-full border border-border px-4 py-2 text-sm font-medium transition-colors hover:border-accent hover:text-accent","children":"Meet the cohort →"}]]}]]}]]}]}] 1f:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -2d:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +2d:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 22:[["$","title","0",{"children":"NYU Ethical Tech CoLab · Emerging Tech, Human Condition"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","link","2",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","3",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","4",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L2d","5",{}]] 2c:null diff --git a/static-site/__next._head.txt b/static-site/__next._head.txt index 1bee38eda..aed0ebfe4 100644 --- a/static-site/__next._head.txt +++ b/static-site/__next._head.txt @@ -1,6 +1,6 @@ 1:"$Sreact.fragment" -2:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -3:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +2:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +3:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 4:"$Sreact.suspense" -5:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +5:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 0:{"rsc":["$","$1","h",{"children":[null,["$","$L2",null,{"children":[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]]}],["$","div",null,{"hidden":true,"children":["$","$L3",null,{"children":["$","$4",null,{"name":"Next.Metadata","children":[["$","title","0",{"children":"NYU Ethical Tech CoLab · Emerging Tech, Human Condition"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","link","2",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","3",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","4",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L5","5",{}]]}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/__next._index.txt b/static-site/__next._index.txt index 77a8746c2..e042d4406 100644 --- a/static-site/__next._index.txt +++ b/static-site/__next._index.txt @@ -1,13 +1,13 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -d:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"Link"] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +d:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"Link"] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} 8:["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$Ld",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$Ld",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$Ld",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$Ld",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] diff --git a/static-site/__next._tree.txt b/static-site/__next._tree.txt index 6eafd907a..8116b03ae 100644 --- a/static-site/__next._tree.txt +++ b/static-site/__next._tree.txt @@ -1,7 +1,7 @@ -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] 0:{"tree":{"name":"","param":null,"prefetchHints":16,"slots":{"children":{"name":"__PAGE__","param":null,"prefetchHints":0,"slots":null}}},"staleTime":300,"buildId":"static-snapshot"} diff --git a/static-site/_next/static/chunks/0-5whcwv1z55i.js b/static-site/_next/static/chunks/0-5whcwv1z55i.js deleted file mode 100644 index 77d737250..000000000 --- a/static-site/_next/static/chunks/0-5whcwv1z55i.js +++ /dev/null @@ -1 +0,0 @@ -(globalThis.TURBOPACK||(globalThis.TURBOPACK=[])).push(["object"==typeof document?document.currentScript:void 0,72328,e=>{"use strict";var a=e.i(71164),p=e.i(38544),i=e.i(71645);e.s(["useReducedMotion",0,function(){a.hasReducedMotionListener.current||(0,p.initPrefersReducedMotion)();let[e]=(0,i.useState)(a.prefersReducedMotion.current);return e}])},82987,e=>{"use strict";var a=e.i(43476),p=e.i(32181),i=e.i(72328);let s=[.16,1,.3,1],t={hidden:{},show:{transition:{staggerChildren:.07}}};e.s(["Reveal",0,function({children:e,className:t,delay:r=0,y:o=24,as:n="div"}){let l=(0,i.useReducedMotion)(),c=p.motion[n];return(0,a.jsx)(c,{className:t,initial:{opacity:0,y:l?0:o},whileInView:{opacity:1,y:0},viewport:{once:!0,margin:"-80px"},transition:{duration:.65,ease:s,delay:r},children:e})},"Stagger",0,function({children:e,className:i}){return(0,a.jsx)(p.motion.div,{className:i,variants:t,initial:"hidden",whileInView:"show",viewport:{once:!0,margin:"-60px"},children:e})},"StaggerItem",0,function({children:e,className:t}){let r=(0,i.useReducedMotion)();return(0,a.jsx)(p.motion.div,{className:t,variants:{hidden:{opacity:0,y:20*!r},show:{opacity:1,y:0,transition:{duration:.6,ease:s}}},children:e})}])},18620,e=>{"use strict";var a=e.i(43476),p=e.i(23475),i=e.i(18566);let s=[{label:"Overview",href:"/portfolio"},{label:"Live Demos",href:"/demos"},{label:"Publications",href:"/publications"},{label:"Media",href:"/media"},{label:"Newsletter",href:"/newsletter"}];e.s(["SectionTabs",0,function(){let e=(0,i.usePathname)();return(0,a.jsx)("div",{className:"border-b border-border bg-background/80 backdrop-blur",children:(0,a.jsx)("nav",{"aria-label":"Portfolio sections",className:"mx-auto flex max-w-6xl gap-6 px-6",children:s.map(i=>{let s="/portfolio"===i.href?"/portfolio"===e:e.startsWith(i.href);return(0,a.jsx)(p.Link,{href:i.href,"aria-current":s?"page":void 0,className:`-mb-px border-b-2 py-4 text-sm font-medium uppercase tracking-wider transition-colors ${s?"border-accent text-accent":"border-transparent text-muted hover:text-foreground"}`,children:i.label},i.href)})})})}])},34010,e=>{"use strict";var a=e.i(43476),p=e.i(71645);let i={startPage:0,singlePageBreakpoint:640,chromeHeight:150,chromeWidth:48,maxPageHeight:1100,flippingTime:700,drawShadow:!0,maxShadowOpacity:.4,keyboard:!0};async function s(a,p){let s={...i,...p};if(!s.pages.length)throw Error("mountFlipbook: `pages` is empty");let t=(await e.A(43413)).PageFlip,r=null,o=Math.min(Math.max(s.startPage,0),s.pages.length-1),n=!1,l=null,c=()=>{if(n)return;let e=s.singlePageBreakpoint>0&&window.matchMedia(`(max-width: ${s.singlePageBreakpoint}px)`).matches,{width:p,height:l}=function(e,a){let p=a.chromeHeight??i.chromeHeight,s=a.chromeWidth??i.chromeWidth,t=Math.min(a.viewportHeight-p,a.maxPageHeight??i.maxPageHeight),r=a.viewportWidth-s,o=Math.max(t,120),n=o*e,l=a.single?n:2*n;if(l>r){let e=r/l;n*=e,o*=e}return{width:Math.round(n),height:Math.round(o)}}(s.aspect,{single:e,viewportWidth:window.innerWidth,viewportHeight:window.innerHeight,chromeWidth:s.chromeWidth,chromeHeight:s.chromeHeight,maxPageHeight:s.maxPageHeight});if((r=new t(a,{width:p,height:l,size:"fixed",showCover:!0,usePortrait:e,maxShadowOpacity:s.maxShadowOpacity,mobileScrollSupport:!0,flippingTime:s.flippingTime,drawShadow:s.drawShadow})).loadFromImages(s.pages.slice()),r.on("flip",e=>{o=e.data,s.onFlip?.(o)}),o>0)try{r.turnToPage(o)}catch{}s.onReady?.()},b=()=>{if(r){try{r.destroy()}catch{}r=null,a.innerHTML=""}},u=()=>{l&&clearTimeout(l),l=setTimeout(()=>{b(),c()},200)},d=e=>{"ArrowLeft"===e.key?r?.flipPrev():"ArrowRight"===e.key&&r?.flipNext()};return c(),window.addEventListener("resize",u),s.keyboard&&document.addEventListener("keydown",d),{next:()=>r?.flipNext(),prev:()=>r?.flipPrev(),goTo:e=>r?.turnToPage(e),currentPage:()=>o,pageCount:()=>s.pages.length,destroy:()=>{n||(n=!0,l&&clearTimeout(l),window.removeEventListener("resize",u),document.removeEventListener("keydown",d),b())}}}function t(e,a,p){let i=document.createElement(e);return a&&(i.className=a),null!=p&&(i.textContent=p),i}async function r(e){let a=e.container??document.body,p=e.pages.length,i=t("div",`rab-overlay${e.className?` ${e.className}`:""}`);i.setAttribute("role","dialog"),i.setAttribute("aria-modal","true"),i.setAttribute("aria-label",`${e.title??"Document"} — page view`);let r=t("div","rab-chrome");r.append(t("span","rab-title",e.title??""));let o=t("div","rab-actions"),n=t("span","rab-counter");if(o.append(n),e.pdfUrl){let a=t("a","rab-btn rab-download","Download PDF ↗");a.href=e.pdfUrl,a.target="_blank",a.rel="noopener noreferrer",o.append(a)}let l=t("button","rab-btn","Close ✕");l.type="button",l.setAttribute("aria-label","Close book view"),o.append(l),r.append(o);let c=t("div","rab-stage"),b=t("button","rab-arrow rab-arrow-prev","‹");b.type="button",b.setAttribute("aria-label","Previous page");let u=t("button","rab-arrow rab-arrow-next","›");u.type="button",u.setAttribute("aria-label","Next page");let d=t("div","rab-book"),g=t("p","rab-loading","Opening the book…"),w=t("div","rab-book-wrap");w.append(d,g),c.append(b,w,u);let h=e.hint??"Use the arrows or ← → keys to turn pages · Esc to close",m=t("p","rab-hint",h);i.append(r,c),h&&i.append(m),a.append(i);let f=e=>{n.textContent=p?`${Math.min(e+1,p)} / ${p}`:""};f(e.startPage??0);let x=document.body.style.overflow;document.body.style.overflow="hidden";let v=null,y=!1,k=()=>{y||(y=!0,document.removeEventListener("keydown",j),document.body.style.overflow=x,v?.destroy(),i.remove(),e.onClose?.())},j=e=>{"Escape"===e.key&&k()};document.addEventListener("keydown",j),l.addEventListener("click",k);try{v=await s(d,{pages:e.pages,aspect:e.aspect,startPage:e.startPage,onFlip:a=>{f(a),e.onFlip?.(a)},onReady:()=>g.remove()})}catch(e){throw k(),e}return y?(v.destroy(),{close:k,next:()=>{},prev:()=>{}}):(b.addEventListener("click",()=>v?.prev()),u.addEventListener("click",()=>v?.next()),{close:k,next:()=>v?.next(),prev:()=>v?.prev()})}function o({pages:e,aspect:i,title:s,pdfUrl:t,className:n,children:l,overlayClassName:c,hint:b,open:u,onOpenChange:d,onFlip:g}){let[w,h]=(0,p.useState)(!1),m=u??w,f=(0,p.useRef)(null),x=(0,p.useRef)(g);x.current=g;let v=(0,p.useCallback)(e=>{void 0===u&&h(e),d?.(e)},[u,d]);return(0,p.useEffect)(()=>{if(!m)return;let a=!1;return r({pages:e,aspect:i,title:s,pdfUrl:t,className:c,hint:b,onFlip:e=>x.current?.(e),onClose:()=>{a||v(!1)}}).then(e=>{a?e.close():f.current=e}).catch(e=>{console.error("read-as-book: failed to open the viewer",e),a||v(!1)}),()=>{a=!0,f.current?.close(),f.current=null}},[m,e,i,s,t,c,b]),(0,a.jsx)("button",{type:"button",onClick:()=>v(!0),className:n??"rab-trigger",children:l??(0,a.jsxs)(a.Fragment,{children:["Read as book ",(0,a.jsx)("span",{"aria-hidden":!0,children:"📖"})]})})}var n=e.i(93209);e.s(["ReportBook",0,function({pages:e,aspect:i,title:s,pdfUrl:t,className:r,children:l}){let c=(0,p.useMemo)(()=>e.map(e=>(0,n.asset)(`/${e}`)),[e]);return(0,a.jsx)(o,{pages:c,aspect:i,title:s,pdfUrl:t,className:r??"btn-sweep inline-flex items-center gap-2 rounded-full border border-border px-5 py-2.5 text-sm font-semibold text-foreground transition-colors hover:border-border-strong",children:l})}],34010)},12243,e=>{"use strict";var a=e.i(43476),p=e.i(32181),i=e.i(87652),s=e.i(91994),t=e.i(72328),r=e.i(71645);e.s(["Tilt3D",0,function({children:e,className:o="",max:n=9,glare:l=!0,disabled:c=!1}){let b=(0,t.useReducedMotion)(),u=(0,r.useRef)(null),d=(0,i.useMotionValue)(0),g=(0,i.useMotionValue)(0),w=(0,s.useSpring)(d,{stiffness:170,damping:16,mass:.5}),h=(0,s.useSpring)(g,{stiffness:170,damping:16,mass:.5});return c?(0,a.jsx)("div",{className:`h-full ${o}`,children:e}):(0,a.jsx)("div",{className:"tilt-scene h-full",children:(0,a.jsx)(p.motion.div,{ref:u,onPointerMove:function(e){let a=u.current;if(b||!a)return;let p=a.getBoundingClientRect(),i=(e.clientX-p.left)/p.width,s=(e.clientY-p.top)/p.height;g.set((i-.5)*2*n),d.set((.5-s)*2*n),a.style.setProperty("--mx",`${100*i}%`),a.style.setProperty("--my",`${100*s}%`),l&&(a.style.setProperty("--gx",`${100*i}%`),a.style.setProperty("--gy",`${100*s}%`),a.style.setProperty("--gop","1"))},onPointerLeave:function(){d.set(0),g.set(0),u.current?.style.setProperty("--gop","0")},style:{rotateX:w,rotateY:h,transformStyle:"preserve-3d"},className:`tilt-card h-full ${l?"tilt-glare":""} ${o}`,children:e})})}])},88302,e=>{"use strict";var a=e.i(43476),p=e.i(71645);e.s(["PosterRail",0,function({title:e,count:i,countNoun:s,ariaLabel:t,children:r}){let o=(0,p.useRef)(null),[n,l]=(0,p.useState)(!0),[c,b]=(0,p.useState)(!0),u=(0,p.useCallback)(()=>{let e=o.current;e&&(l(e.scrollLeft<=1),b(e.scrollLeft>=e.scrollWidth-e.clientWidth-1))},[]);(0,p.useEffect)(()=>{u();let e=o.current;if(!e)return;let a=new ResizeObserver(u);return a.observe(e),()=>a.disconnect()},[u,r]);let d=e=>{let a=o.current;a&&a.scrollBy({left:e*a.clientWidth*.85,behavior:"smooth"})},g=!(n&&c),w=e=>`pointer-events-auto grid h-9 w-9 place-items-center rounded-full border border-border bg-background/90 text-foreground backdrop-blur transition-all ${e?"opacity-100 hover:border-accent hover:text-accent":"cursor-default opacity-0"}`;return(0,a.jsxs)("section",{children:[(0,a.jsxs)("div",{className:"flex items-baseline justify-between gap-4 border-b border-border pb-3",children:[(0,a.jsx)("h2",{className:"font-heading text-2xl uppercase tracking-wide sm:text-3xl",children:e}),(0,a.jsxs)("span",{className:"shrink-0 font-mono text-xs text-muted",children:[i," ",1===i?s:`${s}s`]})]}),(0,a.jsxs)("div",{className:"relative mt-6",children:[(0,a.jsx)("div",{ref:o,onScroll:u,tabIndex:0,role:"group","aria-label":t,className:"no-scrollbar -mx-6 flex snap-x snap-mandatory scroll-pl-6 gap-5 overflow-x-auto px-6 pb-2 focus-visible:outline-none",children:r}),g&&(0,a.jsxs)("div",{className:"pointer-events-none absolute inset-y-0 left-0 right-0 hidden items-center justify-between sm:flex",children:[(0,a.jsx)("button",{type:"button",onClick:()=>d(-1),disabled:n,"aria-label":`Scroll ${e} left`,className:`-ml-4 ${w(!n)}`,children:(0,a.jsx)("span",{"aria-hidden":!0,children:"←"})}),(0,a.jsx)("button",{type:"button",onClick:()=>d(1),disabled:c,"aria-label":`Scroll ${e} right`,className:`-mr-4 ${w(!c)}`,children:(0,a.jsx)("span",{"aria-hidden":!0,children:"→"})})]})]})]})}])},46973,e=>{"use strict";var a=e.i(43476),p=e.i(71645),i=e.i(23475),s=e.i(89042),t=e.i(88302),r=e.i(12243),o=e.i(34010);let n={"after-the-corridor":{pages:["publications/after-the-corridor/pages/p01.webp","publications/after-the-corridor/pages/p02.webp","publications/after-the-corridor/pages/p03.webp","publications/after-the-corridor/pages/p04.webp","publications/after-the-corridor/pages/p05.webp","publications/after-the-corridor/pages/p06.webp","publications/after-the-corridor/pages/p07.webp","publications/after-the-corridor/pages/p08.webp","publications/after-the-corridor/pages/p09.webp","publications/after-the-corridor/pages/p10.webp","publications/after-the-corridor/pages/p11.webp","publications/after-the-corridor/pages/p12.webp","publications/after-the-corridor/pages/p13.webp","publications/after-the-corridor/pages/p14.webp","publications/after-the-corridor/pages/p15.webp","publications/after-the-corridor/pages/p16.webp","publications/after-the-corridor/pages/p17.webp","publications/after-the-corridor/pages/p18.webp","publications/after-the-corridor/pages/p19.webp","publications/after-the-corridor/pages/p20.webp","publications/after-the-corridor/pages/p21.webp","publications/after-the-corridor/pages/p22.webp"],aspect:.7727,pdf:"/publications/after-the-corridor/report.pdf"},"ai-carbon-footprint":{pages:["publications/ai-carbon-footprint/pages/p01.webp","publications/ai-carbon-footprint/pages/p02.webp","publications/ai-carbon-footprint/pages/p03.webp","publications/ai-carbon-footprint/pages/p04.webp","publications/ai-carbon-footprint/pages/p05.webp","publications/ai-carbon-footprint/pages/p06.webp","publications/ai-carbon-footprint/pages/p07.webp","publications/ai-carbon-footprint/pages/p08.webp","publications/ai-carbon-footprint/pages/p09.webp","publications/ai-carbon-footprint/pages/p10.webp","publications/ai-carbon-footprint/pages/p11.webp","publications/ai-carbon-footprint/pages/p12.webp","publications/ai-carbon-footprint/pages/p13.webp","publications/ai-carbon-footprint/pages/p14.webp","publications/ai-carbon-footprint/pages/p15.webp","publications/ai-carbon-footprint/pages/p16.webp","publications/ai-carbon-footprint/pages/p17.webp","publications/ai-carbon-footprint/pages/p18.webp"],aspect:.7067,pdf:"/publications/ai-carbon-footprint/report.pdf"},"ai-models-research":{pages:["publications/ai-models-research/pages/p01.webp","publications/ai-models-research/pages/p02.webp","publications/ai-models-research/pages/p03.webp","publications/ai-models-research/pages/p04.webp","publications/ai-models-research/pages/p05.webp","publications/ai-models-research/pages/p06.webp","publications/ai-models-research/pages/p07.webp","publications/ai-models-research/pages/p08.webp","publications/ai-models-research/pages/p09.webp","publications/ai-models-research/pages/p10.webp","publications/ai-models-research/pages/p11.webp","publications/ai-models-research/pages/p12.webp","publications/ai-models-research/pages/p13.webp","publications/ai-models-research/pages/p14.webp","publications/ai-models-research/pages/p15.webp","publications/ai-models-research/pages/p16.webp","publications/ai-models-research/pages/p17.webp","publications/ai-models-research/pages/p18.webp","publications/ai-models-research/pages/p19.webp"],aspect:.7067,pdf:"/publications/ai-models-research/report.pdf"},"ai-research-assistant":{pages:["publications/ai-research-assistant/pages/p01.webp","publications/ai-research-assistant/pages/p02.webp","publications/ai-research-assistant/pages/p03.webp","publications/ai-research-assistant/pages/p04.webp","publications/ai-research-assistant/pages/p05.webp","publications/ai-research-assistant/pages/p06.webp","publications/ai-research-assistant/pages/p07.webp","publications/ai-research-assistant/pages/p08.webp","publications/ai-research-assistant/pages/p09.webp","publications/ai-research-assistant/pages/p10.webp"],aspect:.7067,pdf:"/publications/ai-research-assistant/report.pdf"},cerai:{pages:["publications/cerai/pages/p01.webp","publications/cerai/pages/p02.webp","publications/cerai/pages/p03.webp","publications/cerai/pages/p04.webp","publications/cerai/pages/p05.webp","publications/cerai/pages/p06.webp","publications/cerai/pages/p07.webp","publications/cerai/pages/p08.webp","publications/cerai/pages/p09.webp","publications/cerai/pages/p10.webp","publications/cerai/pages/p11.webp","publications/cerai/pages/p12.webp","publications/cerai/pages/p13.webp"],aspect:.7067,pdf:"/publications/cerai/report.pdf"},"digital-provenance-passport":{pages:["publications/digital-provenance-passport/pages/p01.webp","publications/digital-provenance-passport/pages/p02.webp","publications/digital-provenance-passport/pages/p03.webp","publications/digital-provenance-passport/pages/p04.webp","publications/digital-provenance-passport/pages/p05.webp","publications/digital-provenance-passport/pages/p06.webp","publications/digital-provenance-passport/pages/p07.webp","publications/digital-provenance-passport/pages/p08.webp","publications/digital-provenance-passport/pages/p09.webp","publications/digital-provenance-passport/pages/p10.webp","publications/digital-provenance-passport/pages/p11.webp","publications/digital-provenance-passport/pages/p12.webp","publications/digital-provenance-passport/pages/p13.webp","publications/digital-provenance-passport/pages/p14.webp","publications/digital-provenance-passport/pages/p15.webp","publications/digital-provenance-passport/pages/p16.webp","publications/digital-provenance-passport/pages/p17.webp","publications/digital-provenance-passport/pages/p18.webp","publications/digital-provenance-passport/pages/p19.webp","publications/digital-provenance-passport/pages/p20.webp","publications/digital-provenance-passport/pages/p21.webp","publications/digital-provenance-passport/pages/p22.webp","publications/digital-provenance-passport/pages/p23.webp","publications/digital-provenance-passport/pages/p24.webp","publications/digital-provenance-passport/pages/p25.webp","publications/digital-provenance-passport/pages/p26.webp","publications/digital-provenance-passport/pages/p27.webp","publications/digital-provenance-passport/pages/p28.webp"],aspect:.7067,pdf:"/publications/digital-provenance-passport/report.pdf"},"diplomatic-simulator":{pages:["publications/diplomatic-simulator/pages/p01.webp","publications/diplomatic-simulator/pages/p02.webp","publications/diplomatic-simulator/pages/p03.webp","publications/diplomatic-simulator/pages/p04.webp","publications/diplomatic-simulator/pages/p05.webp","publications/diplomatic-simulator/pages/p06.webp","publications/diplomatic-simulator/pages/p07.webp","publications/diplomatic-simulator/pages/p08.webp","publications/diplomatic-simulator/pages/p09.webp","publications/diplomatic-simulator/pages/p10.webp","publications/diplomatic-simulator/pages/p11.webp","publications/diplomatic-simulator/pages/p12.webp","publications/diplomatic-simulator/pages/p13.webp","publications/diplomatic-simulator/pages/p14.webp","publications/diplomatic-simulator/pages/p15.webp","publications/diplomatic-simulator/pages/p16.webp","publications/diplomatic-simulator/pages/p17.webp","publications/diplomatic-simulator/pages/p18.webp","publications/diplomatic-simulator/pages/p19.webp","publications/diplomatic-simulator/pages/p20.webp","publications/diplomatic-simulator/pages/p21.webp","publications/diplomatic-simulator/pages/p22.webp","publications/diplomatic-simulator/pages/p23.webp"],aspect:.7067,pdf:"/publications/diplomatic-simulator/report.pdf"},ercf:{pages:["publications/ercf/pages/p01.webp","publications/ercf/pages/p02.webp","publications/ercf/pages/p03.webp","publications/ercf/pages/p04.webp","publications/ercf/pages/p05.webp","publications/ercf/pages/p06.webp","publications/ercf/pages/p07.webp","publications/ercf/pages/p08.webp","publications/ercf/pages/p09.webp","publications/ercf/pages/p10.webp","publications/ercf/pages/p11.webp","publications/ercf/pages/p12.webp","publications/ercf/pages/p13.webp","publications/ercf/pages/p14.webp","publications/ercf/pages/p15.webp","publications/ercf/pages/p16.webp","publications/ercf/pages/p17.webp","publications/ercf/pages/p18.webp","publications/ercf/pages/p19.webp","publications/ercf/pages/p20.webp","publications/ercf/pages/p21.webp","publications/ercf/pages/p22.webp","publications/ercf/pages/p23.webp","publications/ercf/pages/p24.webp","publications/ercf/pages/p25.webp","publications/ercf/pages/p26.webp","publications/ercf/pages/p27.webp","publications/ercf/pages/p28.webp","publications/ercf/pages/p29.webp","publications/ercf/pages/p30.webp","publications/ercf/pages/p31.webp","publications/ercf/pages/p32.webp"],aspect:.7067,pdf:"/publications/ercf/report.pdf"},erus:{pages:["publications/erus/pages/p01.webp","publications/erus/pages/p02.webp","publications/erus/pages/p03.webp","publications/erus/pages/p04.webp","publications/erus/pages/p05.webp","publications/erus/pages/p06.webp","publications/erus/pages/p07.webp","publications/erus/pages/p08.webp","publications/erus/pages/p09.webp","publications/erus/pages/p10.webp","publications/erus/pages/p11.webp","publications/erus/pages/p12.webp","publications/erus/pages/p13.webp","publications/erus/pages/p14.webp","publications/erus/pages/p15.webp","publications/erus/pages/p16.webp","publications/erus/pages/p17.webp","publications/erus/pages/p18.webp","publications/erus/pages/p19.webp","publications/erus/pages/p20.webp","publications/erus/pages/p21.webp","publications/erus/pages/p22.webp","publications/erus/pages/p23.webp","publications/erus/pages/p24.webp"],aspect:.7067,pdf:"/publications/erus/report.pdf"},"evacuation-inform-index":{pages:["publications/evacuation-inform-index/pages/p01.webp","publications/evacuation-inform-index/pages/p02.webp","publications/evacuation-inform-index/pages/p03.webp","publications/evacuation-inform-index/pages/p04.webp","publications/evacuation-inform-index/pages/p05.webp","publications/evacuation-inform-index/pages/p06.webp","publications/evacuation-inform-index/pages/p07.webp","publications/evacuation-inform-index/pages/p08.webp","publications/evacuation-inform-index/pages/p09.webp","publications/evacuation-inform-index/pages/p10.webp","publications/evacuation-inform-index/pages/p11.webp"],aspect:.7067,pdf:"/publications/evacuation-inform-index/report.pdf"},"evacuation-simulation":{pages:["publications/evacuation-simulation/pages/p01.webp","publications/evacuation-simulation/pages/p02.webp","publications/evacuation-simulation/pages/p03.webp","publications/evacuation-simulation/pages/p04.webp","publications/evacuation-simulation/pages/p05.webp","publications/evacuation-simulation/pages/p06.webp","publications/evacuation-simulation/pages/p07.webp","publications/evacuation-simulation/pages/p08.webp","publications/evacuation-simulation/pages/p09.webp","publications/evacuation-simulation/pages/p10.webp","publications/evacuation-simulation/pages/p11.webp","publications/evacuation-simulation/pages/p12.webp","publications/evacuation-simulation/pages/p13.webp","publications/evacuation-simulation/pages/p14.webp","publications/evacuation-simulation/pages/p15.webp","publications/evacuation-simulation/pages/p16.webp","publications/evacuation-simulation/pages/p17.webp","publications/evacuation-simulation/pages/p18.webp","publications/evacuation-simulation/pages/p19.webp","publications/evacuation-simulation/pages/p20.webp","publications/evacuation-simulation/pages/p21.webp","publications/evacuation-simulation/pages/p22.webp"],aspect:.7067,pdf:"/publications/evacuation-simulation/report.pdf"},"forced-labor-structural-risk-index":{pages:["publications/forced-labor-structural-risk-index/pages/p01.webp","publications/forced-labor-structural-risk-index/pages/p02.webp","publications/forced-labor-structural-risk-index/pages/p03.webp","publications/forced-labor-structural-risk-index/pages/p04.webp","publications/forced-labor-structural-risk-index/pages/p05.webp","publications/forced-labor-structural-risk-index/pages/p06.webp","publications/forced-labor-structural-risk-index/pages/p07.webp","publications/forced-labor-structural-risk-index/pages/p08.webp","publications/forced-labor-structural-risk-index/pages/p09.webp","publications/forced-labor-structural-risk-index/pages/p10.webp","publications/forced-labor-structural-risk-index/pages/p11.webp","publications/forced-labor-structural-risk-index/pages/p12.webp","publications/forced-labor-structural-risk-index/pages/p13.webp","publications/forced-labor-structural-risk-index/pages/p14.webp","publications/forced-labor-structural-risk-index/pages/p15.webp","publications/forced-labor-structural-risk-index/pages/p16.webp","publications/forced-labor-structural-risk-index/pages/p17.webp","publications/forced-labor-structural-risk-index/pages/p18.webp","publications/forced-labor-structural-risk-index/pages/p19.webp","publications/forced-labor-structural-risk-index/pages/p20.webp","publications/forced-labor-structural-risk-index/pages/p21.webp","publications/forced-labor-structural-risk-index/pages/p22.webp","publications/forced-labor-structural-risk-index/pages/p23.webp","publications/forced-labor-structural-risk-index/pages/p24.webp","publications/forced-labor-structural-risk-index/pages/p25.webp"],aspect:.7067,pdf:"/publications/forced-labor-structural-risk-index/report.pdf"},haste:{pages:["publications/haste/pages/p01.webp","publications/haste/pages/p02.webp","publications/haste/pages/p03.webp","publications/haste/pages/p04.webp","publications/haste/pages/p05.webp","publications/haste/pages/p06.webp","publications/haste/pages/p07.webp","publications/haste/pages/p08.webp","publications/haste/pages/p09.webp","publications/haste/pages/p10.webp","publications/haste/pages/p11.webp","publications/haste/pages/p12.webp","publications/haste/pages/p13.webp","publications/haste/pages/p14.webp","publications/haste/pages/p15.webp","publications/haste/pages/p16.webp","publications/haste/pages/p17.webp","publications/haste/pages/p18.webp","publications/haste/pages/p19.webp","publications/haste/pages/p20.webp","publications/haste/pages/p21.webp","publications/haste/pages/p22.webp","publications/haste/pages/p23.webp","publications/haste/pages/p24.webp","publications/haste/pages/p25.webp","publications/haste/pages/p26.webp"],aspect:.7067,pdf:"/publications/haste/report.pdf"},"mariupol-severity-model":{pages:["publications/mariupol-severity-model/pages/p01.webp","publications/mariupol-severity-model/pages/p02.webp","publications/mariupol-severity-model/pages/p03.webp","publications/mariupol-severity-model/pages/p04.webp","publications/mariupol-severity-model/pages/p05.webp","publications/mariupol-severity-model/pages/p06.webp","publications/mariupol-severity-model/pages/p07.webp","publications/mariupol-severity-model/pages/p08.webp","publications/mariupol-severity-model/pages/p09.webp","publications/mariupol-severity-model/pages/p10.webp","publications/mariupol-severity-model/pages/p11.webp","publications/mariupol-severity-model/pages/p12.webp","publications/mariupol-severity-model/pages/p13.webp","publications/mariupol-severity-model/pages/p14.webp","publications/mariupol-severity-model/pages/p15.webp","publications/mariupol-severity-model/pages/p16.webp","publications/mariupol-severity-model/pages/p17.webp","publications/mariupol-severity-model/pages/p18.webp","publications/mariupol-severity-model/pages/p19.webp","publications/mariupol-severity-model/pages/p20.webp","publications/mariupol-severity-model/pages/p21.webp","publications/mariupol-severity-model/pages/p22.webp","publications/mariupol-severity-model/pages/p23.webp","publications/mariupol-severity-model/pages/p24.webp","publications/mariupol-severity-model/pages/p25.webp","publications/mariupol-severity-model/pages/p26.webp"],aspect:.7067,pdf:"/publications/mariupol-severity-model/report.pdf"},"provenance-search":{pages:["publications/provenance-search/pages/p01.webp","publications/provenance-search/pages/p02.webp","publications/provenance-search/pages/p03.webp","publications/provenance-search/pages/p04.webp","publications/provenance-search/pages/p05.webp","publications/provenance-search/pages/p06.webp","publications/provenance-search/pages/p07.webp","publications/provenance-search/pages/p08.webp","publications/provenance-search/pages/p09.webp","publications/provenance-search/pages/p10.webp","publications/provenance-search/pages/p11.webp","publications/provenance-search/pages/p12.webp","publications/provenance-search/pages/p13.webp","publications/provenance-search/pages/p14.webp","publications/provenance-search/pages/p15.webp","publications/provenance-search/pages/p16.webp","publications/provenance-search/pages/p17.webp","publications/provenance-search/pages/p18.webp","publications/provenance-search/pages/p19.webp","publications/provenance-search/pages/p20.webp"],aspect:.7067,pdf:"/publications/provenance-search/report.pdf"},vango:{pages:["publications/vango/pages/p01.webp","publications/vango/pages/p02.webp","publications/vango/pages/p03.webp","publications/vango/pages/p04.webp","publications/vango/pages/p05.webp","publications/vango/pages/p06.webp","publications/vango/pages/p07.webp","publications/vango/pages/p08.webp","publications/vango/pages/p09.webp","publications/vango/pages/p10.webp","publications/vango/pages/p11.webp","publications/vango/pages/p12.webp","publications/vango/pages/p13.webp","publications/vango/pages/p14.webp","publications/vango/pages/p15.webp","publications/vango/pages/p16.webp"],aspect:.7067,pdf:"/publications/vango/report.pdf"},"war-games":{pages:["publications/war-games/pages/p01.webp","publications/war-games/pages/p02.webp","publications/war-games/pages/p03.webp","publications/war-games/pages/p04.webp","publications/war-games/pages/p05.webp","publications/war-games/pages/p06.webp","publications/war-games/pages/p07.webp","publications/war-games/pages/p08.webp","publications/war-games/pages/p09.webp","publications/war-games/pages/p10.webp","publications/war-games/pages/p11.webp","publications/war-games/pages/p12.webp","publications/war-games/pages/p13.webp","publications/war-games/pages/p14.webp","publications/war-games/pages/p15.webp"],aspect:.7067,pdf:"/publications/war-games/report.pdf"},"what-is-ethical-ai":{pages:["publications/what-is-ethical-ai/pages/p01.webp","publications/what-is-ethical-ai/pages/p02.webp","publications/what-is-ethical-ai/pages/p03.webp","publications/what-is-ethical-ai/pages/p04.webp","publications/what-is-ethical-ai/pages/p05.webp","publications/what-is-ethical-ai/pages/p06.webp","publications/what-is-ethical-ai/pages/p07.webp","publications/what-is-ethical-ai/pages/p08.webp","publications/what-is-ethical-ai/pages/p09.webp","publications/what-is-ethical-ai/pages/p10.webp","publications/what-is-ethical-ai/pages/p11.webp","publications/what-is-ethical-ai/pages/p12.webp","publications/what-is-ethical-ai/pages/p13.webp","publications/what-is-ethical-ai/pages/p14.webp","publications/what-is-ethical-ai/pages/p15.webp","publications/what-is-ethical-ai/pages/p16.webp","publications/what-is-ethical-ai/pages/p17.webp","publications/what-is-ethical-ai/pages/p18.webp","publications/what-is-ethical-ai/pages/p19.webp","publications/what-is-ethical-ai/pages/p20.webp","publications/what-is-ethical-ai/pages/p21.webp","publications/what-is-ethical-ai/pages/p22.webp","publications/what-is-ethical-ai/pages/p23.webp","publications/what-is-ethical-ai/pages/p24.webp","publications/what-is-ethical-ai/pages/p25.webp","publications/what-is-ethical-ai/pages/p26.webp","publications/what-is-ethical-ai/pages/p27.webp","publications/what-is-ethical-ai/pages/p28.webp","publications/what-is-ethical-ai/pages/p29.webp","publications/what-is-ethical-ai/pages/p30.webp","publications/what-is-ethical-ai/pages/p31.webp","publications/what-is-ethical-ai/pages/p32.webp","publications/what-is-ethical-ai/pages/p33.webp","publications/what-is-ethical-ai/pages/p34.webp","publications/what-is-ethical-ai/pages/p35.webp","publications/what-is-ethical-ai/pages/p36.webp","publications/what-is-ethical-ai/pages/p37.webp","publications/what-is-ethical-ai/pages/p38.webp","publications/what-is-ethical-ai/pages/p39.webp","publications/what-is-ethical-ai/pages/p40.webp","publications/what-is-ethical-ai/pages/p41.webp","publications/what-is-ethical-ai/pages/p42.webp","publications/what-is-ethical-ai/pages/p43.webp","publications/what-is-ethical-ai/pages/p44.webp","publications/what-is-ethical-ai/pages/p45.webp","publications/what-is-ethical-ai/pages/p46.webp"],aspect:.7727,pdf:"/publications/what-is-ethical-ai/report.pdf"}};function l(e){if(!e)return;let a=e.match(/\/publications\/([^/?#]+)/);return a?n[a[1]]:void 0}var c=e.i(93209);let b={"Artificial Intelligence":["#0f3b57","#08131f"],Guidelines:["#1d2440","#0d1020"],Evacuation:["#3b1878","#160d1c"],"Cultural heritage":["#4a1d3d","#1a0d18"],Traceability:["#123a3a","#0a1a1c"],Diplomacy:["#2a2160","#100d20"],Sustainability:["#1c3a24","#0b1710"],"Disaster response":["#4a2417","#1a0e0a"]};function u(e,a=0){let[p,i]=b[e]??["#241a35","#120d1c"];return`linear-gradient(${130+31*a%60}deg, ${p} 0%, ${i} ${64+17*a%26}%)`}function d({topic:e,seed:p}){let i="var(--poster-accent)",s="color-mix(in oklab, var(--poster-accent) 26%, transparent)",t="color-mix(in oklab, var(--poster-accent) 12%, transparent)",r=(e,a=1)=>((37*p+11*e)%13-6)/6*a,o=.74+13*p%7/100;return(0,a.jsxs)("svg",{viewBox:"0 0 200 300","aria-hidden":!0,className:"absolute inset-0 h-full w-full",preserveAspectRatio:"xMidYMid slice",children:[(0,a.jsx)("g",{stroke:t,strokeWidth:.6,children:Array.from({length:9}).map((e,p)=>(0,a.jsx)("line",{x1:0,y1:30+30*p,x2:200,y2:30+30*p},p))}),(0,a.jsx)("g",{transform:`translate(${100-100*o+(7*p%11-5)} ${-34-5*p%14}) scale(${o}) rotate(${23*p%9-4} 100 170)`,children:(()=>{switch(e){case"Evacuation":return(0,a.jsxs)(a.Fragment,{children:[(0,a.jsx)("circle",{cx:40,cy:250,r:5,fill:i}),[{d:`M40 250 C 70 ${210+r(1,8)}, 120 ${200+r(2,8)}, 160 150`,dash:!1},{d:`M40 250 C 60 ${190+r(3,8)}, 90 ${140+r(4,8)}, 150 96`,dash:!0},{d:`M40 250 C 80 ${230+r(5,8)}, 130 ${230+r(6,8)}, 168 200`,dash:!1}].map((e,p)=>(0,a.jsx)("path",{d:e.d,fill:"none",stroke:e.dash?s:i,strokeWidth:1.4,strokeDasharray:e.dash?"5 5":void 0,opacity:e.dash?1:.75},p)),[[160,150],[150,96],[168,200]].map(([e,p],t)=>(0,a.jsx)("circle",{cx:e,cy:p,r:4,fill:"none",stroke:1===t?s:i,strokeWidth:1.4},t))]});case"Artificial Intelligence":{let e=[[126,178,230],[110,158,206,254],[166,212]].map((e,a)=>e.map(e=>[58+42*a,e+r(a,10)]));return(0,a.jsxs)(a.Fragment,{children:[e.slice(0,-1).map((i,r)=>i.flatMap((i,o)=>e[r+1].map((e,n)=>{let l=(p+3*r+2*o+n)%5==0;return(0,a.jsx)("line",{x1:i[0],y1:i[1],x2:e[0],y2:e[1],stroke:l?t:s,strokeWidth:l?1.4:1,strokeDasharray:l?"4 4":void 0},`${r}-${o}-${n}`)}))),e.flatMap((p,s)=>p.map(([p,t],r)=>(0,a.jsx)("circle",{cx:p,cy:t,r:s===e.length-1?5.5:4.5,fill:s===e.length-1?i:"none",stroke:i,strokeWidth:1.4,opacity:1===s?.6:1},`n-${s}-${r}`)))]})}case"Guidelines":return(0,a.jsx)(a.Fragment,{children:Array.from({length:6}).map((e,o)=>{let n=110+26*o,l=(p+o)%3!=0;return(0,a.jsxs)("g",{children:[(0,a.jsx)("rect",{x:38,y:n-7,width:14,height:14,rx:3,fill:"none",stroke:l?i:t,strokeWidth:1.4}),l&&(0,a.jsx)("path",{d:`M41.5 ${n} l3 3.5 l6.5 -7.5`,fill:"none",stroke:i,strokeWidth:1.6,strokeLinecap:"round",strokeLinejoin:"round"}),(0,a.jsx)("rect",{x:62,y:n-3,width:70+r(o,26),height:3,rx:1.5,fill:l?s:t})]},o)})});case"Cultural heritage":return(0,a.jsxs)(a.Fragment,{children:[[0,1,2].map(e=>(0,a.jsx)("path",{d:`M${60+14*e} 250 L${60+14*e} ${170-16*e} A${40-14*e} ${40-14*e} 0 0 1 ${140-14*e} ${170-16*e} L${140-14*e} 250`,fill:"none",stroke:0===e?i:s,strokeWidth:1.4,opacity:1-.18*e},e)),(0,a.jsx)("circle",{cx:100,cy:198,r:11,fill:"none",stroke:i,strokeWidth:1.4}),(0,a.jsx)("path",{d:"M95.5 198 l3 3.4 l6 -7",fill:"none",stroke:i,strokeWidth:1.5,strokeLinecap:"round",strokeLinejoin:"round"})]});case"Traceability":return(0,a.jsx)(a.Fragment,{children:Array.from({length:5}).map((e,r)=>{let o=r===2+p%2,n=110+30*r;return(0,a.jsxs)("g",{children:[(0,a.jsx)("rect",{x:62,y:n,width:76,height:20,rx:10,fill:"none",stroke:o?t:i,strokeWidth:1.4,strokeDasharray:o?"4 4":void 0,opacity:o?1:.8}),r<4&&(0,a.jsx)("line",{x1:100,y1:n+20,x2:100,y2:n+30,stroke:o?t:s,strokeWidth:1.4})]},r)})});case"Diplomacy":{let e=[0,1,2,3,4].map(e=>{let a=96+r(e,30),p=104-r(e+3,30);return{y:108+28*e,left:a,right:p,agreed:a>p}});return(0,a.jsxs)(a.Fragment,{children:[(0,a.jsx)("line",{x1:100,y1:92,x2:100,y2:250,stroke:t,strokeWidth:1.4}),e.map((e,p)=>(0,a.jsxs)("g",{children:[(0,a.jsx)("line",{x1:34,y1:e.y,x2:e.left,y2:e.y,stroke:e.agreed?i:s,strokeWidth:e.agreed?2.4:1.6,strokeLinecap:"round"}),(0,a.jsx)("line",{x1:e.right,y1:e.y,x2:166,y2:e.y,stroke:e.agreed?i:s,strokeWidth:e.agreed?2.4:1.6,strokeLinecap:"round"}),e.agreed&&(0,a.jsx)("circle",{cx:(e.left+e.right)/2,cy:e.y,r:3.6,fill:i})]},p))]})}case"Sustainability":return(0,a.jsxs)(a.Fragment,{children:[(0,a.jsx)("path",{d:"M34 250 L34 110 M34 250 L170 250",fill:"none",stroke:t,strokeWidth:1.4}),(0,a.jsx)("path",{d:`M34 ${232+r(1,6)} C 74 ${232+r(1,6)}, 92 130, 170 ${118+r(2,8)}`,fill:"none",stroke:s,strokeWidth:1.4,strokeDasharray:"4 4"}),(0,a.jsx)("path",{d:`M34 ${232+r(1,6)} C 74 ${232+r(1,6)}, 96 168, 170 ${186+r(3,8)}`,fill:"none",stroke:i,strokeWidth:1.6}),[[34,232],[170,186]].map(([e,p],s)=>(0,a.jsx)("circle",{cx:e,cy:p+r(s+1,6),r:4,fill:i},s))]});case"Disaster response":return(0,a.jsxs)(a.Fragment,{children:[[54,40,26].map((e,p)=>(0,a.jsx)("circle",{cx:100,cy:180,r:e,fill:"none",stroke:0===p?s:t,strokeWidth:1.4},p)),(0,a.jsx)("line",{x1:100,y1:180,x2:100+54*Math.cos(p%6-2.4),y2:180+54*Math.sin(p%6-2.4),stroke:i,strokeWidth:1.6}),(0,a.jsx)("circle",{cx:100,cy:180,r:4.5,fill:i})]});default:return null}})()})]})}function g(e){let a=e.indexOf(":");return -1===a||a>48?[e,null]:[e.slice(0,a),e.slice(a+1).trim()]}function w(e){return!!e.url&&"internal"!==e.access}function h({pub:e,onOpen:p}){let[i,s]=g(e.title),t=Number(e.index),o=!!l(e.url);return(0,a.jsx)(r.Tilt3D,{max:5,className:"h-full",children:(0,a.jsxs)("button",{type:"button",onClick:p,"aria-label":w(e)?`Open ${e.title}`:`${e.title}, ${e.status.toLowerCase()}`,className:"group/cover relative flex aspect-[2/3] h-full w-full flex-col justify-end overflow-hidden rounded-xl border border-border p-4 text-left transition-colors hover:border-accent focus-visible:border-accent focus-visible:outline-none",style:{background:u(e.topic,t)},children:[(0,a.jsx)(d,{topic:e.topic,seed:t}),(0,a.jsx)("span",{"aria-hidden":!0,className:"absolute inset-0",style:{background:"linear-gradient(180deg, transparent 34%, color-mix(in oklab, #0c0812 68%, transparent) 58%, #0c0812 92%)"}}),o&&(0,a.jsxs)("span",{className:"absolute left-3 top-3 inline-flex items-center gap-1 rounded-full px-2 py-0.5 font-mono text-[9px] font-semibold uppercase tracking-wider shadow-sm",style:{background:"var(--poster-accent)",color:"var(--poster-accent-ink)"},children:[(0,a.jsx)("span",{"aria-hidden":!0,children:"📖"})," Book"]}),(0,a.jsx)("span",{className:"absolute right-3 top-3",children:"internal"===e.access?(0,a.jsx)("span",{className:"rounded-full border border-white/35 bg-black/45 px-2 py-0.5 font-mono text-[9px] uppercase tracking-wider text-white/85 backdrop-blur",children:"CoLab only"}):e.url?(0,a.jsx)("span",{className:"rounded-full px-2 py-0.5 font-mono text-[9px] font-semibold uppercase tracking-wider",style:{background:"var(--poster-accent)",color:"var(--poster-accent-ink)"},children:"Published"}):(0,a.jsx)("span",{className:"rounded-full border border-border px-2 py-0.5 font-mono text-[9px] uppercase tracking-wider",style:{color:"var(--poster-ink-muted)"},children:"In prep"})}),w(e)&&(0,a.jsx)("span",{className:"absolute inset-0 flex items-center justify-center opacity-0 transition-opacity group-hover/cover:opacity-100 group-focus-visible/cover:opacity-100",children:(0,a.jsx)("span",{className:"inline-flex items-center gap-2 rounded-full px-4 py-2 text-sm font-semibold shadow-lg",style:{background:"var(--poster-accent)",color:"var(--poster-accent-ink)"},children:"Read"})}),(0,a.jsxs)("div",{className:"relative",children:[(0,a.jsx)("p",{className:"font-mono text-[10px] uppercase tracking-wider",style:{color:"var(--poster-accent)"},children:e.topic}),(0,a.jsx)("h3",{className:"mt-1.5 line-clamp-3 font-heading text-xl uppercase leading-[1.05] tracking-[0.02em] text-[color:var(--poster-ink)] transition-colors group-hover/cover:text-[color:var(--poster-accent)]",children:i}),s&&(0,a.jsx)("p",{className:"mt-1.5 line-clamp-2 text-[11px] leading-snug",style:{color:"var(--poster-ink-muted)"},children:s}),(0,a.jsx)("div",{className:"mt-3 font-mono text-[10px] uppercase tracking-wider",style:{color:"var(--poster-ink-muted)"},children:e.date??e.status})]})]})})}function m({pub:e,onClose:s}){(0,p.useEffect)(()=>{let e=e=>"Escape"===e.key&&s();return document.addEventListener("keydown",e),document.body.style.overflow="hidden",()=>{document.removeEventListener("keydown",e),document.body.style.overflow=""}},[s]);let[t,r]=g(e.title),n=e.url.startsWith("/"),b=l(e.url),w="flex w-full items-center justify-between gap-4 rounded-xl border border-border bg-card px-5 py-4 text-left transition-colors hover:border-accent";return(0,a.jsxs)("div",{className:"fixed inset-0 z-[100] flex flex-col bg-background",role:"dialog","aria-modal":"true","aria-label":e.title,children:[(0,a.jsxs)("div",{className:"flex items-center justify-between gap-3 border-b border-border px-4 py-2.5",children:[(0,a.jsxs)("div",{className:"min-w-0",children:[(0,a.jsx)("h3",{className:"truncate font-heading text-xl uppercase tracking-[0.02em] sm:text-2xl",children:t}),(0,a.jsxs)("p",{className:"truncate font-mono text-[10px] text-muted",children:[e.topic,e.date&&` \xb7 ${e.date}`]})]}),(0,a.jsx)("button",{type:"button",onClick:s,className:"shrink-0 rounded-full bg-accent px-4 py-1.5 text-sm font-semibold text-accent-ink transition-transform hover:scale-[1.03]",children:"✕ Close"})]}),(0,a.jsx)("div",{className:"flex-1 overflow-y-auto",children:(0,a.jsxs)("div",{className:"mx-auto grid max-w-4xl gap-10 px-6 py-12 sm:grid-cols-[minmax(0,200px)_1fr] sm:py-14",children:[(0,a.jsx)("div",{className:"relative hidden aspect-[2/3] w-full overflow-hidden rounded-xl border border-border sm:block",style:{background:u(e.topic,Number(e.index))},children:(0,a.jsx)(d,{topic:e.topic,seed:Number(e.index)})}),(0,a.jsxs)("div",{children:[r&&(0,a.jsx)("p",{className:"font-heading text-lg uppercase leading-snug tracking-wide text-muted",children:r}),(0,a.jsx)("p",{className:"mt-4 border-l-2 border-accent pl-5 text-lg leading-relaxed text-foreground/90",children:e.question}),(0,a.jsx)("p",{className:"mt-6 leading-relaxed text-muted",children:e.summary}),(0,a.jsx)("div",{className:"mt-8 space-y-3",children:"internal"===e.access?(0,a.jsxs)("a",{href:e.url,target:"_blank",rel:"noopener noreferrer",className:w,children:[(0,a.jsxs)("span",{className:"font-semibold text-foreground",children:["Open in the CoLab repo"," ",(0,a.jsx)("span",{className:"font-normal text-muted",children:"(sign-in required)"})]}),(0,a.jsx)("span",{"aria-hidden":!0,className:"text-accent",children:"↗"})]}):e.url?(0,a.jsxs)(a.Fragment,{children:[n?(0,a.jsxs)(i.Link,{href:e.url,className:w,children:[(0,a.jsx)("span",{className:"font-semibold text-foreground",children:"Read the report"}),(0,a.jsx)("span",{"aria-hidden":!0,className:"text-accent",children:"→"})]}):(0,a.jsxs)("a",{href:e.url,target:"_blank",rel:"noopener noreferrer",className:w,children:[(0,a.jsx)("span",{className:"font-semibold text-foreground",children:"Read the report"}),(0,a.jsx)("span",{"aria-hidden":!0,className:"text-accent",children:"↗"})]}),b&&(0,a.jsxs)(o.ReportBook,{title:t,pages:b.pages,aspect:b.aspect,pdfUrl:e.pdf??(0,c.asset)(b.pdf),className:w,children:[(0,a.jsxs)("span",{className:"font-semibold text-foreground",children:["Read as book"," ",(0,a.jsx)("span",{className:"font-normal text-muted",children:"(flip the original pages)"})]}),(0,a.jsx)("span",{"aria-hidden":!0,className:"text-accent",children:"📖"})]}),e.pdf&&(0,a.jsxs)("a",{href:e.pdf,target:"_blank",rel:"noopener noreferrer",className:w,children:[(0,a.jsx)("span",{className:"font-semibold text-foreground",children:"Download the PDF"}),(0,a.jsx)("span",{"aria-hidden":!0,className:"text-accent",children:"↗"})]})]}):(0,a.jsxs)("p",{className:"rounded-xl border border-dashed border-border px-5 py-4 text-sm text-muted",children:["This report is ",e.status.toLowerCase(),". The research it writes up is already running, so the demos and the source are the way in until it publishes."]})}),(0,a.jsxs)("div",{className:"mt-8 flex flex-wrap gap-x-4 gap-y-2 border-t border-border pt-6 font-mono text-xs text-muted",children:[(0,a.jsx)("span",{children:e.area}),(0,a.jsx)("span",{children:e.status}),e.date&&(0,a.jsx)("span",{children:e.date}),e.repo&&(0,a.jsx)("a",{href:e.repo,target:"_blank",rel:"noopener noreferrer",className:"ml-auto text-foreground/80 transition-colors hover:text-accent",children:"View source ↗"})]})]})]})})]})}e.s(["PublicationsShowcase",0,function(){let[e,i]=(0,p.useState)(null),[r,o]=(0,p.useState)(null),[n,l]=(0,p.useState)(null),c=s.publications.items.filter(a=>(!e||a.topic===e)&&("published"!==r||!!w(a))),b=c.filter(w).length,u=c.filter(e=>"internal"===e.access).length,d=s.publicationTopics.map(e=>({topic:e,items:c.filter(a=>a.topic===e)})).filter(e=>e.items.length>0),g=e=>`rounded-full border px-3 py-1.5 text-sm transition-colors ${e?"border-accent bg-accent font-semibold text-accent-ink":"border-border text-muted hover:border-accent hover:text-accent"}`;return(0,a.jsxs)("div",{className:"mx-auto max-w-6xl px-6 pb-24",children:[(0,a.jsxs)("div",{className:"space-y-4 border-b border-border pb-8",children:[(0,a.jsxs)("div",{className:"flex flex-wrap items-center gap-2",children:[(0,a.jsx)("span",{className:"mr-2 w-20 shrink-0 text-xs uppercase tracking-wider text-muted",children:"Topic"}),(0,a.jsx)("button",{type:"button",onClick:()=>i(null),className:g(null===e),children:"All"}),s.publicationTopics.filter(e=>s.publications.items.some(a=>a.topic===e)).map(p=>(0,a.jsx)("button",{type:"button",onClick:()=>i(e=>e===p?null:p),className:g(e===p),children:p},p)),(0,a.jsxs)("span",{className:"ml-auto font-mono text-xs text-muted",children:[b," published",u>0&&` \xb7 ${u} CoLab only`," · ",c.length," ","shown"]})]}),(0,a.jsxs)("div",{className:"flex flex-wrap items-center gap-2",children:[(0,a.jsx)("span",{className:"mr-2 w-20 shrink-0 text-xs uppercase tracking-wider text-muted",children:"Status"}),(0,a.jsx)("button",{type:"button",onClick:()=>o(null),className:g(null===r),children:"All"}),(0,a.jsx)("button",{type:"button",onClick:()=>o(e=>"published"===e?null:"published"),className:g("published"===r),children:"Readable now"})]})]}),(0,a.jsx)("div",{className:"mt-12 space-y-14",children:d.map(e=>(0,a.jsx)(t.PosterRail,{title:e.topic,count:e.items.length,countNoun:"report",ariaLabel:`${e.topic} reports`,children:e.items.map(e=>(0,a.jsx)("div",{className:"w-[190px] shrink-0 snap-start scroll-mt-24 sm:w-[215px]",children:(0,a.jsx)(h,{pub:e,onOpen:()=>l(e)})},e.index))},e.topic))}),0===c.length&&(0,a.jsx)("p",{className:"py-16 text-muted",children:"No reports match those filters."}),n&&(0,a.jsx)(m,{pub:n,onClose:()=>l(null)},n.index)]})}],46973)},43413,e=>{e.v(a=>Promise.all(["static/chunks/1qa2_w3ql48wi.js"].map(a=>e.l(a))).then(()=>a(62914)))}]); \ No newline at end of file diff --git a/static-site/_next/static/chunks/0jub-lza-_qsd.js b/static-site/_next/static/chunks/0jub-lza-_qsd.js new file mode 100644 index 000000000..56a2e3558 --- /dev/null +++ b/static-site/_next/static/chunks/0jub-lza-_qsd.js @@ -0,0 +1 @@ +(globalThis.TURBOPACK||(globalThis.TURBOPACK=[])).push(["object"==typeof document?document.currentScript:void 0,72328,e=>{"use strict";var a=e.i(71164),p=e.i(38544),i=e.i(71645);e.s(["useReducedMotion",0,function(){a.hasReducedMotionListener.current||(0,p.initPrefersReducedMotion)();let[e]=(0,i.useState)(a.prefersReducedMotion.current);return e}])},82987,e=>{"use strict";var a=e.i(43476),p=e.i(32181),i=e.i(72328);let s=[.16,1,.3,1],t={hidden:{},show:{transition:{staggerChildren:.07}}};e.s(["Reveal",0,function({children:e,className:t,delay:o=0,y:r=24,as:n="div"}){let l=(0,i.useReducedMotion)(),c=p.motion[n];return(0,a.jsx)(c,{className:t,initial:{opacity:0,y:l?0:r},whileInView:{opacity:1,y:0},viewport:{once:!0,margin:"-80px"},transition:{duration:.65,ease:s,delay:o},children:e})},"Stagger",0,function({children:e,className:i}){return(0,a.jsx)(p.motion.div,{className:i,variants:t,initial:"hidden",whileInView:"show",viewport:{once:!0,margin:"-60px"},children:e})},"StaggerItem",0,function({children:e,className:t}){let o=(0,i.useReducedMotion)();return(0,a.jsx)(p.motion.div,{className:t,variants:{hidden:{opacity:0,y:20*!o},show:{opacity:1,y:0,transition:{duration:.6,ease:s}}},children:e})}])},18620,e=>{"use strict";var a=e.i(43476),p=e.i(23475),i=e.i(18566);let s=[{label:"Overview",href:"/portfolio"},{label:"Live Demos",href:"/demos"},{label:"Publications",href:"/publications"},{label:"Media",href:"/media"},{label:"Newsletter",href:"/newsletter"}];e.s(["SectionTabs",0,function(){let e=(0,i.usePathname)();return(0,a.jsx)("div",{className:"border-b border-border bg-background/80 backdrop-blur",children:(0,a.jsx)("nav",{"aria-label":"Portfolio sections",className:"mx-auto flex max-w-6xl gap-6 px-6",children:s.map(i=>{let s="/portfolio"===i.href?"/portfolio"===e:e.startsWith(i.href);return(0,a.jsx)(p.Link,{href:i.href,"aria-current":s?"page":void 0,className:`-mb-px border-b-2 py-4 text-sm font-medium uppercase tracking-wider transition-colors ${s?"border-accent text-accent":"border-transparent text-muted hover:text-foreground"}`,children:i.label},i.href)})})})}])},34010,e=>{"use strict";var a=e.i(43476),p=e.i(71645);let i={startPage:0,singlePageBreakpoint:640,chromeHeight:150,chromeWidth:48,maxPageHeight:1100,flippingTime:700,drawShadow:!0,maxShadowOpacity:.4,keyboard:!0};async function s(a,p){let s={...i,...p};if(!s.pages.length)throw Error("mountFlipbook: `pages` is empty");let t=(await e.A(43413)).PageFlip,o=null,r=Math.min(Math.max(s.startPage,0),s.pages.length-1),n=!1,l=null,c=()=>{if(n)return;let e=s.singlePageBreakpoint>0&&window.matchMedia(`(max-width: ${s.singlePageBreakpoint}px)`).matches,{width:p,height:l}=function(e,a){let p=a.chromeHeight??i.chromeHeight,s=a.chromeWidth??i.chromeWidth,t=Math.min(a.viewportHeight-p,a.maxPageHeight??i.maxPageHeight),o=a.viewportWidth-s,r=Math.max(t,120),n=r*e,l=a.single?n:2*n;if(l>o){let e=o/l;n*=e,r*=e}return{width:Math.round(n),height:Math.round(r)}}(s.aspect,{single:e,viewportWidth:window.innerWidth,viewportHeight:window.innerHeight,chromeWidth:s.chromeWidth,chromeHeight:s.chromeHeight,maxPageHeight:s.maxPageHeight});if((o=new t(a,{width:p,height:l,size:"fixed",showCover:!0,usePortrait:e,maxShadowOpacity:s.maxShadowOpacity,mobileScrollSupport:!0,flippingTime:s.flippingTime,drawShadow:s.drawShadow})).loadFromImages(s.pages.slice()),o.on("flip",e=>{r=e.data,s.onFlip?.(r)}),r>0)try{o.turnToPage(r)}catch{}s.onReady?.()},b=()=>{if(o){try{o.destroy()}catch{}o=null,a.innerHTML=""}},u=()=>{l&&clearTimeout(l),l=setTimeout(()=>{b(),c()},200)},d=e=>{"ArrowLeft"===e.key?o?.flipPrev():"ArrowRight"===e.key&&o?.flipNext()};return c(),window.addEventListener("resize",u),s.keyboard&&document.addEventListener("keydown",d),{next:()=>o?.flipNext(),prev:()=>o?.flipPrev(),goTo:e=>o?.turnToPage(e),currentPage:()=>r,pageCount:()=>s.pages.length,destroy:()=>{n||(n=!0,l&&clearTimeout(l),window.removeEventListener("resize",u),document.removeEventListener("keydown",d),b())}}}function t(e,a,p){let i=document.createElement(e);return a&&(i.className=a),null!=p&&(i.textContent=p),i}async function o(e){let a=e.container??document.body,p=e.pages.length,i=t("div",`rab-overlay${e.className?` ${e.className}`:""}`);i.setAttribute("role","dialog"),i.setAttribute("aria-modal","true"),i.setAttribute("aria-label",`${e.title??"Document"} — page view`);let o=t("div","rab-chrome");o.append(t("span","rab-title",e.title??""));let r=t("div","rab-actions"),n=t("span","rab-counter");if(r.append(n),e.pdfUrl){let a=t("a","rab-btn rab-download","Download PDF ↗");a.href=e.pdfUrl,a.target="_blank",a.rel="noopener noreferrer",r.append(a)}let l=t("button","rab-btn","Close ✕");l.type="button",l.setAttribute("aria-label","Close book view"),r.append(l),o.append(r);let c=t("div","rab-stage"),b=t("button","rab-arrow rab-arrow-prev","‹");b.type="button",b.setAttribute("aria-label","Previous page");let u=t("button","rab-arrow rab-arrow-next","›");u.type="button",u.setAttribute("aria-label","Next page");let d=t("div","rab-book"),g=t("p","rab-loading","Opening the book…"),w=t("div","rab-book-wrap");w.append(d,g),c.append(b,w,u);let h=e.hint??"Use the arrows or ← → keys to turn pages · Esc to close",m=t("p","rab-hint",h);i.append(o,c),h&&i.append(m),a.append(i);let f=e=>{n.textContent=p?`${Math.min(e+1,p)} / ${p}`:""};f(e.startPage??0);let x=document.body.style.overflow;document.body.style.overflow="hidden";let v=null,y=!1,k=()=>{y||(y=!0,document.removeEventListener("keydown",j),document.body.style.overflow=x,v?.destroy(),i.remove(),e.onClose?.())},j=e=>{"Escape"===e.key&&k()};document.addEventListener("keydown",j),l.addEventListener("click",k);try{v=await s(d,{pages:e.pages,aspect:e.aspect,startPage:e.startPage,onFlip:a=>{f(a),e.onFlip?.(a)},onReady:()=>g.remove()})}catch(e){throw k(),e}return y?(v.destroy(),{close:k,next:()=>{},prev:()=>{}}):(b.addEventListener("click",()=>v?.prev()),u.addEventListener("click",()=>v?.next()),{close:k,next:()=>v?.next(),prev:()=>v?.prev()})}function r({pages:e,aspect:i,title:s,pdfUrl:t,className:n,children:l,overlayClassName:c,hint:b,open:u,onOpenChange:d,onFlip:g}){let[w,h]=(0,p.useState)(!1),m=u??w,f=(0,p.useRef)(null),x=(0,p.useRef)(g);x.current=g;let v=(0,p.useCallback)(e=>{void 0===u&&h(e),d?.(e)},[u,d]);return(0,p.useEffect)(()=>{if(!m)return;let a=!1;return o({pages:e,aspect:i,title:s,pdfUrl:t,className:c,hint:b,onFlip:e=>x.current?.(e),onClose:()=>{a||v(!1)}}).then(e=>{a?e.close():f.current=e}).catch(e=>{console.error("read-as-book: failed to open the viewer",e),a||v(!1)}),()=>{a=!0,f.current?.close(),f.current=null}},[m,e,i,s,t,c,b]),(0,a.jsx)("button",{type:"button",onClick:()=>v(!0),className:n??"rab-trigger",children:l??(0,a.jsxs)(a.Fragment,{children:["Read as book ",(0,a.jsx)("span",{"aria-hidden":!0,children:"📖"})]})})}var n=e.i(93209);e.s(["ReportBook",0,function({pages:e,aspect:i,title:s,pdfUrl:t,className:o,children:l}){let c=(0,p.useMemo)(()=>e.map(e=>(0,n.asset)(`/${e}`)),[e]);return(0,a.jsx)(r,{pages:c,aspect:i,title:s,pdfUrl:t,className:o??"btn-sweep inline-flex items-center gap-2 rounded-full border border-border px-5 py-2.5 text-sm font-semibold text-foreground transition-colors hover:border-border-strong",children:l})}],34010)},12243,e=>{"use strict";var a=e.i(43476),p=e.i(32181),i=e.i(87652),s=e.i(91994),t=e.i(72328),o=e.i(71645);e.s(["Tilt3D",0,function({children:e,className:r="",max:n=9,glare:l=!0,disabled:c=!1}){let b=(0,t.useReducedMotion)(),u=(0,o.useRef)(null),d=(0,i.useMotionValue)(0),g=(0,i.useMotionValue)(0),w=(0,s.useSpring)(d,{stiffness:170,damping:16,mass:.5}),h=(0,s.useSpring)(g,{stiffness:170,damping:16,mass:.5});return c?(0,a.jsx)("div",{className:`h-full ${r}`,children:e}):(0,a.jsx)("div",{className:"tilt-scene h-full",children:(0,a.jsx)(p.motion.div,{ref:u,onPointerMove:function(e){let a=u.current;if(b||!a)return;let p=a.getBoundingClientRect(),i=(e.clientX-p.left)/p.width,s=(e.clientY-p.top)/p.height;g.set((i-.5)*2*n),d.set((.5-s)*2*n),a.style.setProperty("--mx",`${100*i}%`),a.style.setProperty("--my",`${100*s}%`),l&&(a.style.setProperty("--gx",`${100*i}%`),a.style.setProperty("--gy",`${100*s}%`),a.style.setProperty("--gop","1"))},onPointerLeave:function(){d.set(0),g.set(0),u.current?.style.setProperty("--gop","0")},style:{rotateX:w,rotateY:h,transformStyle:"preserve-3d"},className:`tilt-card h-full ${l?"tilt-glare":""} ${r}`,children:e})})}])},88302,e=>{"use strict";var a=e.i(43476),p=e.i(71645);e.s(["PosterRail",0,function({title:e,count:i,countNoun:s,ariaLabel:t,children:o}){let r=(0,p.useRef)(null),[n,l]=(0,p.useState)(!0),[c,b]=(0,p.useState)(!0),u=(0,p.useCallback)(()=>{let e=r.current;e&&(l(e.scrollLeft<=1),b(e.scrollLeft>=e.scrollWidth-e.clientWidth-1))},[]);(0,p.useEffect)(()=>{u();let e=r.current;if(!e)return;let a=new ResizeObserver(u);return a.observe(e),()=>a.disconnect()},[u,o]);let d=e=>{let a=r.current;a&&a.scrollBy({left:e*a.clientWidth*.85,behavior:"smooth"})},g=!(n&&c),w=e=>`pointer-events-auto grid h-9 w-9 place-items-center rounded-full border border-border bg-background/90 text-foreground backdrop-blur transition-all ${e?"opacity-100 hover:border-accent hover:text-accent":"cursor-default opacity-0"}`;return(0,a.jsxs)("section",{children:[(0,a.jsxs)("div",{className:"flex items-baseline justify-between gap-4 border-b border-border pb-3",children:[(0,a.jsx)("h2",{className:"font-heading text-2xl uppercase tracking-wide sm:text-3xl",children:e}),(0,a.jsxs)("span",{className:"shrink-0 font-mono text-xs text-muted",children:[i," ",1===i?s:`${s}s`]})]}),(0,a.jsxs)("div",{className:"relative mt-6",children:[(0,a.jsx)("div",{ref:r,onScroll:u,tabIndex:0,role:"group","aria-label":t,className:"no-scrollbar -mx-6 flex snap-x snap-mandatory scroll-pl-6 gap-5 overflow-x-auto px-6 pb-2 focus-visible:outline-none",children:o}),g&&(0,a.jsxs)("div",{className:"pointer-events-none absolute inset-y-0 left-0 right-0 hidden items-center justify-between sm:flex",children:[(0,a.jsx)("button",{type:"button",onClick:()=>d(-1),disabled:n,"aria-label":`Scroll ${e} left`,className:`-ml-4 ${w(!n)}`,children:(0,a.jsx)("span",{"aria-hidden":!0,children:"←"})}),(0,a.jsx)("button",{type:"button",onClick:()=>d(1),disabled:c,"aria-label":`Scroll ${e} right`,className:`-mr-4 ${w(!c)}`,children:(0,a.jsx)("span",{"aria-hidden":!0,children:"→"})})]})]})]})}])},46973,e=>{"use strict";var a=e.i(43476),p=e.i(71645),i=e.i(23475),s=e.i(89042),t=e.i(88302),o=e.i(12243),r=e.i(34010);let n={"after-the-corridor":{pages:["publications/after-the-corridor/pages/p01.webp","publications/after-the-corridor/pages/p02.webp","publications/after-the-corridor/pages/p03.webp","publications/after-the-corridor/pages/p04.webp","publications/after-the-corridor/pages/p05.webp","publications/after-the-corridor/pages/p06.webp","publications/after-the-corridor/pages/p07.webp","publications/after-the-corridor/pages/p08.webp","publications/after-the-corridor/pages/p09.webp","publications/after-the-corridor/pages/p10.webp","publications/after-the-corridor/pages/p11.webp","publications/after-the-corridor/pages/p12.webp","publications/after-the-corridor/pages/p13.webp","publications/after-the-corridor/pages/p14.webp","publications/after-the-corridor/pages/p15.webp","publications/after-the-corridor/pages/p16.webp","publications/after-the-corridor/pages/p17.webp","publications/after-the-corridor/pages/p18.webp","publications/after-the-corridor/pages/p19.webp","publications/after-the-corridor/pages/p20.webp","publications/after-the-corridor/pages/p21.webp","publications/after-the-corridor/pages/p22.webp"],aspect:.7727,pdf:"/publications/after-the-corridor/report.pdf"},"agentic-language-development":{pages:["publications/agentic-language-development/pages/p01.webp","publications/agentic-language-development/pages/p02.webp","publications/agentic-language-development/pages/p03.webp","publications/agentic-language-development/pages/p04.webp","publications/agentic-language-development/pages/p05.webp","publications/agentic-language-development/pages/p06.webp","publications/agentic-language-development/pages/p07.webp","publications/agentic-language-development/pages/p08.webp","publications/agentic-language-development/pages/p09.webp","publications/agentic-language-development/pages/p10.webp","publications/agentic-language-development/pages/p11.webp","publications/agentic-language-development/pages/p12.webp","publications/agentic-language-development/pages/p13.webp","publications/agentic-language-development/pages/p14.webp","publications/agentic-language-development/pages/p15.webp","publications/agentic-language-development/pages/p16.webp","publications/agentic-language-development/pages/p17.webp","publications/agentic-language-development/pages/p18.webp","publications/agentic-language-development/pages/p19.webp","publications/agentic-language-development/pages/p20.webp","publications/agentic-language-development/pages/p21.webp","publications/agentic-language-development/pages/p22.webp","publications/agentic-language-development/pages/p23.webp","publications/agentic-language-development/pages/p24.webp","publications/agentic-language-development/pages/p25.webp"],aspect:.7067,pdf:"/publications/agentic-language-development/report.pdf"},"ai-carbon-footprint":{pages:["publications/ai-carbon-footprint/pages/p01.webp","publications/ai-carbon-footprint/pages/p02.webp","publications/ai-carbon-footprint/pages/p03.webp","publications/ai-carbon-footprint/pages/p04.webp","publications/ai-carbon-footprint/pages/p05.webp","publications/ai-carbon-footprint/pages/p06.webp","publications/ai-carbon-footprint/pages/p07.webp","publications/ai-carbon-footprint/pages/p08.webp","publications/ai-carbon-footprint/pages/p09.webp","publications/ai-carbon-footprint/pages/p10.webp","publications/ai-carbon-footprint/pages/p11.webp","publications/ai-carbon-footprint/pages/p12.webp","publications/ai-carbon-footprint/pages/p13.webp","publications/ai-carbon-footprint/pages/p14.webp","publications/ai-carbon-footprint/pages/p15.webp","publications/ai-carbon-footprint/pages/p16.webp","publications/ai-carbon-footprint/pages/p17.webp","publications/ai-carbon-footprint/pages/p18.webp"],aspect:.7067,pdf:"/publications/ai-carbon-footprint/report.pdf"},"ai-models-research":{pages:["publications/ai-models-research/pages/p01.webp","publications/ai-models-research/pages/p02.webp","publications/ai-models-research/pages/p03.webp","publications/ai-models-research/pages/p04.webp","publications/ai-models-research/pages/p05.webp","publications/ai-models-research/pages/p06.webp","publications/ai-models-research/pages/p07.webp","publications/ai-models-research/pages/p08.webp","publications/ai-models-research/pages/p09.webp","publications/ai-models-research/pages/p10.webp","publications/ai-models-research/pages/p11.webp","publications/ai-models-research/pages/p12.webp","publications/ai-models-research/pages/p13.webp","publications/ai-models-research/pages/p14.webp","publications/ai-models-research/pages/p15.webp","publications/ai-models-research/pages/p16.webp","publications/ai-models-research/pages/p17.webp","publications/ai-models-research/pages/p18.webp","publications/ai-models-research/pages/p19.webp"],aspect:.7067,pdf:"/publications/ai-models-research/report.pdf"},"ai-research-assistant":{pages:["publications/ai-research-assistant/pages/p01.webp","publications/ai-research-assistant/pages/p02.webp","publications/ai-research-assistant/pages/p03.webp","publications/ai-research-assistant/pages/p04.webp","publications/ai-research-assistant/pages/p05.webp","publications/ai-research-assistant/pages/p06.webp","publications/ai-research-assistant/pages/p07.webp","publications/ai-research-assistant/pages/p08.webp","publications/ai-research-assistant/pages/p09.webp","publications/ai-research-assistant/pages/p10.webp"],aspect:.7067,pdf:"/publications/ai-research-assistant/report.pdf"},cerai:{pages:["publications/cerai/pages/p01.webp","publications/cerai/pages/p02.webp","publications/cerai/pages/p03.webp","publications/cerai/pages/p04.webp","publications/cerai/pages/p05.webp","publications/cerai/pages/p06.webp","publications/cerai/pages/p07.webp","publications/cerai/pages/p08.webp","publications/cerai/pages/p09.webp","publications/cerai/pages/p10.webp","publications/cerai/pages/p11.webp","publications/cerai/pages/p12.webp","publications/cerai/pages/p13.webp"],aspect:.7067,pdf:"/publications/cerai/report.pdf"},"digital-provenance-passport":{pages:["publications/digital-provenance-passport/pages/p01.webp","publications/digital-provenance-passport/pages/p02.webp","publications/digital-provenance-passport/pages/p03.webp","publications/digital-provenance-passport/pages/p04.webp","publications/digital-provenance-passport/pages/p05.webp","publications/digital-provenance-passport/pages/p06.webp","publications/digital-provenance-passport/pages/p07.webp","publications/digital-provenance-passport/pages/p08.webp","publications/digital-provenance-passport/pages/p09.webp","publications/digital-provenance-passport/pages/p10.webp","publications/digital-provenance-passport/pages/p11.webp","publications/digital-provenance-passport/pages/p12.webp","publications/digital-provenance-passport/pages/p13.webp","publications/digital-provenance-passport/pages/p14.webp","publications/digital-provenance-passport/pages/p15.webp","publications/digital-provenance-passport/pages/p16.webp","publications/digital-provenance-passport/pages/p17.webp","publications/digital-provenance-passport/pages/p18.webp","publications/digital-provenance-passport/pages/p19.webp","publications/digital-provenance-passport/pages/p20.webp","publications/digital-provenance-passport/pages/p21.webp","publications/digital-provenance-passport/pages/p22.webp","publications/digital-provenance-passport/pages/p23.webp","publications/digital-provenance-passport/pages/p24.webp","publications/digital-provenance-passport/pages/p25.webp","publications/digital-provenance-passport/pages/p26.webp","publications/digital-provenance-passport/pages/p27.webp","publications/digital-provenance-passport/pages/p28.webp"],aspect:.7067,pdf:"/publications/digital-provenance-passport/report.pdf"},"diplomatic-simulator":{pages:["publications/diplomatic-simulator/pages/p01.webp","publications/diplomatic-simulator/pages/p02.webp","publications/diplomatic-simulator/pages/p03.webp","publications/diplomatic-simulator/pages/p04.webp","publications/diplomatic-simulator/pages/p05.webp","publications/diplomatic-simulator/pages/p06.webp","publications/diplomatic-simulator/pages/p07.webp","publications/diplomatic-simulator/pages/p08.webp","publications/diplomatic-simulator/pages/p09.webp","publications/diplomatic-simulator/pages/p10.webp","publications/diplomatic-simulator/pages/p11.webp","publications/diplomatic-simulator/pages/p12.webp","publications/diplomatic-simulator/pages/p13.webp","publications/diplomatic-simulator/pages/p14.webp","publications/diplomatic-simulator/pages/p15.webp","publications/diplomatic-simulator/pages/p16.webp","publications/diplomatic-simulator/pages/p17.webp","publications/diplomatic-simulator/pages/p18.webp","publications/diplomatic-simulator/pages/p19.webp","publications/diplomatic-simulator/pages/p20.webp","publications/diplomatic-simulator/pages/p21.webp","publications/diplomatic-simulator/pages/p22.webp","publications/diplomatic-simulator/pages/p23.webp"],aspect:.7067,pdf:"/publications/diplomatic-simulator/report.pdf"},ercf:{pages:["publications/ercf/pages/p01.webp","publications/ercf/pages/p02.webp","publications/ercf/pages/p03.webp","publications/ercf/pages/p04.webp","publications/ercf/pages/p05.webp","publications/ercf/pages/p06.webp","publications/ercf/pages/p07.webp","publications/ercf/pages/p08.webp","publications/ercf/pages/p09.webp","publications/ercf/pages/p10.webp","publications/ercf/pages/p11.webp","publications/ercf/pages/p12.webp","publications/ercf/pages/p13.webp","publications/ercf/pages/p14.webp","publications/ercf/pages/p15.webp","publications/ercf/pages/p16.webp","publications/ercf/pages/p17.webp","publications/ercf/pages/p18.webp","publications/ercf/pages/p19.webp","publications/ercf/pages/p20.webp","publications/ercf/pages/p21.webp","publications/ercf/pages/p22.webp","publications/ercf/pages/p23.webp","publications/ercf/pages/p24.webp","publications/ercf/pages/p25.webp","publications/ercf/pages/p26.webp","publications/ercf/pages/p27.webp","publications/ercf/pages/p28.webp","publications/ercf/pages/p29.webp","publications/ercf/pages/p30.webp","publications/ercf/pages/p31.webp","publications/ercf/pages/p32.webp"],aspect:.7067,pdf:"/publications/ercf/report.pdf"},erus:{pages:["publications/erus/pages/p01.webp","publications/erus/pages/p02.webp","publications/erus/pages/p03.webp","publications/erus/pages/p04.webp","publications/erus/pages/p05.webp","publications/erus/pages/p06.webp","publications/erus/pages/p07.webp","publications/erus/pages/p08.webp","publications/erus/pages/p09.webp","publications/erus/pages/p10.webp","publications/erus/pages/p11.webp","publications/erus/pages/p12.webp","publications/erus/pages/p13.webp","publications/erus/pages/p14.webp","publications/erus/pages/p15.webp","publications/erus/pages/p16.webp","publications/erus/pages/p17.webp","publications/erus/pages/p18.webp","publications/erus/pages/p19.webp","publications/erus/pages/p20.webp","publications/erus/pages/p21.webp","publications/erus/pages/p22.webp","publications/erus/pages/p23.webp","publications/erus/pages/p24.webp"],aspect:.7067,pdf:"/publications/erus/report.pdf"},"evacuation-inform-index":{pages:["publications/evacuation-inform-index/pages/p01.webp","publications/evacuation-inform-index/pages/p02.webp","publications/evacuation-inform-index/pages/p03.webp","publications/evacuation-inform-index/pages/p04.webp","publications/evacuation-inform-index/pages/p05.webp","publications/evacuation-inform-index/pages/p06.webp","publications/evacuation-inform-index/pages/p07.webp","publications/evacuation-inform-index/pages/p08.webp","publications/evacuation-inform-index/pages/p09.webp","publications/evacuation-inform-index/pages/p10.webp","publications/evacuation-inform-index/pages/p11.webp"],aspect:.7067,pdf:"/publications/evacuation-inform-index/report.pdf"},"evacuation-simulation":{pages:["publications/evacuation-simulation/pages/p01.webp","publications/evacuation-simulation/pages/p02.webp","publications/evacuation-simulation/pages/p03.webp","publications/evacuation-simulation/pages/p04.webp","publications/evacuation-simulation/pages/p05.webp","publications/evacuation-simulation/pages/p06.webp","publications/evacuation-simulation/pages/p07.webp","publications/evacuation-simulation/pages/p08.webp","publications/evacuation-simulation/pages/p09.webp","publications/evacuation-simulation/pages/p10.webp","publications/evacuation-simulation/pages/p11.webp","publications/evacuation-simulation/pages/p12.webp","publications/evacuation-simulation/pages/p13.webp","publications/evacuation-simulation/pages/p14.webp","publications/evacuation-simulation/pages/p15.webp","publications/evacuation-simulation/pages/p16.webp","publications/evacuation-simulation/pages/p17.webp","publications/evacuation-simulation/pages/p18.webp","publications/evacuation-simulation/pages/p19.webp","publications/evacuation-simulation/pages/p20.webp","publications/evacuation-simulation/pages/p21.webp","publications/evacuation-simulation/pages/p22.webp"],aspect:.7067,pdf:"/publications/evacuation-simulation/report.pdf"},"forced-labor-structural-risk-index":{pages:["publications/forced-labor-structural-risk-index/pages/p01.webp","publications/forced-labor-structural-risk-index/pages/p02.webp","publications/forced-labor-structural-risk-index/pages/p03.webp","publications/forced-labor-structural-risk-index/pages/p04.webp","publications/forced-labor-structural-risk-index/pages/p05.webp","publications/forced-labor-structural-risk-index/pages/p06.webp","publications/forced-labor-structural-risk-index/pages/p07.webp","publications/forced-labor-structural-risk-index/pages/p08.webp","publications/forced-labor-structural-risk-index/pages/p09.webp","publications/forced-labor-structural-risk-index/pages/p10.webp","publications/forced-labor-structural-risk-index/pages/p11.webp","publications/forced-labor-structural-risk-index/pages/p12.webp","publications/forced-labor-structural-risk-index/pages/p13.webp","publications/forced-labor-structural-risk-index/pages/p14.webp","publications/forced-labor-structural-risk-index/pages/p15.webp","publications/forced-labor-structural-risk-index/pages/p16.webp","publications/forced-labor-structural-risk-index/pages/p17.webp","publications/forced-labor-structural-risk-index/pages/p18.webp","publications/forced-labor-structural-risk-index/pages/p19.webp","publications/forced-labor-structural-risk-index/pages/p20.webp","publications/forced-labor-structural-risk-index/pages/p21.webp","publications/forced-labor-structural-risk-index/pages/p22.webp","publications/forced-labor-structural-risk-index/pages/p23.webp","publications/forced-labor-structural-risk-index/pages/p24.webp","publications/forced-labor-structural-risk-index/pages/p25.webp"],aspect:.7067,pdf:"/publications/forced-labor-structural-risk-index/report.pdf"},haste:{pages:["publications/haste/pages/p01.webp","publications/haste/pages/p02.webp","publications/haste/pages/p03.webp","publications/haste/pages/p04.webp","publications/haste/pages/p05.webp","publications/haste/pages/p06.webp","publications/haste/pages/p07.webp","publications/haste/pages/p08.webp","publications/haste/pages/p09.webp","publications/haste/pages/p10.webp","publications/haste/pages/p11.webp","publications/haste/pages/p12.webp","publications/haste/pages/p13.webp","publications/haste/pages/p14.webp","publications/haste/pages/p15.webp","publications/haste/pages/p16.webp","publications/haste/pages/p17.webp","publications/haste/pages/p18.webp","publications/haste/pages/p19.webp","publications/haste/pages/p20.webp","publications/haste/pages/p21.webp","publications/haste/pages/p22.webp","publications/haste/pages/p23.webp","publications/haste/pages/p24.webp","publications/haste/pages/p25.webp","publications/haste/pages/p26.webp"],aspect:.7067,pdf:"/publications/haste/report.pdf"},"mariupol-severity-model":{pages:["publications/mariupol-severity-model/pages/p01.webp","publications/mariupol-severity-model/pages/p02.webp","publications/mariupol-severity-model/pages/p03.webp","publications/mariupol-severity-model/pages/p04.webp","publications/mariupol-severity-model/pages/p05.webp","publications/mariupol-severity-model/pages/p06.webp","publications/mariupol-severity-model/pages/p07.webp","publications/mariupol-severity-model/pages/p08.webp","publications/mariupol-severity-model/pages/p09.webp","publications/mariupol-severity-model/pages/p10.webp","publications/mariupol-severity-model/pages/p11.webp","publications/mariupol-severity-model/pages/p12.webp","publications/mariupol-severity-model/pages/p13.webp","publications/mariupol-severity-model/pages/p14.webp","publications/mariupol-severity-model/pages/p15.webp","publications/mariupol-severity-model/pages/p16.webp","publications/mariupol-severity-model/pages/p17.webp","publications/mariupol-severity-model/pages/p18.webp","publications/mariupol-severity-model/pages/p19.webp","publications/mariupol-severity-model/pages/p20.webp","publications/mariupol-severity-model/pages/p21.webp","publications/mariupol-severity-model/pages/p22.webp","publications/mariupol-severity-model/pages/p23.webp","publications/mariupol-severity-model/pages/p24.webp","publications/mariupol-severity-model/pages/p25.webp","publications/mariupol-severity-model/pages/p26.webp"],aspect:.7067,pdf:"/publications/mariupol-severity-model/report.pdf"},"provenance-search":{pages:["publications/provenance-search/pages/p01.webp","publications/provenance-search/pages/p02.webp","publications/provenance-search/pages/p03.webp","publications/provenance-search/pages/p04.webp","publications/provenance-search/pages/p05.webp","publications/provenance-search/pages/p06.webp","publications/provenance-search/pages/p07.webp","publications/provenance-search/pages/p08.webp","publications/provenance-search/pages/p09.webp","publications/provenance-search/pages/p10.webp","publications/provenance-search/pages/p11.webp","publications/provenance-search/pages/p12.webp","publications/provenance-search/pages/p13.webp","publications/provenance-search/pages/p14.webp","publications/provenance-search/pages/p15.webp","publications/provenance-search/pages/p16.webp","publications/provenance-search/pages/p17.webp","publications/provenance-search/pages/p18.webp","publications/provenance-search/pages/p19.webp","publications/provenance-search/pages/p20.webp"],aspect:.7067,pdf:"/publications/provenance-search/report.pdf"},vango:{pages:["publications/vango/pages/p01.webp","publications/vango/pages/p02.webp","publications/vango/pages/p03.webp","publications/vango/pages/p04.webp","publications/vango/pages/p05.webp","publications/vango/pages/p06.webp","publications/vango/pages/p07.webp","publications/vango/pages/p08.webp","publications/vango/pages/p09.webp","publications/vango/pages/p10.webp","publications/vango/pages/p11.webp","publications/vango/pages/p12.webp","publications/vango/pages/p13.webp","publications/vango/pages/p14.webp","publications/vango/pages/p15.webp","publications/vango/pages/p16.webp"],aspect:.7067,pdf:"/publications/vango/report.pdf"},"war-games":{pages:["publications/war-games/pages/p01.webp","publications/war-games/pages/p02.webp","publications/war-games/pages/p03.webp","publications/war-games/pages/p04.webp","publications/war-games/pages/p05.webp","publications/war-games/pages/p06.webp","publications/war-games/pages/p07.webp","publications/war-games/pages/p08.webp","publications/war-games/pages/p09.webp","publications/war-games/pages/p10.webp","publications/war-games/pages/p11.webp","publications/war-games/pages/p12.webp","publications/war-games/pages/p13.webp","publications/war-games/pages/p14.webp","publications/war-games/pages/p15.webp"],aspect:.7067,pdf:"/publications/war-games/report.pdf"},"what-is-ethical-ai":{pages:["publications/what-is-ethical-ai/pages/p01.webp","publications/what-is-ethical-ai/pages/p02.webp","publications/what-is-ethical-ai/pages/p03.webp","publications/what-is-ethical-ai/pages/p04.webp","publications/what-is-ethical-ai/pages/p05.webp","publications/what-is-ethical-ai/pages/p06.webp","publications/what-is-ethical-ai/pages/p07.webp","publications/what-is-ethical-ai/pages/p08.webp","publications/what-is-ethical-ai/pages/p09.webp","publications/what-is-ethical-ai/pages/p10.webp","publications/what-is-ethical-ai/pages/p11.webp","publications/what-is-ethical-ai/pages/p12.webp","publications/what-is-ethical-ai/pages/p13.webp","publications/what-is-ethical-ai/pages/p14.webp","publications/what-is-ethical-ai/pages/p15.webp","publications/what-is-ethical-ai/pages/p16.webp","publications/what-is-ethical-ai/pages/p17.webp","publications/what-is-ethical-ai/pages/p18.webp","publications/what-is-ethical-ai/pages/p19.webp","publications/what-is-ethical-ai/pages/p20.webp","publications/what-is-ethical-ai/pages/p21.webp","publications/what-is-ethical-ai/pages/p22.webp","publications/what-is-ethical-ai/pages/p23.webp","publications/what-is-ethical-ai/pages/p24.webp","publications/what-is-ethical-ai/pages/p25.webp","publications/what-is-ethical-ai/pages/p26.webp","publications/what-is-ethical-ai/pages/p27.webp","publications/what-is-ethical-ai/pages/p28.webp","publications/what-is-ethical-ai/pages/p29.webp","publications/what-is-ethical-ai/pages/p30.webp","publications/what-is-ethical-ai/pages/p31.webp","publications/what-is-ethical-ai/pages/p32.webp","publications/what-is-ethical-ai/pages/p33.webp","publications/what-is-ethical-ai/pages/p34.webp","publications/what-is-ethical-ai/pages/p35.webp","publications/what-is-ethical-ai/pages/p36.webp","publications/what-is-ethical-ai/pages/p37.webp","publications/what-is-ethical-ai/pages/p38.webp","publications/what-is-ethical-ai/pages/p39.webp","publications/what-is-ethical-ai/pages/p40.webp","publications/what-is-ethical-ai/pages/p41.webp","publications/what-is-ethical-ai/pages/p42.webp","publications/what-is-ethical-ai/pages/p43.webp","publications/what-is-ethical-ai/pages/p44.webp","publications/what-is-ethical-ai/pages/p45.webp","publications/what-is-ethical-ai/pages/p46.webp"],aspect:.7727,pdf:"/publications/what-is-ethical-ai/report.pdf"}};function l(e){if(!e)return;let a=e.match(/\/publications\/([^/?#]+)/);return a?n[a[1]]:void 0}var c=e.i(93209);let b={"Artificial Intelligence":["#0f3b57","#08131f"],Guidelines:["#1d2440","#0d1020"],Evacuation:["#3b1878","#160d1c"],"Cultural heritage":["#4a1d3d","#1a0d18"],Traceability:["#123a3a","#0a1a1c"],Diplomacy:["#2a2160","#100d20"],Sustainability:["#1c3a24","#0b1710"],"Disaster response":["#4a2417","#1a0e0a"]};function u(e,a=0){let[p,i]=b[e]??["#241a35","#120d1c"];return`linear-gradient(${130+31*a%60}deg, ${p} 0%, ${i} ${64+17*a%26}%)`}function d({topic:e,seed:p}){let i="var(--poster-accent)",s="color-mix(in oklab, var(--poster-accent) 26%, transparent)",t="color-mix(in oklab, var(--poster-accent) 12%, transparent)",o=(e,a=1)=>((37*p+11*e)%13-6)/6*a,r=.74+13*p%7/100;return(0,a.jsxs)("svg",{viewBox:"0 0 200 300","aria-hidden":!0,className:"absolute inset-0 h-full w-full",preserveAspectRatio:"xMidYMid slice",children:[(0,a.jsx)("g",{stroke:t,strokeWidth:.6,children:Array.from({length:9}).map((e,p)=>(0,a.jsx)("line",{x1:0,y1:30+30*p,x2:200,y2:30+30*p},p))}),(0,a.jsx)("g",{transform:`translate(${100-100*r+(7*p%11-5)} ${-34-5*p%14}) scale(${r}) rotate(${23*p%9-4} 100 170)`,children:(()=>{switch(e){case"Evacuation":return(0,a.jsxs)(a.Fragment,{children:[(0,a.jsx)("circle",{cx:40,cy:250,r:5,fill:i}),[{d:`M40 250 C 70 ${210+o(1,8)}, 120 ${200+o(2,8)}, 160 150`,dash:!1},{d:`M40 250 C 60 ${190+o(3,8)}, 90 ${140+o(4,8)}, 150 96`,dash:!0},{d:`M40 250 C 80 ${230+o(5,8)}, 130 ${230+o(6,8)}, 168 200`,dash:!1}].map((e,p)=>(0,a.jsx)("path",{d:e.d,fill:"none",stroke:e.dash?s:i,strokeWidth:1.4,strokeDasharray:e.dash?"5 5":void 0,opacity:e.dash?1:.75},p)),[[160,150],[150,96],[168,200]].map(([e,p],t)=>(0,a.jsx)("circle",{cx:e,cy:p,r:4,fill:"none",stroke:1===t?s:i,strokeWidth:1.4},t))]});case"Artificial Intelligence":{let e=[[126,178,230],[110,158,206,254],[166,212]].map((e,a)=>e.map(e=>[58+42*a,e+o(a,10)]));return(0,a.jsxs)(a.Fragment,{children:[e.slice(0,-1).map((i,o)=>i.flatMap((i,r)=>e[o+1].map((e,n)=>{let l=(p+3*o+2*r+n)%5==0;return(0,a.jsx)("line",{x1:i[0],y1:i[1],x2:e[0],y2:e[1],stroke:l?t:s,strokeWidth:l?1.4:1,strokeDasharray:l?"4 4":void 0},`${o}-${r}-${n}`)}))),e.flatMap((p,s)=>p.map(([p,t],o)=>(0,a.jsx)("circle",{cx:p,cy:t,r:s===e.length-1?5.5:4.5,fill:s===e.length-1?i:"none",stroke:i,strokeWidth:1.4,opacity:1===s?.6:1},`n-${s}-${o}`)))]})}case"Guidelines":return(0,a.jsx)(a.Fragment,{children:Array.from({length:6}).map((e,r)=>{let n=110+26*r,l=(p+r)%3!=0;return(0,a.jsxs)("g",{children:[(0,a.jsx)("rect",{x:38,y:n-7,width:14,height:14,rx:3,fill:"none",stroke:l?i:t,strokeWidth:1.4}),l&&(0,a.jsx)("path",{d:`M41.5 ${n} l3 3.5 l6.5 -7.5`,fill:"none",stroke:i,strokeWidth:1.6,strokeLinecap:"round",strokeLinejoin:"round"}),(0,a.jsx)("rect",{x:62,y:n-3,width:70+o(r,26),height:3,rx:1.5,fill:l?s:t})]},r)})});case"Cultural heritage":return(0,a.jsxs)(a.Fragment,{children:[[0,1,2].map(e=>(0,a.jsx)("path",{d:`M${60+14*e} 250 L${60+14*e} ${170-16*e} A${40-14*e} ${40-14*e} 0 0 1 ${140-14*e} ${170-16*e} L${140-14*e} 250`,fill:"none",stroke:0===e?i:s,strokeWidth:1.4,opacity:1-.18*e},e)),(0,a.jsx)("circle",{cx:100,cy:198,r:11,fill:"none",stroke:i,strokeWidth:1.4}),(0,a.jsx)("path",{d:"M95.5 198 l3 3.4 l6 -7",fill:"none",stroke:i,strokeWidth:1.5,strokeLinecap:"round",strokeLinejoin:"round"})]});case"Traceability":return(0,a.jsx)(a.Fragment,{children:Array.from({length:5}).map((e,o)=>{let r=o===2+p%2,n=110+30*o;return(0,a.jsxs)("g",{children:[(0,a.jsx)("rect",{x:62,y:n,width:76,height:20,rx:10,fill:"none",stroke:r?t:i,strokeWidth:1.4,strokeDasharray:r?"4 4":void 0,opacity:r?1:.8}),o<4&&(0,a.jsx)("line",{x1:100,y1:n+20,x2:100,y2:n+30,stroke:r?t:s,strokeWidth:1.4})]},o)})});case"Diplomacy":{let e=[0,1,2,3,4].map(e=>{let a=96+o(e,30),p=104-o(e+3,30);return{y:108+28*e,left:a,right:p,agreed:a>p}});return(0,a.jsxs)(a.Fragment,{children:[(0,a.jsx)("line",{x1:100,y1:92,x2:100,y2:250,stroke:t,strokeWidth:1.4}),e.map((e,p)=>(0,a.jsxs)("g",{children:[(0,a.jsx)("line",{x1:34,y1:e.y,x2:e.left,y2:e.y,stroke:e.agreed?i:s,strokeWidth:e.agreed?2.4:1.6,strokeLinecap:"round"}),(0,a.jsx)("line",{x1:e.right,y1:e.y,x2:166,y2:e.y,stroke:e.agreed?i:s,strokeWidth:e.agreed?2.4:1.6,strokeLinecap:"round"}),e.agreed&&(0,a.jsx)("circle",{cx:(e.left+e.right)/2,cy:e.y,r:3.6,fill:i})]},p))]})}case"Sustainability":return(0,a.jsxs)(a.Fragment,{children:[(0,a.jsx)("path",{d:"M34 250 L34 110 M34 250 L170 250",fill:"none",stroke:t,strokeWidth:1.4}),(0,a.jsx)("path",{d:`M34 ${232+o(1,6)} C 74 ${232+o(1,6)}, 92 130, 170 ${118+o(2,8)}`,fill:"none",stroke:s,strokeWidth:1.4,strokeDasharray:"4 4"}),(0,a.jsx)("path",{d:`M34 ${232+o(1,6)} C 74 ${232+o(1,6)}, 96 168, 170 ${186+o(3,8)}`,fill:"none",stroke:i,strokeWidth:1.6}),[[34,232],[170,186]].map(([e,p],s)=>(0,a.jsx)("circle",{cx:e,cy:p+o(s+1,6),r:4,fill:i},s))]});case"Disaster response":return(0,a.jsxs)(a.Fragment,{children:[[54,40,26].map((e,p)=>(0,a.jsx)("circle",{cx:100,cy:180,r:e,fill:"none",stroke:0===p?s:t,strokeWidth:1.4},p)),(0,a.jsx)("line",{x1:100,y1:180,x2:100+54*Math.cos(p%6-2.4),y2:180+54*Math.sin(p%6-2.4),stroke:i,strokeWidth:1.6}),(0,a.jsx)("circle",{cx:100,cy:180,r:4.5,fill:i})]});default:return null}})()})]})}function g(e){let a=e.indexOf(":");return -1===a||a>48?[e,null]:[e.slice(0,a),e.slice(a+1).trim()]}function w(e){return!!e.url&&"internal"!==e.access}function h({pub:e,onOpen:p}){let[i,s]=g(e.title),t=Number(e.index),r=!!l(e.url);return(0,a.jsx)(o.Tilt3D,{max:5,className:"h-full",children:(0,a.jsxs)("button",{type:"button",onClick:p,"aria-label":w(e)?`Open ${e.title}`:`${e.title}, ${e.status.toLowerCase()}`,className:"group/cover relative flex aspect-[2/3] h-full w-full flex-col justify-end overflow-hidden rounded-xl border border-border p-4 text-left transition-colors hover:border-accent focus-visible:border-accent focus-visible:outline-none",style:{background:u(e.topic,t)},children:[(0,a.jsx)(d,{topic:e.topic,seed:t}),(0,a.jsx)("span",{"aria-hidden":!0,className:"absolute inset-0",style:{background:"linear-gradient(180deg, transparent 34%, color-mix(in oklab, #0c0812 68%, transparent) 58%, #0c0812 92%)"}}),r&&(0,a.jsxs)("span",{className:"absolute left-3 top-3 inline-flex items-center gap-1 rounded-full px-2 py-0.5 font-mono text-[9px] font-semibold uppercase tracking-wider shadow-sm",style:{background:"var(--poster-accent)",color:"var(--poster-accent-ink)"},children:[(0,a.jsx)("span",{"aria-hidden":!0,children:"📖"})," Book"]}),(0,a.jsx)("span",{className:"absolute right-3 top-3",children:"internal"===e.access?(0,a.jsx)("span",{className:"rounded-full border border-white/35 bg-black/45 px-2 py-0.5 font-mono text-[9px] uppercase tracking-wider text-white/85 backdrop-blur",children:"CoLab only"}):e.url?(0,a.jsx)("span",{className:"rounded-full px-2 py-0.5 font-mono text-[9px] font-semibold uppercase tracking-wider",style:{background:"var(--poster-accent)",color:"var(--poster-accent-ink)"},children:"Published"}):(0,a.jsx)("span",{className:"rounded-full border border-border px-2 py-0.5 font-mono text-[9px] uppercase tracking-wider",style:{color:"var(--poster-ink-muted)"},children:"In prep"})}),w(e)&&(0,a.jsx)("span",{className:"absolute inset-0 flex items-center justify-center opacity-0 transition-opacity group-hover/cover:opacity-100 group-focus-visible/cover:opacity-100",children:(0,a.jsx)("span",{className:"inline-flex items-center gap-2 rounded-full px-4 py-2 text-sm font-semibold shadow-lg",style:{background:"var(--poster-accent)",color:"var(--poster-accent-ink)"},children:"Read"})}),(0,a.jsxs)("div",{className:"relative",children:[(0,a.jsx)("p",{className:"font-mono text-[10px] uppercase tracking-wider",style:{color:"var(--poster-accent)"},children:e.topic}),(0,a.jsx)("h3",{className:"mt-1.5 line-clamp-3 font-heading text-xl uppercase leading-[1.05] tracking-[0.02em] text-[color:var(--poster-ink)] transition-colors group-hover/cover:text-[color:var(--poster-accent)]",children:i}),s&&(0,a.jsx)("p",{className:"mt-1.5 line-clamp-2 text-[11px] leading-snug",style:{color:"var(--poster-ink-muted)"},children:s}),(0,a.jsx)("div",{className:"mt-3 font-mono text-[10px] uppercase tracking-wider",style:{color:"var(--poster-ink-muted)"},children:e.date??e.status})]})]})})}function m({pub:e,onClose:s}){(0,p.useEffect)(()=>{let e=e=>"Escape"===e.key&&s();return document.addEventListener("keydown",e),document.body.style.overflow="hidden",()=>{document.removeEventListener("keydown",e),document.body.style.overflow=""}},[s]);let[t,o]=g(e.title),n=e.url.startsWith("/"),b=l(e.url),w="flex w-full items-center justify-between gap-4 rounded-xl border border-border bg-card px-5 py-4 text-left transition-colors hover:border-accent";return(0,a.jsxs)("div",{className:"fixed inset-0 z-[100] flex flex-col bg-background",role:"dialog","aria-modal":"true","aria-label":e.title,children:[(0,a.jsxs)("div",{className:"flex items-center justify-between gap-3 border-b border-border px-4 py-2.5",children:[(0,a.jsxs)("div",{className:"min-w-0",children:[(0,a.jsx)("h3",{className:"truncate font-heading text-xl uppercase tracking-[0.02em] sm:text-2xl",children:t}),(0,a.jsxs)("p",{className:"truncate font-mono text-[10px] text-muted",children:[e.topic,e.date&&` \xb7 ${e.date}`]})]}),(0,a.jsx)("button",{type:"button",onClick:s,className:"shrink-0 rounded-full bg-accent px-4 py-1.5 text-sm font-semibold text-accent-ink transition-transform hover:scale-[1.03]",children:"✕ Close"})]}),(0,a.jsx)("div",{className:"flex-1 overflow-y-auto",children:(0,a.jsxs)("div",{className:"mx-auto grid max-w-4xl gap-10 px-6 py-12 sm:grid-cols-[minmax(0,200px)_1fr] sm:py-14",children:[(0,a.jsx)("div",{className:"relative hidden aspect-[2/3] w-full overflow-hidden rounded-xl border border-border sm:block",style:{background:u(e.topic,Number(e.index))},children:(0,a.jsx)(d,{topic:e.topic,seed:Number(e.index)})}),(0,a.jsxs)("div",{children:[o&&(0,a.jsx)("p",{className:"font-heading text-lg uppercase leading-snug tracking-wide text-muted",children:o}),(0,a.jsx)("p",{className:"mt-4 border-l-2 border-accent pl-5 text-lg leading-relaxed text-foreground/90",children:e.question}),(0,a.jsx)("p",{className:"mt-6 leading-relaxed text-muted",children:e.summary}),(0,a.jsx)("div",{className:"mt-8 space-y-3",children:"internal"===e.access?(0,a.jsxs)("a",{href:e.url,target:"_blank",rel:"noopener noreferrer",className:w,children:[(0,a.jsxs)("span",{className:"font-semibold text-foreground",children:["Open in the CoLab repo"," ",(0,a.jsx)("span",{className:"font-normal text-muted",children:"(sign-in required)"})]}),(0,a.jsx)("span",{"aria-hidden":!0,className:"text-accent",children:"↗"})]}):e.url?(0,a.jsxs)(a.Fragment,{children:[n?(0,a.jsxs)(i.Link,{href:e.url,className:w,children:[(0,a.jsx)("span",{className:"font-semibold text-foreground",children:"Read the report"}),(0,a.jsx)("span",{"aria-hidden":!0,className:"text-accent",children:"→"})]}):(0,a.jsxs)("a",{href:e.url,target:"_blank",rel:"noopener noreferrer",className:w,children:[(0,a.jsx)("span",{className:"font-semibold text-foreground",children:"Read the report"}),(0,a.jsx)("span",{"aria-hidden":!0,className:"text-accent",children:"↗"})]}),b&&(0,a.jsxs)(r.ReportBook,{title:t,pages:b.pages,aspect:b.aspect,pdfUrl:e.pdf??(0,c.asset)(b.pdf),className:w,children:[(0,a.jsxs)("span",{className:"font-semibold text-foreground",children:["Read as book"," ",(0,a.jsx)("span",{className:"font-normal text-muted",children:"(flip the original pages)"})]}),(0,a.jsx)("span",{"aria-hidden":!0,className:"text-accent",children:"📖"})]}),e.pdf&&(0,a.jsxs)("a",{href:e.pdf,target:"_blank",rel:"noopener noreferrer",className:w,children:[(0,a.jsx)("span",{className:"font-semibold text-foreground",children:"Download the PDF"}),(0,a.jsx)("span",{"aria-hidden":!0,className:"text-accent",children:"↗"})]})]}):(0,a.jsxs)("p",{className:"rounded-xl border border-dashed border-border px-5 py-4 text-sm text-muted",children:["This report is ",e.status.toLowerCase(),". The research it writes up is already running, so the demos and the source are the way in until it publishes."]})}),(0,a.jsxs)("div",{className:"mt-8 flex flex-wrap gap-x-4 gap-y-2 border-t border-border pt-6 font-mono text-xs text-muted",children:[(0,a.jsx)("span",{children:e.area}),(0,a.jsx)("span",{children:e.status}),e.date&&(0,a.jsx)("span",{children:e.date}),e.repo&&(0,a.jsx)("a",{href:e.repo,target:"_blank",rel:"noopener noreferrer",className:"ml-auto text-foreground/80 transition-colors hover:text-accent",children:"View source ↗"})]})]})]})})]})}e.s(["PublicationsShowcase",0,function(){let[e,i]=(0,p.useState)(null),[o,r]=(0,p.useState)(null),[n,l]=(0,p.useState)(null),c=s.publications.items.filter(a=>(!e||a.topic===e)&&("published"!==o||!!w(a))),b=c.filter(w).length,u=c.filter(e=>"internal"===e.access).length,d=s.publicationTopics.map(e=>({topic:e,items:c.filter(a=>a.topic===e)})).filter(e=>e.items.length>0),g=e=>`rounded-full border px-3 py-1.5 text-sm transition-colors ${e?"border-accent bg-accent font-semibold text-accent-ink":"border-border text-muted hover:border-accent hover:text-accent"}`;return(0,a.jsxs)("div",{className:"mx-auto max-w-6xl px-6 pb-24",children:[(0,a.jsxs)("div",{className:"space-y-4 border-b border-border pb-8",children:[(0,a.jsxs)("div",{className:"flex flex-wrap items-center gap-2",children:[(0,a.jsx)("span",{className:"mr-2 w-20 shrink-0 text-xs uppercase tracking-wider text-muted",children:"Topic"}),(0,a.jsx)("button",{type:"button",onClick:()=>i(null),className:g(null===e),children:"All"}),s.publicationTopics.filter(e=>s.publications.items.some(a=>a.topic===e)).map(p=>(0,a.jsx)("button",{type:"button",onClick:()=>i(e=>e===p?null:p),className:g(e===p),children:p},p)),(0,a.jsxs)("span",{className:"ml-auto font-mono text-xs text-muted",children:[b," published",u>0&&` \xb7 ${u} CoLab only`," · ",c.length," ","shown"]})]}),(0,a.jsxs)("div",{className:"flex flex-wrap items-center gap-2",children:[(0,a.jsx)("span",{className:"mr-2 w-20 shrink-0 text-xs uppercase tracking-wider text-muted",children:"Status"}),(0,a.jsx)("button",{type:"button",onClick:()=>r(null),className:g(null===o),children:"All"}),(0,a.jsx)("button",{type:"button",onClick:()=>r(e=>"published"===e?null:"published"),className:g("published"===o),children:"Readable now"})]})]}),(0,a.jsx)("div",{className:"mt-12 space-y-14",children:d.map(e=>(0,a.jsx)(t.PosterRail,{title:e.topic,count:e.items.length,countNoun:"report",ariaLabel:`${e.topic} reports`,children:e.items.map(e=>(0,a.jsx)("div",{className:"w-[190px] shrink-0 snap-start scroll-mt-24 sm:w-[215px]",children:(0,a.jsx)(h,{pub:e,onOpen:()=>l(e)})},e.index))},e.topic))}),0===c.length&&(0,a.jsx)("p",{className:"py-16 text-muted",children:"No reports match those filters."}),n&&(0,a.jsx)(m,{pub:n,onClose:()=>l(null)},n.index)]})}],46973)},43413,e=>{e.v(a=>Promise.all(["static/chunks/1qa2_w3ql48wi.js"].map(a=>e.l(a))).then(()=>a(62914)))}]); \ No newline at end of file diff --git a/static-site/_next/static/chunks/1-kh6uz_49jxr.js b/static-site/_next/static/chunks/1-kh6uz_49jxr.js deleted file mode 100644 index a5d25fc90..000000000 --- a/static-site/_next/static/chunks/1-kh6uz_49jxr.js +++ /dev/null @@ -1 +0,0 @@ -(globalThis.TURBOPACK||(globalThis.TURBOPACK=[])).push(["object"==typeof document?document.currentScript:void 0,89042,e=>{"use strict";let a=["NYU School of Professional Studies","Center for Global Affairs (CGA)","Microsoft","Generative AI for Good","The Microsoft Garage","Tavily","Tradeverifyd","Sayari","B3IQ","Art & Antiquities Blockchain Consortium","100x","D_ID","Apne Aap Women Worldwide","Gaia","OSCE / ODIHR Anti-Trafficking","Coinbase","Rivr","x402 Foundation","Blockchain for Social Impact"];[{name:"NYU School of Professional Studies",about:"The NYU school that houses the Center for Global Affairs and the Ethical Tech CoLab.",url:"https://www.sps.nyu.edu",logo:"/logos/nyu-sps.jpg"},{name:"Center for Global Affairs (CGA)",about:"NYU SPS's Center for Global Affairs — the CoLab's academic home, focused on global policy and the human condition.",url:"https://www.sps.nyu.edu/homepage/academics/divisions-and-departments/center-for-global-affairs.html",logo:"/nyu-sps-cga-logo.jpg"},{name:"Microsoft",about:"Technology partner across proof-of-human, AI, web3, and privacy-preserving systems such as zero-knowledge proofs.",url:"https://www.microsoft.com",logo:"/logos/microsoft.svg"},{name:"Generative AI for Good",about:"Uses generative media for human-condition storytelling — making lived experience legible to audiences and decision-makers.",url:"",logo:"/logos/generative-ai-for-good.jpg"},{name:"D_ID",about:"Avatars and digital-human tooling for synthetic media, used to prototype human-condition storytelling.",url:"https://www.d-id.com",logo:"/logos/d-id.jpg"},{name:"Tradeverifyd",about:"Ask your supply chain anything — agentic AI that resolves suppliers across 200+ global data sources, maps ownership through tier three, and monitors exposure to forced labor and sanctions regimes. Founded as Mesur.io in 2016; rebranded in 2025.",url:"https://tradeverifyd.com",logo:"/logos/tradeverifyd.svg"},{name:"Rivr",about:"Twitch livestreams, keyword tags, and chat-room activity as a real-time data source for detecting at-risk behavior online.",url:"",logo:"/logos/rivr.jpg"},{name:"100x",about:"",url:"",logo:"/logos/100x.jpg"},{name:"Apne Aap Women Worldwide",about:"",url:"https://apneaap.org",logo:"/logos/apne-aap-women-worldwide.png"},{name:"Gaia",about:"",url:"",logo:"/logos/gaia.png"},{name:"OSCE / ODIHR Anti-Trafficking",about:"Multilateral partner on anti-trafficking, platform safety, and human rights — a recurring voice at the Ethical Tech Summit.",url:"https://www.osce.org/odihr",logo:""},{name:"Coinbase",about:"Industry partner on agentic commerce and micropayment rails, engaged through the Ethical Tech Summit.",url:"https://www.coinbase.com",logo:"/logos/coinbase.png"},{name:"x402 Foundation",about:"Stewards the x402 payment standard behind the CoLab's agentic-commerce and micropayment work, engaged through the Ethical Tech Summit.",url:"https://www.x402.org",logo:"/logos/x402-foundation.jpg"},{name:"Sayari",about:"Maps global corporate ownership and supply-chain relationships across more than 250 jurisdictions, surfacing hidden counterparty and compliance risk.",url:"https://sayari.com",logo:"/logos/sayari.jpg"},{name:"B3IQ",about:"US-based AI infrastructure: NVIDIA-powered GPU servers that buyers own, racked and run in a data center in Eugene, Oregon. B3IQ hosts the CoLab's GPU machine, which the cohort reaches over SSH for model work.",url:"https://b3iq.org",logo:"/logos/b3iq.png",logoTile:"dark"},{name:"Art & Antiquities Blockchain Consortium",about:"A consortium working on cultural-heritage ownership and repatriation, bringing stakeholders, technologies, and cultural artifacts together.",url:"https://aabconsortium.org",logo:"/logos/art-antiquities-blockchain-consortium.png"},{name:"Blockchain for Social Impact",about:"A not-for-profit coalition that incubates and develops blockchain solutions addressing environmental and social challenges, convening nonprofits, governments, investors, and technologists.",url:"https://blockchainforsocialimpact.com",logo:"/logos/blockchain-for-social-impact.png"},{name:"The Microsoft Garage",about:"Microsoft's worldwide innovation program — local Garage sites, the Global Hackathon, and the Innovation Studio platform — where employees and customers turn experimental ideas into working projects.",url:"https://www.microsoft.com/en-us/garage/",logo:"/logos/microsoft-garage.png"},{name:"Tavily",about:"A web search and extraction API built for AI agents rather than people, returning clean extracted text with the source address of every result. The CoLab uses it as the restricted research engine behind the provenance work and the tavily-research agent skill.",url:"https://tavily.com",logo:"/logos/tavily.svg"}].sort((e,t)=>{let i=a.indexOf(e.name),o=a.indexOf(t.name);return(-1===i?Number.MAX_SAFE_INTEGER:i)-(-1===o?Number.MAX_SAFE_INTEGER:o)}),e.s(["areaSlug",0,e=>e.toLowerCase().replace(/\s+/g,"-"),"nav",0,[{label:"Home",href:"/"},{label:"Portfolio",href:"/portfolio",children:[{label:"Overview",href:"/portfolio"},{label:"Live Demos",href:"/demos"},{label:"Publications",href:"/publications"},{label:"Media",href:"/media"},{label:"Newsletter",href:"/newsletter"}]},{label:"Team",href:"/team"}],"productTerms",0,["Summer 2026","Fall 2025","Spring 2025"],"productThemes",0,["Evacuation","Cultural heritage","Traceability","Early warning","Diplomacy","Research","Storytelling"],"products",0,[{name:"Diplomatic Simulator",repoName:"diplomatic-simulator",term:"Summer 2026",repo:"https://github.com/Ethical-Tech-CoLab/diplomatic-simulator",demo:"https://ethical-tech-colab.github.io/diplomatic-simulator/live.html",demos:[{label:"Watch a live negotiation",href:"https://ethical-tech-colab.github.io/diplomatic-simulator/live.html"},{label:"Interactive dashboard",href:"https://ethical-tech-colab.github.io/diplomatic-simulator/dashboard.html"},{label:"Monte Carlo analysis",href:"https://ethical-tech-colab.github.io/diplomatic-simulator/montecarlo.html"},{label:"Methodology & AI use",href:"https://ethical-tech-colab.github.io/diplomatic-simulator/methodology.html"}],blurb:"A multi-party AI negotiation table where each delegation argues from its own confidential 'privileged instructions.' Five crises — Arctic, Central Asia, Cyprus, South China Sea, and Iran–US Strait of Hormuz — each with a live-negotiation replay, an interactive analysis dashboard (parties' strategy, proposals, detector hits), and Monte Carlo runs over randomized conditions.",language:"HTML",theme:"Diplomacy",featured:!0,publication:"/publications/diplomatic-simulator"},{name:"AI Models Research",repoName:"AI-Models-Research",term:"Summer 2026",repo:"https://github.com/Ethical-Tech-CoLab/AI-Models-Research",demo:"https://ethical-tech-colab.github.io/AI-Models-Research/",demos:[{label:"Read the handbook",href:"https://ethical-tech-colab.github.io/AI-Models-Research/"},{label:"Cost-per-accepted-task calculator",href:"https://ethical-tech-colab.github.io/AI-Models-Research/interactive/"}],blurb:"A comparative review of frontier, open-weight, and efficient model families across benchmark performance, factual accuracy, latency, token economics, and energy use. Three evidence grades separate independently verified measurement from institutional research and from provider marketing, and every figure names the party that produced it. Ships with a live cost-per-accepted-task calculator over a validated data layer.",language:"Python",theme:"Research",featured:!0,publication:"/publications/ai-models-research"},{name:"Exodus — Civilian Evacuation Risk Platform",repoName:"Exodus",term:"Summer 2026",repo:"https://github.com/Ethical-Tech-CoLab/Exodus",demo:"https://exodus-ruddy.vercel.app",blurb:"Three civilian-evacuation risk tools behind one backend and design language: a crisis map of the INFORM Severity Index with live news and conflict timelines, a seven-dimension scenario risk model, and an endangerment-and-feasibility assessment.",language:"TypeScript",theme:"Evacuation",featured:!0},{name:"Evacuation Inform Index",repoName:"evacuation-inform-index-carolina",term:"Summer 2026",repo:"https://github.com/Ethical-Tech-CoLab/evacuation-inform-index-carolina",demo:"https://ethical-tech-colab.github.io/evacuation-inform-index-carolina/",blurb:"A composite index scoring 104 active crises on the risk of staying and the risk of leaving, side by side, so the comparison between them can be argued with. Built on INFORM Severity data with live news and an ACLED conflict timeline, filterable by crisis type, and with every one of its thirteen legal citations explained — including how far each one is actually justified.",language:"HTML",theme:"Evacuation",featured:!0,publication:"/publications/evacuation-inform-index"},{name:"Evacuation Routing Simulator",repoName:"India-EvacSimulation",term:"Summer 2026",repo:"https://github.com/Ethical-Tech-CoLab/India-EvacSimulation",demo:"https://ethical-tech-colab.github.io/India-EvacSimulation/",blurb:"An interactive simulator modeling how routing and assignment choices shape who reaches safety, with shareable scenario state and a built-in explainer walkthrough.",language:"HTML",theme:"Evacuation",publication:"/publications/erus"},{name:"Mariupol 2022 — Corridor Severity Model",repoName:"mariupol-evacuation-model",term:"Summer 2026",repo:"https://github.com/Ethical-Tech-CoLab/mariupol-evacuation-model",demo:"https://ethical-tech-colab.github.io/mariupol-evacuation-model/",blurb:"A corridor-severity model reconstructing the 2022 Mariupol evacuation, scoring the risk and viability of humanitarian corridors under siege conditions.",language:"HTML",theme:"Evacuation",publication:"/publications/mariupol-severity-model"},{name:"Evacuation Behavior Simulator",repoName:"Evac-Sim-Melanie",term:"Summer 2026",repo:"https://github.com/Ethical-Tech-CoLab/Evac-Sim-Melanie",demo:"https://ethical-tech-colab.github.io/Evac-Sim-Melanie/",blurb:"An agent-based model of community evacuation behavior — family clusters, information-seeking, neighbor influence, and elder and child delays across an UNAWARE → SEEKING → MILLING → EVACUATING → DONE lifecycle.",language:"JavaScript",theme:"Evacuation",publication:"/publications/evacuation-simulation"},{name:"Digital Art Passport (VANGO)",repoName:"VANGO",term:"Summer 2026",repo:"https://github.com/Ethical-Tech-CoLab/VANGO",demo:"https://ethical-tech-colab.github.io/VANGO/",blurb:"A digital provenance passport for cultural artifacts, pairing verifiable credentials with collaborative cataloguing to support repatriation, attribution, and ethical stewardship.",language:"HTML",theme:"Cultural heritage",publication:"/publications/vango"},{name:"ERCF — Evacuation Risk & Cost Framework",repoName:"ercf",term:"Summer 2026",repo:"https://github.com/Ethical-Tech-CoLab/ercf",demo:"https://ercf-production.up.railway.app",blurb:"A decision-support framework estimating the human and financial cost of civilian evacuation in armed conflict, scoring scenarios across seven risk dimensions and comparing evacuating against staying.",language:"Python",theme:"Evacuation",publication:"/publications/ercf"},{name:"CERAI — Civilian Evacuation Risk Anticipation Index",repoName:"CERAI_AR",term:"Summer 2026",repo:"https://github.com/Ethical-Tech-CoLab/CERAI_AR",demo:"https://ethical-tech-colab.github.io/CERAI_AR/",blurb:"A composite index anticipating civilian risk during evacuation operations, scoring weighted categories — hazards, infrastructure, population vulnerability, information environment, and weather — with Monte Carlo sensitivity analysis, ACLED conflict-event lookup, and source-credibility tagging on every indicator, so evacuation-risk reasoning stays auditable for IHL compliance review. Built by Alana Robertson in the Summer 2026 cohort, on Teresa Cantero's doctoral research into AI and the protection and evacuation of civilians under IHL.",language:"HTML",theme:"Evacuation",publication:"/publications/cerai"},{name:"Arts Provenance Agent",repoName:"arts-provenance-agent",term:"Summer 2026",repo:"https://github.com/Ethical-Tech-CoLab/arts-provenance-agent",demo:"https://ethical-tech-colab.github.io/arts-provenance-agent/",blurb:"An x402-native agent that traces artwork provenance, flags looting and repatriation risk, and issues a signed, tamper-evident Passport for each object. This live demo runs on mock data — catalog dashboards, passport issue/verify, and a replayed agent trace.",language:"TypeScript",theme:"Cultural heritage",featured:!0,publication:"/publications/digital-provenance-passport"},{name:"Provenance Search — Arts & Artifacts",repoName:"provenance-search",term:"Summer 2026",repo:"https://github.com/Ethical-Tech-CoLab/provenance-search",demo:"https://provenance-search-production.up.railway.app",blurb:"Traces the ownership chain of an artwork across museum collections, cultural archives and loss registries, then issues a provenance passport carrying a rule-based confidence score you can inspect rather than a verdict you have to trust. Seven sources checked, three ways to search, no sign-up.",language:"HTML",theme:"Cultural heritage",publication:"/publications/provenance-search"},{name:"Digital Passport for Artworks",repoName:"digital-passport-artworks",term:"Summer 2026",repo:"https://github.com/Ethical-Tech-CoLab/digital-passport-artworks",demo:"https://ethical-tech-colab.github.io/digital-passport-artworks/",blurb:"Upload one photograph and watch the whole chain run in order: fingerprint the image, search for a prior record, score it for forgery risk, and only once it clears — or a reviewer overrides — sign and issue the passport. Real ECDSA cryptography, entirely in the browser. Originally by Christine Lumen.",language:"HTML",theme:"Cultural heritage"},{name:"Forced Labor Structural Risk Index",repoName:"forced-labor-structural-risk-index",term:"Fall 2025",repo:"https://github.com/Ethical-Tech-CoLab/forced-labor-structural-risk-index",demo:"https://ethical-tech-colab.github.io/forced-labor-structural-risk-index/",blurb:"An interactive index mapping the structural conditions that enable forced labor across 184 countries on a 0–1 risk scale — with a choropleth world map, national and sub-national layers, rankings, and sources. Built by Amanda Lindsey in the Fall 2025 cohort.",language:"HTML",theme:"Traceability",featured:!0,publication:"/publications/forced-labor-structural-risk-index"},{name:"AI Research Question Assistant",repoName:"ai-research-question-assistant",term:"Fall 2025",repo:"https://github.com/Ethical-Tech-CoLab/ai-research-question-assistant",demo:"https://ethical-tech-colab.github.io/ai-research-question-assistant/",blurb:"A survey and toolkit for how AI helps researchers move from a broad interest to a well-formed research question — finding gaps, generating candidate questions, summarizing the state of the art, and flagging contradictions, with guardrails against hallucinated citations. Ships with a reusable Copilot 'Researcher' prompt and a journal-credibility rubric. Built by the Fall 2025 cohort.",language:"HTML",theme:"Research",publication:"/publications/ai-research-assistant"},{name:"AI's Carbon Footprint",repoName:"ai-carbon-footprint",term:"Spring 2025",repo:"https://github.com/Ethical-Tech-CoLab/ai-carbon-footprint",demo:"https://ethical-tech-colab.github.io/ai-carbon-footprint/",blurb:"A report on what artificial intelligence costs the environment: energy drawn across training and inference, the data-center and hardware toll behind it, and the mitigations, regulations and policies that could bend the curve. 1,287 MWh went into training GPT-3 alone — roughly what 130 US households use in a year.",language:"HTML",theme:"Research",publication:"/publications/ai-carbon-footprint"},{name:"GitHub 101",repoName:"github-101",term:"Summer 2026",access:"internal",repo:"https://github.com/Ethical-Tech-CoLab/Getting-Going/blob/main/docs/github-101.md",blurb:"A beginner's reference for Git and GitHub from day one: what the two actually are and how they differ, the key concepts, setting up, and the handful of commands that cover almost everything you will do in the first months.",language:"Markdown",theme:"Research"},{name:"GitHub Workflow",repoName:"github-workflow",term:"Summer 2026",access:"internal",repo:"https://github.com/Ethical-Tech-CoLab/Getting-Going/blob/main/docs/github-workflow.md",blurb:"The collaborative loop, step by step: the GitHub flow at a glance, creating and managing branches, opening a pull request that can actually be reviewed, and what to look for when reviewing someone else's.",language:"Markdown",theme:"Research"},{name:"GitHub Pages 101",repoName:"github-pages-101",term:"Summer 2026",access:"internal",repo:"https://github.com/Ethical-Tech-CoLab/Getting-Going/blob/main/docs/github-pages-101.md",blurb:"Publishing a free site straight from a repository, for first-time users — no server to rent, nothing to install. Includes an honest account of what GitHub Pages can and cannot do, which is where most first attempts come unstuck.",language:"Markdown",theme:"Research"},{name:"Synthetic Data Guide",repoName:"synthetic-data-guide",term:"Summer 2026",access:"internal",repo:"https://github.com/Ethical-Tech-CoLab/Getting-Going/blob/main/synthetic-data/SYNTHETIC_DATA_GUIDE.md",blurb:"Generating data algorithmically rather than collecting it: why you would — developing before real data exists, forcing edge cases, sharing something realistic without privacy exposure, reproducing a dataset from a seed — and how to do it without fooling yourself about what the results mean.",language:"Markdown",theme:"Research"},{name:"Synthetic Data — Training Deck",repoName:"synthetic-data-training",term:"Summer 2026",access:"internal",repo:"https://github.com/Ethical-Tech-CoLab/Getting-Going/blob/main/synthetic-data/synthetic-data-training.html",blurb:"The companion training material to the synthetic-data guide — a self-contained walkthrough for running the topic as a session rather than reading it alone.",language:"HTML",theme:"Research"},{name:"tavily-research",repoName:"tavily-research",term:"Summer 2026",access:"internal",repo:"https://github.com/Ethical-Tech-CoLab/Getting-Going/tree/main/tavily-research",blurb:"A cross-agent skill wrapping the Tavily REST API — search, extract, crawl and map — with a dependency-free client in both Python and Node. It calls the API directly, so it still works where a Tavily MCP server cannot be launched, such as under IT policy that blocks local npx.",language:"Python · Node",theme:"Research"},{name:"Getting Started on Cloud GPU",repoName:"cloud-gpu-guide",term:"Summer 2026",access:"internal",repo:"https://github.com/Ethical-Tech-CoLab/Getting-started-on-CloudGPU-/blob/main/remote-gpu-how-to-guide.md",blurb:"Getting a terminal on the CoLab's B3IQ machine over SSH. One-time setup, about five minutes: installing cloudflared for the secure tunnel, creating an SSH key if you do not have one, and connecting from macOS, Windows or Linux.",language:"Markdown",theme:"Research"},{name:"War Games",repoName:"War-Games",term:"Summer 2026",repo:"https://github.com/Ethical-Tech-CoLab/War-Games",demo:"https://ethical-tech-colab.github.io/War-Games/",blurb:"A terminal thriller in the shape of WarGames (1983), reframed around a modern AI agent: the only winning move is to understand the machine. Ships with a Monte Carlo simulation harness and a written case study behind the fiction.",language:"JavaScript",theme:"Diplomacy"},{name:"Malawi Prevention Platform (MVDC)",repoName:"MVDC",term:"Summer 2026",repo:"https://github.com/carolina-moron/MVDC",demo:"https://hilarious-llama-c8e989.netlify.app",blurb:"Gathers publicly reported conflict and human-rights incidents across Malawi, classifies each against a rights-based taxonomy, traces it to a likely root cause, and matches it to the Malawian organisation best placed to respond. 258 incidents across 26 of 28 districts, with 46 response actors. The working half of the Malawi Voice Data Commons, which aims to let someone report by voice, in Chichewa, from a basic phone.",language:"TypeScript",theme:"Early warning",access:"internal"},{name:"Generative AI for Good — Avatar Storytelling",repoName:"avatar-storytelling",term:"Spring 2025",repo:"https://github.com/Ethical-Tech-CoLab/Avatar-Impact-Stories",demo:"https://ethical-tech-colab.github.io/Avatar-Impact-Stories/",blurb:"Human-condition storytelling with culturally grounded digital-human avatars — short pieces produced with generative media, now gathered into Many Voices, a browser kiosk where a scattered grid of faces plays each survivor testimony in place. The stories are written from real-world human-trafficking and child-labour cases; the avatars are built with D-ID.",language:"D-ID",theme:"Storytelling"}],"publicationTopics",0,["Artificial Intelligence","Guidelines","Evacuation","Cultural heritage","Traceability","Diplomacy","Sustainability","Disaster response"],"publications",0,{eyebrow:"Publications · Academic reports",heading:"The research, written up.",intro:"Each research question the CoLab takes on is written up as an academic report. Browse the catalogue by topic, pick a title to read what it asks and what it found, then open the report itself. Titles still in preparation are shelved alongside the published ones, so you can see where the work is going.",items:[{index:"28",area:"Foundations",topic:"Artificial Intelligence",question:"What is ethical AI, and what makes any technology, or any exercise of power, ethical at all?",title:"What Is Ethical AI? Ethics, Ethical Technology, and Ethical International Relations for the Age of Intelligent Machines",summary:"The Ethical Tech CoLab's foundational paper. It traces ethics from the earliest civilizations through international affairs and human rights law to the responsible AI movement, the humanitarian sector, and the United Nations system, and defends an institutional answer: ethical AI is not a product feature but artificial intelligence whose whole lifecycle stays accountable to ethical deliberation, a human rights floor, the do no harm obligation, and the participation of those it affects. Closes with the CoLab's motivation and research philosophy.",status:"Published",date:"July 2026",url:"/publications/what-is-ethical-ai",pdf:"https://ethical-tech-colab.github.io/what-is-ethical-ai/pdf/What-Is-Ethical-AI-ETC-Report.pdf",repo:"https://github.com/Ethical-Tech-CoLab/what-is-ethical-ai"},{index:"06",area:"Research",topic:"Artificial Intelligence",question:"How can AI help researchers formulate rigorous research questions?",title:"AI-Powered Assistance in Formulating Research Questions",summary:"A survey of how AI supports researchers across the question-formulation workflow — finding gaps, generating candidate questions, summarizing the state of the art, and flagging contradictions — with guardrails against hallucinated citations and a red-team verification model. Ships with a reusable Copilot 'Researcher' prompt and a journal-credibility rubric.",status:"Published",date:"October 2025",url:"/publications/ai-research-assistant"},{index:"19",area:"AI systems",topic:"Guidelines",question:"Which AI model should be selected for a task, and on what evidence?",title:"AI Model Performance: Capabilities, Accuracy, Speed, Energy Use, and Token Economics",summary:"A comparative review of frontier, open-weight, and efficient model families, separating independently verified measurement from provider marketing. Every figure carries its source, its date, and its evidence grade. Ships with a live cost-per-accepted-task calculator and a validated data layer.",status:"Published",date:"July 2026",url:"/publications/ai-models-research"},{index:"05",area:"Sustainability",topic:"Artificial Intelligence",question:"How large is AI's environmental footprint — and how can it be reduced?",title:"AI's Carbon Footprint: The Environmental Impact of AI",summary:"A CoLab report on AI's energy use across training and inference, the data-center and hardware toll, and the mitigation strategies, regulations, and policies that could bend the curve.",status:"Published",date:"May 2025",url:"/publications/ai-carbon-footprint"},{index:"27",area:"Evacuation",topic:"Evacuation",question:"How can AI inform evacuation decisions, and what happens after?",title:"After the Corridor: From AI-Informed Evacuation to Digital Public Goods for Refugee Economic Inclusion",summary:"The Summer 2026 cohort's synthesis report. It reads the five fielded evacuation prototypes and the forced-labour risk index together, states plainly what is validated and what is not, and extends the arc to the camp: the measurement, financial-modeling, and rights infrastructure that can shorten protracted displacement, grounded at Dzaleka Refugee Camp, Malawi. It also names the one thing the lab refuses to build, biometric identity for displaced people, and says why that refusal is itself a public good.",status:"Published",date:"July 2026",url:"/publications/after-the-corridor",pdf:"https://ethical-tech-colab.github.io/after-the-corridor-report/pdf/After-the-Corridor-ETC-Research-Report.pdf",repo:"https://github.com/Ethical-Tech-CoLab/after-the-corridor-report"},{index:"07",area:"Evacuation",topic:"Evacuation",question:"Is leaving actually safer than staying, and how would anyone know?",title:"The Evacuation Inform Index: Weighing the Risk of Leaving Against the Risk of Staying",summary:"A plain-language report on a decision support prototype that scores 104 active crises on two risks at once, keeping danger and feasibility deliberately apart so that operational difficulty cannot quietly cancel out a legal obligation. Prepared as masters research at the NYU Center for Global Affairs.",status:"Published",date:"July 2026",url:"/publications/evacuation-inform-index"},{index:"08",area:"Evacuation",topic:"Evacuation",question:"How dangerous is it to stay, and is leaving actually possible?",title:"The Civilian Evacuation Risk Anticipation Index: A Decision-Support Tool for Protecting Civilians During Evacuations in Armed Conflict",summary:"A plain-language report on a prototype that keeps two questions apart: how dangerous it is for civilians to remain, and whether an organised evacuation is possible at all. Separating them matters legally, because the duty to protect arises from danger, not from operational convenience.",status:"Published",date:"July 2026",url:"/publications/cerai"},{index:"09",area:"Evacuation",topic:"Evacuation",question:"What does it actually cost to move a besieged population, and what does staying cost?",title:"The Evacuation Risk and Cost Framework: Estimating the Human and Financial Cost of Civilian Evacuation in Armed Conflict",summary:"A report on a framework that scores seven weighted risk dimensions, estimates the cost of evacuating against the cost of staying, and finds the point at which those two curves cross. Calibrated against 31 documented historical operations.",status:"Published",date:"July 2026",url:"/publications/ercf"},{index:"10",area:"Evacuation",topic:"Evacuation",question:"How much does poor field information degrade an evacuation decision?",title:"The Evacuation Readiness and Uncertainty Simulator: Showing How Poor Field Information Degrades Evacuation Decisions",summary:"A report on a simulator that models what happens when planners act on incomplete intelligence. All data is synthetic and reproducible from a seed, so the mechanism rather than any particular place is what is on display.",status:"Published",date:"July 2026",url:"/publications/erus"},{index:"11",area:"Evacuation",topic:"Evacuation",question:"How do information and demographics decide who actually leaves?",title:"The Evacuation Simulator: An Agent-Based Model of How Civilians Decide to Leave During Armed Conflict",summary:"A report on an agent-based model of household evacuation behaviour, tracing how warnings spread through a community and how age, mobility, and trust shape who reaches safety and who does not.",status:"Published",date:"July 2026",url:"/publications/evacuation-simulation"},{index:"12",area:"Evacuation",topic:"Evacuation",question:"Could a daily severity measure have shown the danger in Mariupol as it unfolded?",title:"The Mariupol Corridor Severity Model: A Daily Measure of Civilian Danger During the Siege of Mariupol, March to May 2022",summary:"A retrospective report scoring 77 days of the siege on a daily severity measure built from open sources. Conduct descriptions are attributed to their sources rather than adjudicated, and contested figures are left contested.",status:"Published",date:"July 2026",url:"/publications/mariupol-severity-model"},{index:"13",area:"Cultural heritage",topic:"Cultural heritage",question:"Can an artwork's ownership history be checked automatically, and how far should that be trusted?",title:"The Digital Provenance Passport: An Automated Assistant for Tracing the Ownership History of Artworks and Cultural Objects",summary:"A report on an assistant that assembles a provenance timeline, flags looting and valuation risk, and issues a signed record. Candid about how much of the demonstration runs on fixtures rather than live sources.",status:"Published",date:"July 2026",url:"/publications/digital-provenance-passport"},{index:"14",area:"Cultural heritage",topic:"Cultural heritage",question:"What can be established about an object's history from free public sources alone?",title:"Provenance Search: An Automated Ownership-History Check for Artworks and Cultural Objects",summary:"A report on a tool that queries seven public sources and scores confidence with a fixed, inspectable algorithm, deliberately flagging what cannot be verified rather than filling the gaps.",status:"Published",date:"July 2026",url:"/publications/provenance-search"},{index:"15",area:"Cultural heritage",topic:"Cultural heritage",question:"What is worth recording about a visit to a work of art?",title:"VANGO: A Digital Passport for Recording Visits to Works of Art",summary:"A report on a digital souvenir passport that records attendance at art experiences. It performs no provenance or authenticity checking, and the report is explicit that nothing it records is evidence of ownership or lawful origin.",status:"Published",date:"July 2026",url:"/publications/vango"},{index:"16",area:"Traceability",topic:"Traceability",question:"Where are the conditions that make forced labour likely, before any case is detected?",title:"The Forced Labor Structural Risk Index: A Country-Level Measure of the Conditions Under Which Forced Labour Becomes More Likely",summary:"A report on an index scoring 184 countries across 11 domains and 43 indicators, measuring the conditions that enable forced labour rather than counting cases. Ships with uncertainty bands and is meant to be read in tiers, not as a league table.",status:"Published",date:"December 2025",url:"/publications/forced-labor-structural-risk-index"},{index:"17",area:"Diplomacy",topic:"Diplomacy",question:"Can practitioners rehearse a multi-party negotiation against AI agents?",title:"The Diplomatic Simulator: A Multi-Party Negotiation Simulator Driven by Artificial Intelligence Agents",summary:"A report on a simulator in which AI delegations negotiate a live crisis, producing a structured record of positions, concessions, and tactics. Across every Monte Carlo trial run, no comprehensive settlement was reached.",status:"Published",date:"July 2026",url:"/publications/diplomatic-simulator"},{index:"29",area:"Diplomacy",topic:"Diplomacy",question:"What happens when the machine in a nuclear thriller is a real language model?",title:"The Only Winning Move: Rebuilding WarGames (1983) as a Playable Study of Autonomous AI Agents",summary:"A report on rebuilding the 1983 film as a browser game whose antagonist is a real language model held to a strict output contract, and on what four tracks of Monte Carlo evaluation found. Structured output turned out to be solved; the real problem was that models will not escalate, stalling a quarter of games. On owned hardware the smallest model beat every cloud model on speed and the largest returned no valid output at all.",status:"Published",date:"July 2026",url:"/publications/war-games",repo:"https://github.com/Ethical-Tech-CoLab/War-Games"},{index:"18",area:"Disaster response",topic:"Disaster response",question:"How fast can building damage be assessed after a disaster, and with how few labels?",title:"HASTE: High-speed Assessment and Satellite Tracking for Emergencies",summary:"A plain-language report on a satellite damage-assessment platform built by the Microsoft AI for Good Lab, used in 31 field deployments. The software is Microsoft's; only this report is the CoLab's.",status:"Published",date:"July 2026",url:"/publications/haste"},{index:"20",area:"Guidelines",topic:"Guidelines",question:"How does someone with no Git background start contributing?",title:"GitHub 101: Getting Started with Git and GitHub",summary:"A beginner's reference for Git and GitHub from day one — what the two actually are and how they differ, the key concepts, setting up, and the handful of commands that cover almost everything you will do in the first months.",status:"Internal guide",date:"July 2026",access:"internal",url:"https://github.com/Ethical-Tech-CoLab/Getting-Going/blob/main/docs/github-101.md"},{index:"21",area:"Guidelines",topic:"Guidelines",question:"How does a team actually collaborate in a shared repository?",title:"GitHub Workflow: Branching, Pull Requests and Code Review",summary:"The collaborative loop, step by step: the GitHub flow at a glance, creating and managing branches, opening a pull request that can be reviewed, and what to look for when reviewing someone else's.",status:"Internal guide",date:"July 2026",access:"internal",url:"https://github.com/Ethical-Tech-CoLab/Getting-Going/blob/main/docs/github-workflow.md"},{index:"22",area:"Guidelines",topic:"Guidelines",question:"How do you put a project on the web without paying for hosting?",title:"GitHub Pages 101",summary:"Publishing a free site straight from a repository, written for first-time users — no server to rent, nothing to install. Includes an honest account of what GitHub Pages can and cannot do, which is where most first attempts come unstuck.",status:"Internal guide",date:"July 2026",access:"internal",url:"https://github.com/Ethical-Tech-CoLab/Getting-Going/blob/main/docs/github-pages-101.md"},{index:"23",area:"Guidelines",topic:"Guidelines",question:"How do you build and test a data product before you have real data?",title:"Creating Synthetic Data and Seeding Models: A Beginner's Guide",summary:"Generating data algorithmically rather than collecting it: why you would (developing before real data exists, forcing edge cases, sharing something realistic without privacy exposure, reproducing a dataset from a seed), and how to do it without fooling yourself about what the results mean.",status:"Internal guide",date:"July 2026",access:"internal",url:"https://github.com/Ethical-Tech-CoLab/Getting-Going/blob/main/synthetic-data/SYNTHETIC_DATA_GUIDE.md"},{index:"24",area:"Guidelines",topic:"Guidelines",question:"How do you teach synthetic-data generation to a room of beginners?",title:"Synthetic Data: Training Deck",summary:"The companion training material to the synthetic-data guide — a self-contained walkthrough for running the topic as a session rather than reading it alone.",status:"Internal guide",date:"July 2026",access:"internal",url:"https://github.com/Ethical-Tech-CoLab/Getting-Going/blob/main/synthetic-data/synthetic-data-training.html"},{index:"25",area:"Guidelines",topic:"Guidelines",question:"How can an agent do live web research from any runtime, including locked-down ones?",title:"tavily-research: A Portable Agent Skill for Live Web Research",summary:"A cross-agent skill wrapping the Tavily REST API — search, extract, crawl and map — with a dependency-free client in both Python and Node. It calls the API directly, so it still works where a Tavily MCP server cannot be launched, such as under IT policy that blocks local npx.",status:"Internal guide",date:"July 2026",access:"internal",url:"https://github.com/Ethical-Tech-CoLab/Getting-Going/tree/main/tavily-research"},{index:"26",area:"Guidelines",topic:"Guidelines",question:"How do you get a terminal on the CoLab's GPU machine?",title:"Getting Started on Cloud GPU: SSH Access to the B3IQ Machine",summary:"One-time setup, about five minutes: installing cloudflared for the secure tunnel, creating an SSH key if you do not have one, and connecting from macOS, Windows or Linux.",status:"Internal guide",date:"July 2026",access:"internal",url:"https://github.com/Ethical-Tech-CoLab/Getting-started-on-CloudGPU-/blob/main/remote-gpu-how-to-guide.md"}]},"researchAreas",0,[{index:"01",key:"Evacuation",question:"How can AI inform evacuation decisions?",tags:["Disaster response","Civic tech","Information equity"],summary:"Turning fragmented crisis signals into decisions, so responders and residents can act before information gaps cost lives.",detail:"Projects under this question measure and improve the quality, accessibility, and equity of the information people rely on when they have to leave.",stack:["Geospatial analytics","LLM signal extraction","Open data pipelines","React dashboards"],projects:[{name:"Evacuation Inform Index",summary:"Humanitarian practice tends to treat evacuation as the obvious good, but the journey carries its own risk, and a crisis can be both catastrophic to endure and impossible to escape. This index scores 104 active crises on those two risks separately, drawing on INFORM Severity data, a twelve-factor vulnerability profile, and live conflict reporting, so that the comparison between leaving and staying becomes visible rather than assumed.",status:"Active",repo:"https://github.com/Ethical-Tech-CoLab/evacuation-inform-index-carolina",publication:"/publications/evacuation-inform-index"},{name:"Exodus — Civilian Evacuation Risk Platform",summary:"A platform unifying three civilian-evacuation risk tools: a crisis map of the INFORM Severity Index with live news and conflict timelines, a seven-dimension scenario risk model, and an endangerment-and-feasibility risk assessment, sharing one backend and design language.",status:"Active",repo:"https://github.com/Ethical-Tech-CoLab/Exodus"},{name:"ERCF — Evacuation Risk & Cost Framework",summary:"A decision-support tool that estimates the human and financial cost of civilian evacuation in armed conflict, scoring scenarios across seven risk dimensions and comparing the cost of evacuating against the cost of staying in the zone.",status:"Active",repo:"https://github.com/Ethical-Tech-CoLab/ercf",publication:"/publications/ercf"},{name:"Evacuation Behavior Simulator",summary:"An interactive, agent-based model of community evacuation behavior, modeling family clusters, information-seeking, neighbor social influence, and elder and child delays across an UNAWARE → SEEKING → MILLING → EVACUATING → DONE lifecycle.",status:"Active",repo:"https://github.com/Ethical-Tech-CoLab/Evac-Sim-Melanie",publication:"/publications/evacuation-simulation"},{name:"Evacuation Routing Simulator",summary:"An interactive evacuation simulator that models how routing and assignment choices shape who reaches safety, with shareable scenario state and a built-in explainer walkthrough.",status:"Active",repo:"https://github.com/Ethical-Tech-CoLab/India-EvacSimulation",publication:"/publications/erus"},{name:"Mariupol 2022 — Corridor Severity Model",summary:"A corridor-severity model reconstructing the 2022 Mariupol evacuation, scoring the risk and viability of humanitarian corridors under siege conditions.",status:"Active",repo:"https://github.com/Ethical-Tech-CoLab/mariupol-evacuation-model",demo:"https://ethical-tech-colab.github.io/mariupol-evacuation-model/",publication:"/publications/mariupol-severity-model"}]},{index:"02",key:"Cultural heritage",question:"How can technology support the ethical return of cultural artifacts?",tags:["Cultural heritage","Provenance","Restorative justice"],summary:"Giving institutions, source communities, and researchers a shared, verifiable record of an artifact's life across borders.",detail:"Projects under this question pair verifiable credentials with collaborative cataloguing to support repatriation, attribution, and ethical stewardship.",stack:["Verifiable credentials","IIIF imaging","Ledger anchoring","Knowledge graphs"],projects:[{name:"Arts Provenance Agent — Digital Provenance Passport",summary:"An x402-native agent that traces the provenance of artworks and artifacts, flags looting, repatriation, and valuation risk, and issues a cryptographically signed, tamper-evident Passport (a JSON-LD Verifiable Credential) for each object. Every claim is grounded in an allowlist of authoritative sources — the Met, UNESCO, ICOM, the Art Loss Register — so a fact without a citation is never produced. Includes a dashboard for tracing where an object has been, its risk score, and repatriation status.",status:"Active",repo:"https://github.com/Ethical-Tech-CoLab/arts-provenance-agent",demo:"https://ethical-tech-colab.github.io/arts-provenance-agent/",publication:"/publications/digital-provenance-passport"},{name:"Provenance Search — Provenance Intelligence",summary:"Looks up an artwork's ownership history across free public sources — Tavily search (restricted to the Met, Getty, INTERPOL, UNESCO, Art Loss Register and more), the Met and Art Institute of Chicago APIs, MoMA, Wikidata, and Europeana — and emits a provenance passport with a confidence score computed by a fixed algorithm, not the AI. Supports text, image upload, and live camera capture for use inside a museum.",status:"In development",repo:"https://github.com/Ethical-Tech-CoLab/provenance-search",publication:"/publications/provenance-search"},{name:"Digital Art Passport (VANGO)",summary:"A mobile-style digital passport for art experiences: scan a QR code or enter an artwork's code to collect a stamp — a vintage illustration of the piece with the artist, venue, and date. Stamps persist across sessions, with a collector bio and passport number, and a fully multilingual interface (English, French, Italian).",status:"Active",repo:"https://github.com/Ethical-Tech-CoLab/VANGO",demo:"https://ethical-tech-colab.github.io/VANGO/",publication:"/publications/vango"},{name:"3D Online Gallery",summary:"An immersive 3D gallery for exploring provenance-tracked and repatriated works in a navigable virtual space. Placeholder for an upcoming build.",status:"Coming soon"}]},{index:"03",key:"Traceability",question:"How can ethical claims in supply chains be made verifiable?",tags:["Labor rights","Climate accountability","Trust infrastructure"],summary:"Auditable traceability for materials, labor, and impact, from origin to shelf.",detail:"Projects under this question build interoperable trace records that travel with goods, letting buyers and regulators verify ethical claims without exposing sensitive supplier data.",stack:["Zero-knowledge proofs","ERP integrations","Standards (GS1, EPCIS)","Edge attestation"],projects:[{name:"Ethical Supply Chain & Traceability",summary:"Interoperable trace records that travel with goods, letting buyers and regulators verify ethical claims about materials, labor, and impact without exposing sensitive supplier data.",status:"Active"},{name:"Forced Labor Structural Risk Index",summary:"An interactive index that scores structural forced-labor risk across regions and sectors, turning scattered supply-chain and labor signals into a comparable measure buyers and regulators can act on. Built by Amanda Lindsey in the Fall 2025 cohort.",status:"Active",demo:"https://ethical-tech-colab.github.io/forced-labor-structural-risk-index/",publication:"/publications/forced-labor-structural-risk-index"}]},{index:"04",key:"Diplomacy",question:"How can AI help practitioners rehearse high-stakes diplomacy?",tags:["Diplomacy","AI safety","Pedagogy"],summary:"An AI-mediated environment where negotiators rehearse high-stakes diplomacy with culturally grounded agents.",detail:"Projects under this question pair domain experts with adaptive agents that model historical context, incentives, and red lines, training practitioners for situations textbooks cannot capture.",stack:["Multi-agent LLMs","Scenario authoring tools","Evaluation harnesses","Voice interfaces"],projects:[{name:"Diplomatic Simulator",summary:"A multi-party AI negotiation table where each delegation argues from its own confidential 'privileged instructions.' Five crises — Arctic, Central Asia (Fergana Valley), Cyprus, South China Sea, and Iran–US Strait of Hormuz — each with a live-negotiation replay, an interactive analysis dashboard (parties' strategy, proposals, detector hits), and Monte Carlo runs over randomized conditions.",status:"Active",repo:"https://github.com/Ethical-Tech-CoLab/diplomatic-simulator",demo:"https://ethical-tech-colab.github.io/diplomatic-simulator/",publication:"/publications/diplomatic-simulator"},{name:"War Games",summary:"A terminal thriller in the shape of WarGames (1983), reframed around a modern AI agent: the only winning move is to understand the machine. Ships with a Monte Carlo simulation harness and a written case study behind the fiction.",status:"Active",repo:"https://github.com/Ethical-Tech-CoLab/War-Games",demo:"https://ethical-tech-colab.github.io/War-Games/",publication:"/publications/war-games"}]}],"site",0,{name:"Ethical Tech CoLab",tagline:"Emerging tech, human condition.",linkedin:"https://www.linkedin.com/company/ethical-tech-lab/",email:"ethical-tech-colab@nyu.edu",partnersLine:"NYU SPS · CGA · Microsoft · New York",footerBlurb:"Exploring intervention opportunities at the intersection of emerging technologies and the human condition.",cohortRange:"Four cohorts · est. 2024-2026",social:{github:"https://github.com/Ethical-Tech-CoLab",instagram:"",twitter:""},legal:["The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.","Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners."]}])}]); \ No newline at end of file diff --git a/static-site/_next/static/chunks/1l1xxdmwck4_f.js b/static-site/_next/static/chunks/1l1xxdmwck4_f.js new file mode 100644 index 000000000..50791151e --- /dev/null +++ b/static-site/_next/static/chunks/1l1xxdmwck4_f.js @@ -0,0 +1 @@ +(globalThis.TURBOPACK||(globalThis.TURBOPACK=[])).push(["object"==typeof document?document.currentScript:void 0,89042,e=>{"use strict";let a=["NYU School of Professional Studies","Center for Global Affairs (CGA)","Microsoft","Generative AI for Good","The Microsoft Garage","Tavily","Tradeverifyd","Sayari","B3IQ","Art & Antiquities Blockchain Consortium","100x","D_ID","Apne Aap Women Worldwide","Gaia","OSCE / ODIHR Anti-Trafficking","Coinbase","Rivr","x402 Foundation","Blockchain for Social Impact"];[{name:"NYU School of Professional Studies",about:"The NYU school that houses the Center for Global Affairs and the Ethical Tech CoLab.",url:"https://www.sps.nyu.edu",logo:"/logos/nyu-sps.jpg"},{name:"Center for Global Affairs (CGA)",about:"NYU SPS's Center for Global Affairs — the CoLab's academic home, focused on global policy and the human condition.",url:"https://www.sps.nyu.edu/homepage/academics/divisions-and-departments/center-for-global-affairs.html",logo:"/nyu-sps-cga-logo.jpg"},{name:"Microsoft",about:"Technology partner across proof-of-human, AI, web3, and privacy-preserving systems such as zero-knowledge proofs.",url:"https://www.microsoft.com",logo:"/logos/microsoft.svg"},{name:"Generative AI for Good",about:"Uses generative media for human-condition storytelling — making lived experience legible to audiences and decision-makers.",url:"",logo:"/logos/generative-ai-for-good.jpg"},{name:"D_ID",about:"Avatars and digital-human tooling for synthetic media, used to prototype human-condition storytelling.",url:"https://www.d-id.com",logo:"/logos/d-id.jpg"},{name:"Tradeverifyd",about:"Ask your supply chain anything — agentic AI that resolves suppliers across 200+ global data sources, maps ownership through tier three, and monitors exposure to forced labor and sanctions regimes. Founded as Mesur.io in 2016; rebranded in 2025.",url:"https://tradeverifyd.com",logo:"/logos/tradeverifyd.svg"},{name:"Rivr",about:"Twitch livestreams, keyword tags, and chat-room activity as a real-time data source for detecting at-risk behavior online.",url:"",logo:"/logos/rivr.jpg"},{name:"100x",about:"",url:"",logo:"/logos/100x.jpg"},{name:"Apne Aap Women Worldwide",about:"",url:"https://apneaap.org",logo:"/logos/apne-aap-women-worldwide.png"},{name:"Gaia",about:"",url:"",logo:"/logos/gaia.png"},{name:"OSCE / ODIHR Anti-Trafficking",about:"Multilateral partner on anti-trafficking, platform safety, and human rights — a recurring voice at the Ethical Tech Summit.",url:"https://www.osce.org/odihr",logo:""},{name:"Coinbase",about:"Industry partner on agentic commerce and micropayment rails, engaged through the Ethical Tech Summit.",url:"https://www.coinbase.com",logo:"/logos/coinbase.png"},{name:"x402 Foundation",about:"Stewards the x402 payment standard behind the CoLab's agentic-commerce and micropayment work, engaged through the Ethical Tech Summit.",url:"https://www.x402.org",logo:"/logos/x402-foundation.jpg"},{name:"Sayari",about:"Maps global corporate ownership and supply-chain relationships across more than 250 jurisdictions, surfacing hidden counterparty and compliance risk.",url:"https://sayari.com",logo:"/logos/sayari.jpg"},{name:"B3IQ",about:"US-based AI infrastructure: NVIDIA-powered GPU servers that buyers own, racked and run in a data center in Eugene, Oregon. B3IQ hosts the CoLab's GPU machine, which the cohort reaches over SSH for model work.",url:"https://b3iq.org",logo:"/logos/b3iq.png",logoTile:"dark"},{name:"Art & Antiquities Blockchain Consortium",about:"A consortium working on cultural-heritage ownership and repatriation, bringing stakeholders, technologies, and cultural artifacts together.",url:"https://aabconsortium.org",logo:"/logos/art-antiquities-blockchain-consortium.png"},{name:"Blockchain for Social Impact",about:"A not-for-profit coalition that incubates and develops blockchain solutions addressing environmental and social challenges, convening nonprofits, governments, investors, and technologists.",url:"https://blockchainforsocialimpact.com",logo:"/logos/blockchain-for-social-impact.png"},{name:"The Microsoft Garage",about:"Microsoft's worldwide innovation program — local Garage sites, the Global Hackathon, and the Innovation Studio platform — where employees and customers turn experimental ideas into working projects.",url:"https://www.microsoft.com/en-us/garage/",logo:"/logos/microsoft-garage.png"},{name:"Tavily",about:"A web search and extraction API built for AI agents rather than people, returning clean extracted text with the source address of every result. The CoLab uses it as the restricted research engine behind the provenance work and the tavily-research agent skill.",url:"https://tavily.com",logo:"/logos/tavily.svg"}].sort((e,t)=>{let i=a.indexOf(e.name),o=a.indexOf(t.name);return(-1===i?Number.MAX_SAFE_INTEGER:i)-(-1===o?Number.MAX_SAFE_INTEGER:o)}),e.s(["areaSlug",0,e=>e.toLowerCase().replace(/\s+/g,"-"),"nav",0,[{label:"Home",href:"/"},{label:"Portfolio",href:"/portfolio",children:[{label:"Overview",href:"/portfolio"},{label:"Live Demos",href:"/demos"},{label:"Publications",href:"/publications"},{label:"Media",href:"/media"},{label:"Newsletter",href:"/newsletter"}]},{label:"Team",href:"/team"}],"productTerms",0,["Summer 2026","Fall 2025","Spring 2025"],"productThemes",0,["Evacuation","Cultural heritage","Traceability","Early warning","Diplomacy","Research","Storytelling"],"products",0,[{name:"Agentic Language Development",repoName:"agentic-language-development",term:"Summer 2026",repo:"https://github.com/Ethical-Tech-CoLab/agentic-language-development",demo:"https://ethical-tech-colab.github.io/agentic-language-development/",demos:[{label:"Open the project site",href:"https://ethical-tech-colab.github.io/agentic-language-development/"},{label:"The concept in full",href:"https://github.com/Ethical-Tech-CoLab/agentic-language-development/blob/main/CONCEPT-IDEA.md"},{label:"Ledger integrity design",href:"https://github.com/Ethical-Tech-CoLab/agentic-language-development/blob/main/LEDGER-INTEGRITY-DESIGN.md"},{label:"Experiment notebook",href:"https://github.com/Ethical-Tech-CoLab/agentic-language-development/blob/main/EXPERIMENT-NOTEBOOK.md"}],blurb:"Two isolated agents, one channel that carries no human language, and shared tasks neither can finish alone. Each keeps its own chronological ledger of what it believes a symbol means, hash-chained and anchored so an outsider can prove nothing was edited afterwards. Concept phase: nineteen experiments are pre-registered and none has been run, which the site says on its own front page.",language:"HTML",theme:"Research",publication:"/publications/agentic-language-development"},{name:"Agentic Behavior Observatory",repoName:"agentic-behavior-observatory",term:"Summer 2026",repo:"https://github.com/Ethical-Tech-CoLab/agentic-behavior-observatory",demo:"https://ethical-tech-colab.github.io/agentic-behavior-observatory/",blurb:"Paste a GitHub repository URL and watch it get read: five axes of agentic behaviour scored from the code itself, every point linked to the file and line where its signal fired. The analysis runs in your browser against the public GitHub API, so nothing is uploaded and no token is needed. The population lab shows which demographic dimensions the analysed corpus models, and which nobody models at all.",language:"Python",theme:"Research"},{name:"Cyber Dictionary and Library",repoName:"cyber-dictionary",term:"Summer 2026",repo:"https://github.com/Ethical-Tech-CoLab/cyber-dictionary",demo:"https://ethical-tech-colab.github.io/cyber-dictionary/",blurb:"Two rooms in one site. A dictionary of 423 technology and cybersecurity terms across 11 domains, each defined in a sentence or two of plain English, and a library of 105 open data sources and open-source technologies across 11 shelves, each with what it gives you, how to reach its API, and what it costs.",language:"JavaScript",theme:"Research"},{name:"Diplomatic Simulator",repoName:"diplomatic-simulator",term:"Summer 2026",repo:"https://github.com/Ethical-Tech-CoLab/diplomatic-simulator",demo:"https://ethical-tech-colab.github.io/diplomatic-simulator/live.html",demos:[{label:"Watch a live negotiation",href:"https://ethical-tech-colab.github.io/diplomatic-simulator/live.html"},{label:"Interactive dashboard",href:"https://ethical-tech-colab.github.io/diplomatic-simulator/dashboard.html"},{label:"Monte Carlo analysis",href:"https://ethical-tech-colab.github.io/diplomatic-simulator/montecarlo.html"},{label:"Methodology & AI use",href:"https://ethical-tech-colab.github.io/diplomatic-simulator/methodology.html"}],blurb:"A multi-party AI negotiation table where each delegation argues from its own confidential 'privileged instructions.' Five crises — Arctic, Central Asia, Cyprus, South China Sea, and Iran–US Strait of Hormuz — each with a live-negotiation replay, an interactive analysis dashboard (parties' strategy, proposals, detector hits), and Monte Carlo runs over randomized conditions.",language:"HTML",theme:"Diplomacy",featured:!0,publication:"/publications/diplomatic-simulator"},{name:"AI Models Research",repoName:"AI-Models-Research",term:"Summer 2026",repo:"https://github.com/Ethical-Tech-CoLab/AI-Models-Research",demo:"https://ethical-tech-colab.github.io/AI-Models-Research/",demos:[{label:"Read the handbook",href:"https://ethical-tech-colab.github.io/AI-Models-Research/"},{label:"Cost-per-accepted-task calculator",href:"https://ethical-tech-colab.github.io/AI-Models-Research/interactive/"}],blurb:"A comparative review of frontier, open-weight, and efficient model families across benchmark performance, factual accuracy, latency, token economics, and energy use. Three evidence grades separate independently verified measurement from institutional research and from provider marketing, and every figure names the party that produced it. Ships with a live cost-per-accepted-task calculator over a validated data layer.",language:"Python",theme:"Research",featured:!0,publication:"/publications/ai-models-research"},{name:"Exodus — Civilian Evacuation Risk Platform",repoName:"Exodus",term:"Summer 2026",repo:"https://github.com/Ethical-Tech-CoLab/Exodus",demo:"https://exodus-ruddy.vercel.app",blurb:"Three civilian-evacuation risk tools behind one backend and design language: a crisis map of the INFORM Severity Index with live news and conflict timelines, a seven-dimension scenario risk model, and an endangerment-and-feasibility assessment.",language:"TypeScript",theme:"Evacuation",featured:!0},{name:"Evacuation Inform Index",repoName:"evacuation-inform-index-carolina",term:"Summer 2026",repo:"https://github.com/Ethical-Tech-CoLab/evacuation-inform-index-carolina",demo:"https://ethical-tech-colab.github.io/evacuation-inform-index-carolina/",blurb:"A composite index scoring 104 active crises on the risk of staying and the risk of leaving, side by side, so the comparison between them can be argued with. Built on INFORM Severity data with live news and an ACLED conflict timeline, filterable by crisis type, and with every one of its thirteen legal citations explained — including how far each one is actually justified.",language:"HTML",theme:"Evacuation",featured:!0,publication:"/publications/evacuation-inform-index"},{name:"Evacuation Routing Simulator",repoName:"India-EvacSimulation",term:"Summer 2026",repo:"https://github.com/Ethical-Tech-CoLab/India-EvacSimulation",demo:"https://ethical-tech-colab.github.io/India-EvacSimulation/",blurb:"An interactive simulator modeling how routing and assignment choices shape who reaches safety, with shareable scenario state and a built-in explainer walkthrough.",language:"HTML",theme:"Evacuation",publication:"/publications/erus"},{name:"Mariupol 2022 — Corridor Severity Model",repoName:"mariupol-evacuation-model",term:"Summer 2026",repo:"https://github.com/Ethical-Tech-CoLab/mariupol-evacuation-model",demo:"https://ethical-tech-colab.github.io/mariupol-evacuation-model/",blurb:"A corridor-severity model reconstructing the 2022 Mariupol evacuation, scoring the risk and viability of humanitarian corridors under siege conditions.",language:"HTML",theme:"Evacuation",publication:"/publications/mariupol-severity-model"},{name:"Evacuation Behavior Simulator",repoName:"Evac-Sim-Melanie",term:"Summer 2026",repo:"https://github.com/Ethical-Tech-CoLab/Evac-Sim-Melanie",demo:"https://ethical-tech-colab.github.io/Evac-Sim-Melanie/",blurb:"An agent-based model of community evacuation behavior — family clusters, information-seeking, neighbor influence, and elder and child delays across an UNAWARE → SEEKING → MILLING → EVACUATING → DONE lifecycle.",language:"JavaScript",theme:"Evacuation",publication:"/publications/evacuation-simulation"},{name:"Digital Art Passport (VANGO)",repoName:"VANGO",term:"Summer 2026",repo:"https://github.com/Ethical-Tech-CoLab/VANGO",demo:"https://ethical-tech-colab.github.io/VANGO/",blurb:"A digital provenance passport for cultural artifacts, pairing verifiable credentials with collaborative cataloguing to support repatriation, attribution, and ethical stewardship.",language:"HTML",theme:"Cultural heritage",publication:"/publications/vango"},{name:"ERCF — Evacuation Risk & Cost Framework",repoName:"ercf",term:"Summer 2026",repo:"https://github.com/Ethical-Tech-CoLab/ercf",demo:"https://ercf-production.up.railway.app",blurb:"A decision-support framework estimating the human and financial cost of civilian evacuation in armed conflict, scoring scenarios across seven risk dimensions and comparing evacuating against staying.",language:"Python",theme:"Evacuation",publication:"/publications/ercf"},{name:"CERAI — Civilian Evacuation Risk Anticipation Index",repoName:"CERAI_AR",term:"Summer 2026",repo:"https://github.com/Ethical-Tech-CoLab/CERAI_AR",demo:"https://ethical-tech-colab.github.io/CERAI_AR/",blurb:"A composite index anticipating civilian risk during evacuation operations, scoring weighted categories — hazards, infrastructure, population vulnerability, information environment, and weather — with Monte Carlo sensitivity analysis, ACLED conflict-event lookup, and source-credibility tagging on every indicator, so evacuation-risk reasoning stays auditable for IHL compliance review. Built by Alana Robertson in the Summer 2026 cohort, on Teresa Cantero's doctoral research into AI and the protection and evacuation of civilians under IHL.",language:"HTML",theme:"Evacuation",publication:"/publications/cerai"},{name:"Arts Provenance Agent",repoName:"arts-provenance-agent",term:"Summer 2026",repo:"https://github.com/Ethical-Tech-CoLab/arts-provenance-agent",demo:"https://ethical-tech-colab.github.io/arts-provenance-agent/",blurb:"An x402-native agent that traces artwork provenance, flags looting and repatriation risk, and issues a signed, tamper-evident Passport for each object. This live demo runs on mock data — catalog dashboards, passport issue/verify, and a replayed agent trace.",language:"TypeScript",theme:"Cultural heritage",featured:!0,publication:"/publications/digital-provenance-passport"},{name:"Provenance Search — Arts & Artifacts",repoName:"provenance-search",term:"Summer 2026",repo:"https://github.com/Ethical-Tech-CoLab/provenance-search",demo:"https://provenance-search-production.up.railway.app",blurb:"Traces the ownership chain of an artwork across museum collections, cultural archives and loss registries, then issues a provenance passport carrying a rule-based confidence score you can inspect rather than a verdict you have to trust. Seven sources checked, three ways to search, no sign-up.",language:"HTML",theme:"Cultural heritage",publication:"/publications/provenance-search"},{name:"Digital Passport for Artworks",repoName:"digital-passport-artworks",term:"Summer 2026",repo:"https://github.com/Ethical-Tech-CoLab/digital-passport-artworks",demo:"https://ethical-tech-colab.github.io/digital-passport-artworks/",blurb:"Upload one photograph and watch the whole chain run in order: fingerprint the image, search for a prior record, score it for forgery risk, and only once it clears — or a reviewer overrides — sign and issue the passport. Real ECDSA cryptography, entirely in the browser. Originally by Christine Lumen.",language:"HTML",theme:"Cultural heritage"},{name:"Forced Labor Structural Risk Index",repoName:"forced-labor-structural-risk-index",term:"Fall 2025",repo:"https://github.com/Ethical-Tech-CoLab/forced-labor-structural-risk-index",demo:"https://ethical-tech-colab.github.io/forced-labor-structural-risk-index/",blurb:"An interactive index mapping the structural conditions that enable forced labor across 184 countries on a 0–1 risk scale — with a choropleth world map, national and sub-national layers, rankings, and sources. Built by Amanda Lindsey in the Fall 2025 cohort.",language:"HTML",theme:"Traceability",featured:!0,publication:"/publications/forced-labor-structural-risk-index"},{name:"AI Research Question Assistant",repoName:"ai-research-question-assistant",term:"Fall 2025",repo:"https://github.com/Ethical-Tech-CoLab/ai-research-question-assistant",demo:"https://ethical-tech-colab.github.io/ai-research-question-assistant/",blurb:"A survey and toolkit for how AI helps researchers move from a broad interest to a well-formed research question — finding gaps, generating candidate questions, summarizing the state of the art, and flagging contradictions, with guardrails against hallucinated citations. Ships with a reusable Copilot 'Researcher' prompt and a journal-credibility rubric. Built by the Fall 2025 cohort.",language:"HTML",theme:"Research",publication:"/publications/ai-research-assistant"},{name:"AI's Carbon Footprint",repoName:"ai-carbon-footprint",term:"Spring 2025",repo:"https://github.com/Ethical-Tech-CoLab/ai-carbon-footprint",demo:"https://ethical-tech-colab.github.io/ai-carbon-footprint/",blurb:"A report on what artificial intelligence costs the environment: energy drawn across training and inference, the data-center and hardware toll behind it, and the mitigations, regulations and policies that could bend the curve. 1,287 MWh went into training GPT-3 alone — roughly what 130 US households use in a year.",language:"HTML",theme:"Research",publication:"/publications/ai-carbon-footprint"},{name:"GitHub 101",repoName:"github-101",term:"Summer 2026",access:"internal",repo:"https://github.com/Ethical-Tech-CoLab/Getting-Going/blob/main/docs/github-101.md",blurb:"A beginner's reference for Git and GitHub from day one: what the two actually are and how they differ, the key concepts, setting up, and the handful of commands that cover almost everything you will do in the first months.",language:"Markdown",theme:"Research"},{name:"GitHub Workflow",repoName:"github-workflow",term:"Summer 2026",access:"internal",repo:"https://github.com/Ethical-Tech-CoLab/Getting-Going/blob/main/docs/github-workflow.md",blurb:"The collaborative loop, step by step: the GitHub flow at a glance, creating and managing branches, opening a pull request that can actually be reviewed, and what to look for when reviewing someone else's.",language:"Markdown",theme:"Research"},{name:"GitHub Pages 101",repoName:"github-pages-101",term:"Summer 2026",access:"internal",repo:"https://github.com/Ethical-Tech-CoLab/Getting-Going/blob/main/docs/github-pages-101.md",blurb:"Publishing a free site straight from a repository, for first-time users — no server to rent, nothing to install. Includes an honest account of what GitHub Pages can and cannot do, which is where most first attempts come unstuck.",language:"Markdown",theme:"Research"},{name:"Synthetic Data Guide",repoName:"synthetic-data-guide",term:"Summer 2026",access:"internal",repo:"https://github.com/Ethical-Tech-CoLab/Getting-Going/blob/main/synthetic-data/SYNTHETIC_DATA_GUIDE.md",blurb:"Generating data algorithmically rather than collecting it: why you would — developing before real data exists, forcing edge cases, sharing something realistic without privacy exposure, reproducing a dataset from a seed — and how to do it without fooling yourself about what the results mean.",language:"Markdown",theme:"Research"},{name:"Synthetic Data — Training Deck",repoName:"synthetic-data-training",term:"Summer 2026",access:"internal",repo:"https://github.com/Ethical-Tech-CoLab/Getting-Going/blob/main/synthetic-data/synthetic-data-training.html",blurb:"The companion training material to the synthetic-data guide — a self-contained walkthrough for running the topic as a session rather than reading it alone.",language:"HTML",theme:"Research"},{name:"tavily-research",repoName:"tavily-research",term:"Summer 2026",access:"internal",repo:"https://github.com/Ethical-Tech-CoLab/Getting-Going/tree/main/tavily-research",blurb:"A cross-agent skill wrapping the Tavily REST API — search, extract, crawl and map — with a dependency-free client in both Python and Node. It calls the API directly, so it still works where a Tavily MCP server cannot be launched, such as under IT policy that blocks local npx.",language:"Python · Node",theme:"Research"},{name:"Getting Started on Cloud GPU",repoName:"cloud-gpu-guide",term:"Summer 2026",access:"internal",repo:"https://github.com/Ethical-Tech-CoLab/Getting-started-on-CloudGPU-/blob/main/remote-gpu-how-to-guide.md",blurb:"Getting a terminal on the CoLab's B3IQ machine over SSH. One-time setup, about five minutes: installing cloudflared for the secure tunnel, creating an SSH key if you do not have one, and connecting from macOS, Windows or Linux.",language:"Markdown",theme:"Research"},{name:"War Games",repoName:"War-Games",term:"Summer 2026",repo:"https://github.com/Ethical-Tech-CoLab/War-Games",demo:"https://ethical-tech-colab.github.io/War-Games/",blurb:"A terminal thriller in the shape of WarGames (1983), reframed around a modern AI agent: the only winning move is to understand the machine. Ships with a Monte Carlo simulation harness and a written case study behind the fiction.",language:"JavaScript",theme:"Diplomacy"},{name:"Malawi Prevention Platform (MVDC)",repoName:"MVDC",term:"Summer 2026",repo:"https://github.com/carolina-moron/MVDC",demo:"https://hilarious-llama-c8e989.netlify.app",blurb:"Gathers publicly reported conflict and human-rights incidents across Malawi, classifies each against a rights-based taxonomy, traces it to a likely root cause, and matches it to the Malawian organisation best placed to respond. 258 incidents across 26 of 28 districts, with 46 response actors. The working half of the Malawi Voice Data Commons, which aims to let someone report by voice, in Chichewa, from a basic phone.",language:"TypeScript",theme:"Early warning",access:"internal"},{name:"Generative AI for Good — Avatar Storytelling",repoName:"avatar-storytelling",term:"Spring 2025",repo:"https://github.com/Ethical-Tech-CoLab/Avatar-Impact-Stories",demo:"https://ethical-tech-colab.github.io/Avatar-Impact-Stories/",blurb:"Human-condition storytelling with culturally grounded digital-human avatars — short pieces produced with generative media, now gathered into Many Voices, a browser kiosk where a scattered grid of faces plays each survivor testimony in place. The stories are written from real-world human-trafficking and child-labour cases; the avatars are built with D-ID.",language:"D-ID",theme:"Storytelling"}],"publicationTopics",0,["Artificial Intelligence","Guidelines","Evacuation","Cultural heritage","Traceability","Diplomacy","Sustainability","Disaster response"],"publications",0,{eyebrow:"Publications · Academic reports",heading:"The research, written up.",intro:"Each research question the CoLab takes on is written up as an academic report. Browse the catalogue by topic, pick a title to read what it asks and what it found, then open the report itself. Titles still in preparation are shelved alongside the published ones, so you can see where the work is going.",items:[{index:"28",area:"Foundations",topic:"Artificial Intelligence",question:"What is ethical AI, and what makes any technology, or any exercise of power, ethical at all?",title:"What Is Ethical AI? Ethics, Ethical Technology, and Ethical International Relations for the Age of Intelligent Machines",summary:"The Ethical Tech CoLab's foundational paper. It traces ethics from the earliest civilizations through international affairs and human rights law to the responsible AI movement, the humanitarian sector, and the United Nations system, and defends an institutional answer: ethical AI is not a product feature but artificial intelligence whose whole lifecycle stays accountable to ethical deliberation, a human rights floor, the do no harm obligation, and the participation of those it affects. Closes with the CoLab's motivation and research philosophy.",status:"Published",date:"July 2026",url:"/publications/what-is-ethical-ai",pdf:"https://ethical-tech-colab.github.io/what-is-ethical-ai/pdf/What-Is-Ethical-AI-ETC-Report.pdf",repo:"https://github.com/Ethical-Tech-CoLab/what-is-ethical-ai"},{index:"06",area:"Research",topic:"Artificial Intelligence",question:"How can AI help researchers formulate rigorous research questions?",title:"AI-Powered Assistance in Formulating Research Questions",summary:"A survey of how AI supports researchers across the question-formulation workflow — finding gaps, generating candidate questions, summarizing the state of the art, and flagging contradictions — with guardrails against hallucinated citations and a red-team verification model. Ships with a reusable Copilot 'Researcher' prompt and a journal-credibility rubric.",status:"Published",date:"October 2025",url:"/publications/ai-research-assistant"},{index:"19",area:"AI systems",topic:"Guidelines",question:"Which AI model should be selected for a task, and on what evidence?",title:"AI Model Performance: Capabilities, Accuracy, Speed, Energy Use, and Token Economics",summary:"A comparative review of frontier, open-weight, and efficient model families, separating independently verified measurement from provider marketing. Every figure carries its source, its date, and its evidence grade. Ships with a live cost-per-accepted-task calculator and a validated data layer.",status:"Published",date:"July 2026",url:"/publications/ai-models-research"},{index:"29",area:"Research",topic:"Artificial Intelligence",question:"Can two isolated agents invent a grounded, auditable language through shared experience alone?",title:"Agentic Language Development: Emergent Communication Between Isolated Agents, and the Evidence That Would Prove It",summary:"Two agents, one channel that carries no human language, and shared tasks neither can finish alone. The report sets out the nursery architecture, the independent chronological ledgers each agent keeps of what it thinks a symbol means, the hash-chained and publicly anchored evidence trail behind them, and the nineteen ordered experiments that would test the whole thing. It also draws the line the field's own literature insists on: solving the task is not evidence that the messages were read, and a fluent ledger may be a story told after the fact.",status:"Concept",date:"August 2026",url:"/publications/agentic-language-development",repo:"https://github.com/Ethical-Tech-CoLab/agentic-language-development"},{index:"05",area:"Sustainability",topic:"Artificial Intelligence",question:"How large is AI's environmental footprint — and how can it be reduced?",title:"AI's Carbon Footprint: The Environmental Impact of AI",summary:"A CoLab report on AI's energy use across training and inference, the data-center and hardware toll, and the mitigation strategies, regulations, and policies that could bend the curve.",status:"Published",date:"May 2025",url:"/publications/ai-carbon-footprint"},{index:"27",area:"Evacuation",topic:"Evacuation",question:"How can AI inform evacuation decisions, and what happens after?",title:"After the Corridor: From AI-Informed Evacuation to Digital Public Goods for Refugee Economic Inclusion",summary:"The Summer 2026 cohort's synthesis report. It reads the five fielded evacuation prototypes and the forced-labour risk index together, states plainly what is validated and what is not, and extends the arc to the camp: the measurement, financial-modeling, and rights infrastructure that can shorten protracted displacement, grounded at Dzaleka Refugee Camp, Malawi. It also names the one thing the lab refuses to build, biometric identity for displaced people, and says why that refusal is itself a public good.",status:"Published",date:"July 2026",url:"/publications/after-the-corridor",pdf:"https://ethical-tech-colab.github.io/after-the-corridor-report/pdf/After-the-Corridor-ETC-Research-Report.pdf",repo:"https://github.com/Ethical-Tech-CoLab/after-the-corridor-report"},{index:"07",area:"Evacuation",topic:"Evacuation",question:"Is leaving actually safer than staying, and how would anyone know?",title:"The Evacuation Inform Index: Weighing the Risk of Leaving Against the Risk of Staying",summary:"A plain-language report on a decision support prototype that scores 104 active crises on two risks at once, keeping danger and feasibility deliberately apart so that operational difficulty cannot quietly cancel out a legal obligation. Prepared as masters research at the NYU Center for Global Affairs.",status:"Published",date:"July 2026",url:"/publications/evacuation-inform-index"},{index:"08",area:"Evacuation",topic:"Evacuation",question:"How dangerous is it to stay, and is leaving actually possible?",title:"The Civilian Evacuation Risk Anticipation Index: A Decision-Support Tool for Protecting Civilians During Evacuations in Armed Conflict",summary:"A plain-language report on a prototype that keeps two questions apart: how dangerous it is for civilians to remain, and whether an organised evacuation is possible at all. Separating them matters legally, because the duty to protect arises from danger, not from operational convenience.",status:"Published",date:"July 2026",url:"/publications/cerai"},{index:"09",area:"Evacuation",topic:"Evacuation",question:"What does it actually cost to move a besieged population, and what does staying cost?",title:"The Evacuation Risk and Cost Framework: Estimating the Human and Financial Cost of Civilian Evacuation in Armed Conflict",summary:"A report on a framework that scores seven weighted risk dimensions, estimates the cost of evacuating against the cost of staying, and finds the point at which those two curves cross. Calibrated against 31 documented historical operations.",status:"Published",date:"July 2026",url:"/publications/ercf"},{index:"10",area:"Evacuation",topic:"Evacuation",question:"How much does poor field information degrade an evacuation decision?",title:"The Evacuation Readiness and Uncertainty Simulator: Showing How Poor Field Information Degrades Evacuation Decisions",summary:"A report on a simulator that models what happens when planners act on incomplete intelligence. All data is synthetic and reproducible from a seed, so the mechanism rather than any particular place is what is on display.",status:"Published",date:"July 2026",url:"/publications/erus"},{index:"11",area:"Evacuation",topic:"Evacuation",question:"How do information and demographics decide who actually leaves?",title:"The Evacuation Simulator: An Agent-Based Model of How Civilians Decide to Leave During Armed Conflict",summary:"A report on an agent-based model of household evacuation behaviour, tracing how warnings spread through a community and how age, mobility, and trust shape who reaches safety and who does not.",status:"Published",date:"July 2026",url:"/publications/evacuation-simulation"},{index:"12",area:"Evacuation",topic:"Evacuation",question:"Could a daily severity measure have shown the danger in Mariupol as it unfolded?",title:"The Mariupol Corridor Severity Model: A Daily Measure of Civilian Danger During the Siege of Mariupol, March to May 2022",summary:"A retrospective report scoring 77 days of the siege on a daily severity measure built from open sources. Conduct descriptions are attributed to their sources rather than adjudicated, and contested figures are left contested.",status:"Published",date:"July 2026",url:"/publications/mariupol-severity-model"},{index:"13",area:"Cultural heritage",topic:"Cultural heritage",question:"Can an artwork's ownership history be checked automatically, and how far should that be trusted?",title:"The Digital Provenance Passport: An Automated Assistant for Tracing the Ownership History of Artworks and Cultural Objects",summary:"A report on an assistant that assembles a provenance timeline, flags looting and valuation risk, and issues a signed record. Candid about how much of the demonstration runs on fixtures rather than live sources.",status:"Published",date:"July 2026",url:"/publications/digital-provenance-passport"},{index:"14",area:"Cultural heritage",topic:"Cultural heritage",question:"What can be established about an object's history from free public sources alone?",title:"Provenance Search: An Automated Ownership-History Check for Artworks and Cultural Objects",summary:"A report on a tool that queries seven public sources and scores confidence with a fixed, inspectable algorithm, deliberately flagging what cannot be verified rather than filling the gaps.",status:"Published",date:"July 2026",url:"/publications/provenance-search"},{index:"15",area:"Cultural heritage",topic:"Cultural heritage",question:"What is worth recording about a visit to a work of art?",title:"VANGO: A Digital Passport for Recording Visits to Works of Art",summary:"A report on a digital souvenir passport that records attendance at art experiences. It performs no provenance or authenticity checking, and the report is explicit that nothing it records is evidence of ownership or lawful origin.",status:"Published",date:"July 2026",url:"/publications/vango"},{index:"16",area:"Traceability",topic:"Traceability",question:"Where are the conditions that make forced labour likely, before any case is detected?",title:"The Forced Labor Structural Risk Index: A Country-Level Measure of the Conditions Under Which Forced Labour Becomes More Likely",summary:"A report on an index scoring 184 countries across 11 domains and 43 indicators, measuring the conditions that enable forced labour rather than counting cases. Ships with uncertainty bands and is meant to be read in tiers, not as a league table.",status:"Published",date:"December 2025",url:"/publications/forced-labor-structural-risk-index"},{index:"17",area:"Diplomacy",topic:"Diplomacy",question:"Can practitioners rehearse a multi-party negotiation against AI agents?",title:"The Diplomatic Simulator: A Multi-Party Negotiation Simulator Driven by Artificial Intelligence Agents",summary:"A report on a simulator in which AI delegations negotiate a live crisis, producing a structured record of positions, concessions, and tactics. Across every Monte Carlo trial run, no comprehensive settlement was reached.",status:"Published",date:"July 2026",url:"/publications/diplomatic-simulator"},{index:"29",area:"Diplomacy",topic:"Diplomacy",question:"What happens when the machine in a nuclear thriller is a real language model?",title:"The Only Winning Move: Rebuilding WarGames (1983) as a Playable Study of Autonomous AI Agents",summary:"A report on rebuilding the 1983 film as a browser game whose antagonist is a real language model held to a strict output contract, and on what four tracks of Monte Carlo evaluation found. Structured output turned out to be solved; the real problem was that models will not escalate, stalling a quarter of games. On owned hardware the smallest model beat every cloud model on speed and the largest returned no valid output at all.",status:"Published",date:"July 2026",url:"/publications/war-games",repo:"https://github.com/Ethical-Tech-CoLab/War-Games"},{index:"18",area:"Disaster response",topic:"Disaster response",question:"How fast can building damage be assessed after a disaster, and with how few labels?",title:"HASTE: High-speed Assessment and Satellite Tracking for Emergencies",summary:"A plain-language report on a satellite damage-assessment platform built by the Microsoft AI for Good Lab, used in 31 field deployments. The software is Microsoft's; only this report is the CoLab's.",status:"Published",date:"July 2026",url:"/publications/haste"},{index:"30",area:"Guidelines",topic:"Guidelines",question:"When a repository claims to model human behaviour at scale, what is it actually modelling?",title:"The Agentic Behavior Observatory: Reading What a Repository Models, and What It Leaves Out",summary:"Paste a GitHub repository URL and get an evidence-linked read of how it models agentic behaviour, scored on five axes: agent-based simulation, synthetic data generation, language-model behavioural modelling, reinforcement learning, and evaluation. Every point is traceable to the file and line where its signal fired, and the population lab surfaces which demographic dimensions the analysed corpus models at all. A score is signal coverage, not a quality judgement.",status:"Published",date:"August 2026",url:"https://ethical-tech-colab.github.io/agentic-behavior-observatory/",repo:"https://github.com/Ethical-Tech-CoLab/agentic-behavior-observatory"},{index:"31",area:"Guidelines",topic:"Guidelines",question:"What does that term mean, and where do you get open data for the thing you are building?",title:"The Cyber Dictionary and Library",summary:"Two rooms in one site. The dictionary defines 423 technology and cybersecurity terms across 11 domains in a sentence or two of plain English, written for the moment you actually looked the term up. The library is 105 open data sources and open-source technologies across 11 shelves, from satellite imagery to conflict and rights data, each with what it gives you, how to reach its API, and what it costs.",status:"Published",date:"August 2026",url:"https://ethical-tech-colab.github.io/cyber-dictionary/",repo:"https://github.com/Ethical-Tech-CoLab/cyber-dictionary"},{index:"20",area:"Guidelines",topic:"Guidelines",question:"How does someone with no Git background start contributing?",title:"GitHub 101: Getting Started with Git and GitHub",summary:"A beginner's reference for Git and GitHub from day one — what the two actually are and how they differ, the key concepts, setting up, and the handful of commands that cover almost everything you will do in the first months.",status:"Internal guide",date:"July 2026",access:"internal",url:"https://github.com/Ethical-Tech-CoLab/Getting-Going/blob/main/docs/github-101.md"},{index:"21",area:"Guidelines",topic:"Guidelines",question:"How does a team actually collaborate in a shared repository?",title:"GitHub Workflow: Branching, Pull Requests and Code Review",summary:"The collaborative loop, step by step: the GitHub flow at a glance, creating and managing branches, opening a pull request that can be reviewed, and what to look for when reviewing someone else's.",status:"Internal guide",date:"July 2026",access:"internal",url:"https://github.com/Ethical-Tech-CoLab/Getting-Going/blob/main/docs/github-workflow.md"},{index:"22",area:"Guidelines",topic:"Guidelines",question:"How do you put a project on the web without paying for hosting?",title:"GitHub Pages 101",summary:"Publishing a free site straight from a repository, written for first-time users — no server to rent, nothing to install. Includes an honest account of what GitHub Pages can and cannot do, which is where most first attempts come unstuck.",status:"Internal guide",date:"July 2026",access:"internal",url:"https://github.com/Ethical-Tech-CoLab/Getting-Going/blob/main/docs/github-pages-101.md"},{index:"23",area:"Guidelines",topic:"Guidelines",question:"How do you build and test a data product before you have real data?",title:"Creating Synthetic Data and Seeding Models: A Beginner's Guide",summary:"Generating data algorithmically rather than collecting it: why you would (developing before real data exists, forcing edge cases, sharing something realistic without privacy exposure, reproducing a dataset from a seed), and how to do it without fooling yourself about what the results mean.",status:"Internal guide",date:"July 2026",access:"internal",url:"https://github.com/Ethical-Tech-CoLab/Getting-Going/blob/main/synthetic-data/SYNTHETIC_DATA_GUIDE.md"},{index:"24",area:"Guidelines",topic:"Guidelines",question:"How do you teach synthetic-data generation to a room of beginners?",title:"Synthetic Data: Training Deck",summary:"The companion training material to the synthetic-data guide — a self-contained walkthrough for running the topic as a session rather than reading it alone.",status:"Internal guide",date:"July 2026",access:"internal",url:"https://github.com/Ethical-Tech-CoLab/Getting-Going/blob/main/synthetic-data/synthetic-data-training.html"},{index:"25",area:"Guidelines",topic:"Guidelines",question:"How can an agent do live web research from any runtime, including locked-down ones?",title:"tavily-research: A Portable Agent Skill for Live Web Research",summary:"A cross-agent skill wrapping the Tavily REST API — search, extract, crawl and map — with a dependency-free client in both Python and Node. It calls the API directly, so it still works where a Tavily MCP server cannot be launched, such as under IT policy that blocks local npx.",status:"Internal guide",date:"July 2026",access:"internal",url:"https://github.com/Ethical-Tech-CoLab/Getting-Going/tree/main/tavily-research"},{index:"26",area:"Guidelines",topic:"Guidelines",question:"How do you get a terminal on the CoLab's GPU machine?",title:"Getting Started on Cloud GPU: SSH Access to the B3IQ Machine",summary:"One-time setup, about five minutes: installing cloudflared for the secure tunnel, creating an SSH key if you do not have one, and connecting from macOS, Windows or Linux.",status:"Internal guide",date:"July 2026",access:"internal",url:"https://github.com/Ethical-Tech-CoLab/Getting-started-on-CloudGPU-/blob/main/remote-gpu-how-to-guide.md"}]},"researchAreas",0,[{index:"01",key:"Evacuation",question:"How can AI inform evacuation decisions?",tags:["Disaster response","Civic tech","Information equity"],summary:"Turning fragmented crisis signals into decisions, so responders and residents can act before information gaps cost lives.",detail:"Projects under this question measure and improve the quality, accessibility, and equity of the information people rely on when they have to leave.",stack:["Geospatial analytics","LLM signal extraction","Open data pipelines","React dashboards"],projects:[{name:"Evacuation Inform Index",summary:"Humanitarian practice tends to treat evacuation as the obvious good, but the journey carries its own risk, and a crisis can be both catastrophic to endure and impossible to escape. This index scores 104 active crises on those two risks separately, drawing on INFORM Severity data, a twelve-factor vulnerability profile, and live conflict reporting, so that the comparison between leaving and staying becomes visible rather than assumed.",status:"Active",repo:"https://github.com/Ethical-Tech-CoLab/evacuation-inform-index-carolina",publication:"/publications/evacuation-inform-index"},{name:"Exodus — Civilian Evacuation Risk Platform",summary:"A platform unifying three civilian-evacuation risk tools: a crisis map of the INFORM Severity Index with live news and conflict timelines, a seven-dimension scenario risk model, and an endangerment-and-feasibility risk assessment, sharing one backend and design language.",status:"Active",repo:"https://github.com/Ethical-Tech-CoLab/Exodus"},{name:"ERCF — Evacuation Risk & Cost Framework",summary:"A decision-support tool that estimates the human and financial cost of civilian evacuation in armed conflict, scoring scenarios across seven risk dimensions and comparing the cost of evacuating against the cost of staying in the zone.",status:"Active",repo:"https://github.com/Ethical-Tech-CoLab/ercf",publication:"/publications/ercf"},{name:"Evacuation Behavior Simulator",summary:"An interactive, agent-based model of community evacuation behavior, modeling family clusters, information-seeking, neighbor social influence, and elder and child delays across an UNAWARE → SEEKING → MILLING → EVACUATING → DONE lifecycle.",status:"Active",repo:"https://github.com/Ethical-Tech-CoLab/Evac-Sim-Melanie",publication:"/publications/evacuation-simulation"},{name:"Evacuation Routing Simulator",summary:"An interactive evacuation simulator that models how routing and assignment choices shape who reaches safety, with shareable scenario state and a built-in explainer walkthrough.",status:"Active",repo:"https://github.com/Ethical-Tech-CoLab/India-EvacSimulation",publication:"/publications/erus"},{name:"Mariupol 2022 — Corridor Severity Model",summary:"A corridor-severity model reconstructing the 2022 Mariupol evacuation, scoring the risk and viability of humanitarian corridors under siege conditions.",status:"Active",repo:"https://github.com/Ethical-Tech-CoLab/mariupol-evacuation-model",demo:"https://ethical-tech-colab.github.io/mariupol-evacuation-model/",publication:"/publications/mariupol-severity-model"}]},{index:"02",key:"Cultural heritage",question:"How can technology support the ethical return of cultural artifacts?",tags:["Cultural heritage","Provenance","Restorative justice"],summary:"Giving institutions, source communities, and researchers a shared, verifiable record of an artifact's life across borders.",detail:"Projects under this question pair verifiable credentials with collaborative cataloguing to support repatriation, attribution, and ethical stewardship.",stack:["Verifiable credentials","IIIF imaging","Ledger anchoring","Knowledge graphs"],projects:[{name:"Arts Provenance Agent — Digital Provenance Passport",summary:"An x402-native agent that traces the provenance of artworks and artifacts, flags looting, repatriation, and valuation risk, and issues a cryptographically signed, tamper-evident Passport (a JSON-LD Verifiable Credential) for each object. Every claim is grounded in an allowlist of authoritative sources — the Met, UNESCO, ICOM, the Art Loss Register — so a fact without a citation is never produced. Includes a dashboard for tracing where an object has been, its risk score, and repatriation status.",status:"Active",repo:"https://github.com/Ethical-Tech-CoLab/arts-provenance-agent",demo:"https://ethical-tech-colab.github.io/arts-provenance-agent/",publication:"/publications/digital-provenance-passport"},{name:"Provenance Search — Provenance Intelligence",summary:"Looks up an artwork's ownership history across free public sources — Tavily search (restricted to the Met, Getty, INTERPOL, UNESCO, Art Loss Register and more), the Met and Art Institute of Chicago APIs, MoMA, Wikidata, and Europeana — and emits a provenance passport with a confidence score computed by a fixed algorithm, not the AI. Supports text, image upload, and live camera capture for use inside a museum.",status:"In development",repo:"https://github.com/Ethical-Tech-CoLab/provenance-search",publication:"/publications/provenance-search"},{name:"Digital Art Passport (VANGO)",summary:"A mobile-style digital passport for art experiences: scan a QR code or enter an artwork's code to collect a stamp — a vintage illustration of the piece with the artist, venue, and date. Stamps persist across sessions, with a collector bio and passport number, and a fully multilingual interface (English, French, Italian).",status:"Active",repo:"https://github.com/Ethical-Tech-CoLab/VANGO",demo:"https://ethical-tech-colab.github.io/VANGO/",publication:"/publications/vango"},{name:"3D Online Gallery",summary:"An immersive 3D gallery for exploring provenance-tracked and repatriated works in a navigable virtual space. Placeholder for an upcoming build.",status:"Coming soon"}]},{index:"03",key:"Traceability",question:"How can ethical claims in supply chains be made verifiable?",tags:["Labor rights","Climate accountability","Trust infrastructure"],summary:"Auditable traceability for materials, labor, and impact, from origin to shelf.",detail:"Projects under this question build interoperable trace records that travel with goods, letting buyers and regulators verify ethical claims without exposing sensitive supplier data.",stack:["Zero-knowledge proofs","ERP integrations","Standards (GS1, EPCIS)","Edge attestation"],projects:[{name:"Ethical Supply Chain & Traceability",summary:"Interoperable trace records that travel with goods, letting buyers and regulators verify ethical claims about materials, labor, and impact without exposing sensitive supplier data.",status:"Active"},{name:"Forced Labor Structural Risk Index",summary:"An interactive index that scores structural forced-labor risk across regions and sectors, turning scattered supply-chain and labor signals into a comparable measure buyers and regulators can act on. Built by Amanda Lindsey in the Fall 2025 cohort.",status:"Active",demo:"https://ethical-tech-colab.github.io/forced-labor-structural-risk-index/",publication:"/publications/forced-labor-structural-risk-index"}]},{index:"04",key:"Diplomacy",question:"How can AI help practitioners rehearse high-stakes diplomacy?",tags:["Diplomacy","AI safety","Pedagogy"],summary:"An AI-mediated environment where negotiators rehearse high-stakes diplomacy with culturally grounded agents.",detail:"Projects under this question pair domain experts with adaptive agents that model historical context, incentives, and red lines, training practitioners for situations textbooks cannot capture.",stack:["Multi-agent LLMs","Scenario authoring tools","Evaluation harnesses","Voice interfaces"],projects:[{name:"Diplomatic Simulator",summary:"A multi-party AI negotiation table where each delegation argues from its own confidential 'privileged instructions.' Five crises — Arctic, Central Asia (Fergana Valley), Cyprus, South China Sea, and Iran–US Strait of Hormuz — each with a live-negotiation replay, an interactive analysis dashboard (parties' strategy, proposals, detector hits), and Monte Carlo runs over randomized conditions.",status:"Active",repo:"https://github.com/Ethical-Tech-CoLab/diplomatic-simulator",demo:"https://ethical-tech-colab.github.io/diplomatic-simulator/",publication:"/publications/diplomatic-simulator"},{name:"War Games",summary:"A terminal thriller in the shape of WarGames (1983), reframed around a modern AI agent: the only winning move is to understand the machine. Ships with a Monte Carlo simulation harness and a written case study behind the fiction.",status:"Active",repo:"https://github.com/Ethical-Tech-CoLab/War-Games",demo:"https://ethical-tech-colab.github.io/War-Games/",publication:"/publications/war-games"}]}],"site",0,{name:"Ethical Tech CoLab",tagline:"Emerging tech, human condition.",linkedin:"https://www.linkedin.com/company/ethical-tech-lab/",email:"ethical-tech-colab@nyu.edu",partnersLine:"NYU SPS · CGA · Microsoft · New York",footerBlurb:"Exploring intervention opportunities at the intersection of emerging technologies and the human condition.",cohortRange:"Four cohorts · est. 2024-2026",social:{github:"https://github.com/Ethical-Tech-CoLab",instagram:"",twitter:""},legal:["The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.","Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners."]}])}]); \ No newline at end of file diff --git a/static-site/_next/static/chunks/1w0nyqw5bqniv.css b/static-site/_next/static/chunks/1sag07fioh6gb.css similarity index 75% rename from static-site/_next/static/chunks/1w0nyqw5bqniv.css rename to static-site/_next/static/chunks/1sag07fioh6gb.css index c89879b5a..e0dac3fdd 100644 --- a/static-site/_next/static/chunks/1w0nyqw5bqniv.css +++ b/static-site/_next/static/chunks/1sag07fioh6gb.css @@ -1,3 +1,3 @@ @font-face{font-family:Bebas Neue;font-style:normal;font-weight:400;font-display:swap;src:url(../media/2039e8342bda6056-s.1hiu0hj9qpr9g.woff2)format("woff2");unicode-range:U+100-2BA,U+2BD-2C5,U+2C7-2CC,U+2CE-2D7,U+2DD-2FF,U+304,U+308,U+329,U+1D00-1DBF,U+1E00-1E9F,U+1EF2-1EFF,U+2020,U+20A0-20AB,U+20AD-20C0,U+2113,U+2C60-2C7F,U+A720-A7FF}@font-face{font-family:Bebas Neue;font-style:normal;font-weight:400;font-display:swap;src:url(../media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2)format("woff2");unicode-range:U+??,U+131,U+152-153,U+2BB-2BC,U+2C6,U+2DA,U+2DC,U+304,U+308,U+329,U+2000-206F,U+20AC,U+2122,U+2191,U+2193,U+2212,U+2215,U+FEFF,U+FFFD}@font-face{font-family:Bebas Neue Fallback;src:local(Arial);ascent-override:117.32%;descent-override:39.11%;line-gap-override:0.0%;size-adjust:76.72%}.bebas_neue_4d1b6b20-module__PolqSW__className{font-family:Bebas Neue,Bebas Neue Fallback;font-style:normal;font-weight:400}.bebas_neue_4d1b6b20-module__PolqSW__variable{--font-bebas:"Bebas Neue", "Bebas Neue Fallback"} -@font-face{font-family:Space Mono;font-style:italic;font-weight:400;font-display:swap;src:url(../media/417912ede2e82152-s.1umk1654cmfxm.woff2)format("woff2");unicode-range:U+102-103,U+110-111,U+128-129,U+168-169,U+1A0-1A1,U+1AF-1B0,U+300-301,U+303-304,U+308-309,U+323,U+329,U+1EA0-1EF9,U+20AB}@font-face{font-family:Space Mono;font-style:italic;font-weight:400;font-display:swap;src:url(../media/51f2519120a6b711-s.038kgsu-hd-vh.woff2)format("woff2");unicode-range:U+100-2BA,U+2BD-2C5,U+2C7-2CC,U+2CE-2D7,U+2DD-2FF,U+304,U+308,U+329,U+1D00-1DBF,U+1E00-1E9F,U+1EF2-1EFF,U+2020,U+20A0-20AB,U+20AD-20C0,U+2113,U+2C60-2C7F,U+A720-A7FF}@font-face{font-family:Space Mono;font-style:italic;font-weight:400;font-display:swap;src:url(../media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2)format("woff2");unicode-range:U+??,U+131,U+152-153,U+2BB-2BC,U+2C6,U+2DA,U+2DC,U+304,U+308,U+329,U+2000-206F,U+20AC,U+2122,U+2191,U+2193,U+2212,U+2215,U+FEFF,U+FFFD}@font-face{font-family:Space Mono;font-style:italic;font-weight:700;font-display:swap;src:url(../media/2c4e8b1b42841af7-s.2p_e9n24plgne.woff2)format("woff2");unicode-range:U+102-103,U+110-111,U+128-129,U+168-169,U+1A0-1A1,U+1AF-1B0,U+300-301,U+303-304,U+308-309,U+323,U+329,U+1EA0-1EF9,U+20AB}@font-face{font-family:Space Mono;font-style:italic;font-weight:700;font-display:swap;src:url(../media/7663dec790de7af8-s.2b96g58xb1jk9.woff2)format("woff2");unicode-range:U+100-2BA,U+2BD-2C5,U+2C7-2CC,U+2CE-2D7,U+2DD-2FF,U+304,U+308,U+329,U+1D00-1DBF,U+1E00-1E9F,U+1EF2-1EFF,U+2020,U+20A0-20AB,U+20AD-20C0,U+2113,U+2C60-2C7F,U+A720-A7FF}@font-face{font-family:Space Mono;font-style:italic;font-weight:700;font-display:swap;src:url(../media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2)format("woff2");unicode-range:U+??,U+131,U+152-153,U+2BB-2BC,U+2C6,U+2DA,U+2DC,U+304,U+308,U+329,U+2000-206F,U+20AC,U+2122,U+2191,U+2193,U+2212,U+2215,U+FEFF,U+FFFD}@font-face{font-family:Space Mono;font-style:normal;font-weight:400;font-display:swap;src:url(../media/4ba802ed8e67eac5-s.0fuxtgzazslqx.woff2)format("woff2");unicode-range:U+102-103,U+110-111,U+128-129,U+168-169,U+1A0-1A1,U+1AF-1B0,U+300-301,U+303-304,U+308-309,U+323,U+329,U+1EA0-1EF9,U+20AB}@font-face{font-family:Space Mono;font-style:normal;font-weight:400;font-display:swap;src:url(../media/d7a0600e467cf0bd-s.451yigaaz_4mg.woff2)format("woff2");unicode-range:U+100-2BA,U+2BD-2C5,U+2C7-2CC,U+2CE-2D7,U+2DD-2FF,U+304,U+308,U+329,U+1D00-1DBF,U+1E00-1E9F,U+1EF2-1EFF,U+2020,U+20A0-20AB,U+20AD-20C0,U+2113,U+2C60-2C7F,U+A720-A7FF}@font-face{font-family:Space Mono;font-style:normal;font-weight:400;font-display:swap;src:url(../media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2)format("woff2");unicode-range:U+??,U+131,U+152-153,U+2BB-2BC,U+2C6,U+2DA,U+2DC,U+304,U+308,U+329,U+2000-206F,U+20AC,U+2122,U+2191,U+2193,U+2212,U+2215,U+FEFF,U+FFFD}@font-face{font-family:Space Mono;font-style:normal;font-weight:700;font-display:swap;src:url(../media/b8f2b92a9960dd69-s.0we3rzhp_h228.woff2)format("woff2");unicode-range:U+102-103,U+110-111,U+128-129,U+168-169,U+1A0-1A1,U+1AF-1B0,U+300-301,U+303-304,U+308-309,U+323,U+329,U+1EA0-1EF9,U+20AB}@font-face{font-family:Space Mono;font-style:normal;font-weight:700;font-display:swap;src:url(../media/28e60ca39c9ae554-s.44izq0ia6orkf.woff2)format("woff2");unicode-range:U+100-2BA,U+2BD-2C5,U+2C7-2CC,U+2CE-2D7,U+2DD-2FF,U+304,U+308,U+329,U+1D00-1DBF,U+1E00-1E9F,U+1EF2-1EFF,U+2020,U+20A0-20AB,U+20AD-20C0,U+2113,U+2C60-2C7F,U+A720-A7FF}@font-face{font-family:Space Mono;font-style:normal;font-weight:700;font-display:swap;src:url(../media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2)format("woff2");unicode-range:U+??,U+131,U+152-153,U+2BB-2BC,U+2C6,U+2DA,U+2DC,U+304,U+308,U+329,U+2000-206F,U+20AC,U+2122,U+2191,U+2193,U+2212,U+2215,U+FEFF,U+FFFD}@font-face{font-family:Space Mono Fallback;src:local(Arial);ascent-override:81.58%;descent-override:26.3%;line-gap-override:0.0%;size-adjust:137.28%}.space_mono_6ab7c80c-module__dmwzCa__className{font-family:Space Mono,Space Mono Fallback}.space_mono_6ab7c80c-module__dmwzCa__variable{--font-space-mono:"Space Mono", "Space Mono Fallback"} -@layer properties{@supports (((-webkit-hyphens:none)) and (not (margin-trim:inline))) or ((-moz-orient:inline) and (not (color:rgb(from red r g b)))){*,:before,:after,::backdrop{--tw-translate-x:0;--tw-translate-y:0;--tw-translate-z:0;--tw-rotate-x:initial;--tw-rotate-y:initial;--tw-rotate-z:initial;--tw-skew-x:initial;--tw-skew-y:initial;--tw-scroll-snap-strictness:proximity;--tw-space-y-reverse:0;--tw-divide-y-reverse:0;--tw-border-style:solid;--tw-gradient-position:initial;--tw-gradient-from:#0000;--tw-gradient-via:#0000;--tw-gradient-to:#0000;--tw-gradient-stops:initial;--tw-gradient-via-stops:initial;--tw-gradient-from-position:0%;--tw-gradient-via-position:50%;--tw-gradient-to-position:100%;--tw-leading:initial;--tw-font-weight:initial;--tw-tracking:initial;--tw-ordinal:initial;--tw-slashed-zero:initial;--tw-numeric-figure:initial;--tw-numeric-spacing:initial;--tw-numeric-fraction:initial;--tw-shadow:0 0 #0000;--tw-shadow-color:initial;--tw-shadow-alpha:100%;--tw-inset-shadow:0 0 #0000;--tw-inset-shadow-color:initial;--tw-inset-shadow-alpha:100%;--tw-ring-color:initial;--tw-ring-shadow:0 0 #0000;--tw-inset-ring-color:initial;--tw-inset-ring-shadow:0 0 #0000;--tw-ring-inset:initial;--tw-ring-offset-width:0px;--tw-ring-offset-color:#fff;--tw-ring-offset-shadow:0 0 #0000;--tw-outline-style:solid;--tw-blur:initial;--tw-brightness:initial;--tw-contrast:initial;--tw-grayscale:initial;--tw-hue-rotate:initial;--tw-invert:initial;--tw-opacity:initial;--tw-saturate:initial;--tw-sepia:initial;--tw-drop-shadow:initial;--tw-drop-shadow-color:initial;--tw-drop-shadow-alpha:100%;--tw-drop-shadow-size:initial;--tw-backdrop-blur:initial;--tw-backdrop-brightness:initial;--tw-backdrop-contrast:initial;--tw-backdrop-grayscale:initial;--tw-backdrop-hue-rotate:initial;--tw-backdrop-invert:initial;--tw-backdrop-opacity:initial;--tw-backdrop-saturate:initial;--tw-backdrop-sepia:initial;--tw-duration:initial;--tw-ease:initial;--tw-scale-x:1;--tw-scale-y:1;--tw-scale-z:1;--tw-content:"";--aura-rot:0deg}}}@layer theme{:root,:host{--font-sans:var(--font-space-mono);--font-serif:var(--font-bebas);--font-mono:var(--font-space-mono);--color-black:#000;--color-white:#fff;--spacing:.25rem;--container-sm:24rem;--container-md:28rem;--container-xl:36rem;--container-2xl:42rem;--container-3xl:48rem;--container-4xl:56rem;--container-5xl:64rem;--container-6xl:72rem;--text-xs:.75rem;--text-xs--line-height:calc(1 / .75);--text-sm:.875rem;--text-sm--line-height:calc(1.25 / .875);--text-base:1rem;--text-base--line-height:calc(1.5 / 1);--text-lg:1.125rem;--text-lg--line-height:calc(1.75 / 1.125);--text-xl:1.25rem;--text-xl--line-height:calc(1.75 / 1.25);--text-2xl:1.5rem;--text-2xl--line-height:calc(2 / 1.5);--text-3xl:1.875rem;--text-3xl--line-height:calc(2.25 / 1.875);--text-4xl:2.25rem;--text-4xl--line-height:calc(2.5 / 2.25);--text-5xl:3rem;--text-5xl--line-height:1;--text-6xl:3.75rem;--text-6xl--line-height:1;--text-7xl:4.5rem;--text-7xl--line-height:1;--font-weight-normal:400;--font-weight-medium:500;--font-weight-semibold:600;--tracking-tight:-.025em;--tracking-wide:.025em;--tracking-wider:.05em;--leading-tight:1.25;--leading-snug:1.375;--leading-relaxed:1.625;--radius-sm:.25rem;--radius-lg:.5rem;--radius-xl:.75rem;--radius-2xl:1rem;--radius-3xl:1.5rem;--ease-in:cubic-bezier(.4, 0, 1, 1);--ease-out:cubic-bezier(0, 0, .2, 1);--ease-in-out:cubic-bezier(.4, 0, .2, 1);--animate-pulse:pulse 2s cubic-bezier(.4, 0, .6, 1) infinite;--blur-sm:8px;--blur-md:12px;--aspect-video:16 / 9;--default-transition-duration:.15s;--default-transition-timing-function:cubic-bezier(.4, 0, .2, 1);--default-font-family:var(--font-space-mono);--default-mono-font-family:var(--font-space-mono);--font-heading:var(--font-bebas)}}@layer base{*,:after,:before,::backdrop{box-sizing:border-box;border:0 solid;margin:0;padding:0}::file-selector-button{box-sizing:border-box;border:0 solid;margin:0;padding:0}html,:host{-webkit-text-size-adjust:100%;tab-size:4;line-height:1.5;font-family:var(--default-font-family,ui-sans-serif, system-ui, sans-serif, "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Noto Color Emoji");font-feature-settings:var(--default-font-feature-settings,normal);font-variation-settings:var(--default-font-variation-settings,normal);-webkit-tap-highlight-color:transparent}hr{height:0;color:inherit;border-top-width:1px}abbr:where([title]){-webkit-text-decoration:underline dotted;text-decoration:underline dotted}h1,h2,h3,h4,h5,h6{font-size:inherit;font-weight:inherit}a{color:inherit;-webkit-text-decoration:inherit;-webkit-text-decoration:inherit;-webkit-text-decoration:inherit;-webkit-text-decoration:inherit;text-decoration:inherit}b,strong{font-weight:bolder}code,kbd,samp,pre{font-family:var(--default-mono-font-family,ui-monospace, SFMono-Regular, Menlo, Monaco, Consolas, "Liberation Mono", "Courier New", monospace);font-feature-settings:var(--default-mono-font-feature-settings,normal);font-variation-settings:var(--default-mono-font-variation-settings,normal);font-size:1em}small{font-size:80%}sub,sup{vertical-align:baseline;font-size:75%;line-height:0;position:relative}sub{bottom:-.25em}sup{top:-.5em}table{text-indent:0;border-color:inherit;border-collapse:collapse}:-moz-focusring{outline:auto}progress{vertical-align:baseline}summary{display:list-item}ol,ul,menu{list-style:none}img,svg,video,canvas,audio,iframe,embed,object{vertical-align:middle;display:block}img,video{max-width:100%;height:auto}button,input,select,optgroup,textarea{font:inherit;font-feature-settings:inherit;font-variation-settings:inherit;letter-spacing:inherit;color:inherit;opacity:1;background-color:#0000;border-radius:0}::file-selector-button{font:inherit;font-feature-settings:inherit;font-variation-settings:inherit;letter-spacing:inherit;color:inherit;opacity:1;background-color:#0000;border-radius:0}:where(select:is([multiple],[size])) optgroup{font-weight:bolder}:where(select:is([multiple],[size])) optgroup option{padding-inline-start:20px}::file-selector-button{margin-inline-end:4px}::placeholder{opacity:1}@supports (not ((-webkit-appearance:-apple-pay-button))) or (contain-intrinsic-size:1px){::placeholder{color:currentColor}@supports (color:color-mix(in lab, red, red)){::placeholder{color:color-mix(in oklab, currentcolor 50%, transparent)}}}textarea{resize:vertical}::-webkit-search-decoration{-webkit-appearance:none}::-webkit-date-and-time-value{min-height:1lh;text-align:inherit}::-webkit-datetime-edit{display:inline-flex}::-webkit-datetime-edit-fields-wrapper{padding:0}::-webkit-datetime-edit{padding-block:0}::-webkit-datetime-edit-year-field{padding-block:0}::-webkit-datetime-edit-month-field{padding-block:0}::-webkit-datetime-edit-day-field{padding-block:0}::-webkit-datetime-edit-hour-field{padding-block:0}::-webkit-datetime-edit-minute-field{padding-block:0}::-webkit-datetime-edit-second-field{padding-block:0}::-webkit-datetime-edit-millisecond-field{padding-block:0}::-webkit-datetime-edit-meridiem-field{padding-block:0}::-webkit-calendar-picker-indicator{line-height:1}:-moz-ui-invalid{box-shadow:none}button,input:where([type=button],[type=reset],[type=submit]){appearance:button}::file-selector-button{appearance:button}::-webkit-inner-spin-button{height:auto}::-webkit-outer-spin-button{height:auto}[hidden]:where(:not([hidden=until-found])){display:none!important}h1,h2,h3{font-family:var(--font-heading), "Bebas Neue", sans-serif;letter-spacing:.02em;font-weight:400}}@layer components;@layer utilities{.pointer-events-auto{pointer-events:auto}.pointer-events-none{pointer-events:none}.collapse{visibility:collapse}.invisible{visibility:hidden}.visible{visibility:visible}.sr-only{clip-path:inset(50%);white-space:nowrap;border-width:0;width:1px;height:1px;margin:-1px;padding:0;position:absolute;overflow:hidden}.absolute{position:absolute}.fixed{position:fixed}.relative{position:relative}.static{position:static}.sticky{position:sticky}.inset-0{inset:0}.inset-x-0{inset-inline:0}.inset-x-4{inset-inline:calc(var(--spacing) * 4)}.inset-y-0{inset-block:0}.-top-\[20\%\]{top:-20%}.top-0{top:0}.top-1\/2{top:50%}.top-3{top:calc(var(--spacing) * 3)}.top-4{top:calc(var(--spacing) * 4)}.top-full{top:100%}.right-0{right:0}.right-3{right:calc(var(--spacing) * 3)}.right-4{right:calc(var(--spacing) * 4)}.right-\[-60px\]{right:-60px}.bottom-0{bottom:0}.bottom-4{bottom:calc(var(--spacing) * 4)}.-left-\[20\%\]{left:-20%}.left-0{left:0}.left-3{left:calc(var(--spacing) * 3)}.left-\[-5000px\]{left:-5000px}.z-10{z-index:10}.z-40{z-index:40}.z-50{z-index:50}.z-\[55\]{z-index:55}.z-\[90\]{z-index:90}.z-\[100\]{z-index:100}.col-span-full{grid-column:1/-1}.container{width:100%}@media (min-width:40rem){.container{max-width:40rem}}@media (min-width:48rem){.container{max-width:48rem}}@media (min-width:64rem){.container{max-width:64rem}}@media (min-width:80rem){.container{max-width:80rem}}@media (min-width:96rem){.container{max-width:96rem}}.-mx-6{margin-inline:calc(var(--spacing) * -6)}.mx-auto{margin-inline:auto}.mt-1{margin-top:var(--spacing)}.mt-1\.5{margin-top:calc(var(--spacing) * 1.5)}.mt-2{margin-top:calc(var(--spacing) * 2)}.mt-3{margin-top:calc(var(--spacing) * 3)}.mt-4{margin-top:calc(var(--spacing) * 4)}.mt-5{margin-top:calc(var(--spacing) * 5)}.mt-6{margin-top:calc(var(--spacing) * 6)}.mt-7{margin-top:calc(var(--spacing) * 7)}.mt-8{margin-top:calc(var(--spacing) * 8)}.mt-10{margin-top:calc(var(--spacing) * 10)}.mt-12{margin-top:calc(var(--spacing) * 12)}.mt-14{margin-top:calc(var(--spacing) * 14)}.mt-16{margin-top:calc(var(--spacing) * 16)}.mt-20{margin-top:calc(var(--spacing) * 20)}.mt-auto{margin-top:auto}.-mr-4{margin-right:calc(var(--spacing) * -4)}.mr-2{margin-right:calc(var(--spacing) * 2)}.-mb-px{margin-bottom:-1px}.mb-2{margin-bottom:calc(var(--spacing) * 2)}.mb-3{margin-bottom:calc(var(--spacing) * 3)}.mb-6{margin-bottom:calc(var(--spacing) * 6)}.-ml-4{margin-left:calc(var(--spacing) * -4)}.ml-2{margin-left:calc(var(--spacing) * 2)}.ml-auto{margin-left:auto}.line-clamp-2{-webkit-line-clamp:2;-webkit-box-orient:vertical;display:-webkit-box;overflow:hidden}.line-clamp-3{-webkit-line-clamp:3;-webkit-box-orient:vertical;display:-webkit-box;overflow:hidden}.block{display:block}.contents{display:contents}.flex{display:flex}.grid{display:grid}.hidden{display:none}.inline{display:inline}.inline-block{display:inline-block}.inline-flex{display:inline-flex}.inline-grid{display:inline-grid}.table{display:table}.aspect-\[2\/3\]{aspect-ratio:2/3}.aspect-\[4\/3\]{aspect-ratio:4/3}.aspect-\[4\/5\]{aspect-ratio:4/5}.aspect-\[16\/7\]{aspect-ratio:16/7}.aspect-\[16\/10\]{aspect-ratio:16/10}.aspect-square{aspect-ratio:1}.aspect-video{aspect-ratio:var(--aspect-video)}.h-0\.5{height:calc(var(--spacing) * .5)}.h-1\.5{height:calc(var(--spacing) * 1.5)}.h-2\.5{height:calc(var(--spacing) * 2.5)}.h-3{height:calc(var(--spacing) * 3)}.h-3\.5{height:calc(var(--spacing) * 3.5)}.h-4{height:calc(var(--spacing) * 4)}.h-8{height:calc(var(--spacing) * 8)}.h-9{height:calc(var(--spacing) * 9)}.h-\[140\%\]{height:140%}.h-full{height:100%}.h-px{height:1px}.min-h-\[2\.75rem\]{min-height:2.75rem}.min-h-full{min-height:100%}.w-1\.5{width:calc(var(--spacing) * 1.5)}.w-2\.5{width:calc(var(--spacing) * 2.5)}.w-3\.5{width:calc(var(--spacing) * 3.5)}.w-4{width:calc(var(--spacing) * 4)}.w-6{width:calc(var(--spacing) * 6)}.w-8{width:calc(var(--spacing) * 8)}.w-9{width:calc(var(--spacing) * 9)}.w-20{width:calc(var(--spacing) * 20)}.w-52{width:calc(var(--spacing) * 52)}.w-72{width:calc(var(--spacing) * 72)}.w-\[124px\]{width:124px}.w-\[140\%\]{width:140%}.w-\[190px\]{width:190px}.w-\[min\(72vw\,380px\)\]{width:min(72vw,380px)}.w-fit{width:fit-content}.w-full{width:100%}.max-w-2xl{max-width:var(--container-2xl)}.max-w-3xl{max-width:var(--container-3xl)}.max-w-4xl{max-width:var(--container-4xl)}.max-w-5xl{max-width:var(--container-5xl)}.max-w-6xl{max-width:var(--container-6xl)}.max-w-\[40em\]{max-width:40em}.max-w-\[80vw\]{max-width:80vw}.max-w-md{max-width:var(--container-md)}.max-w-sm{max-width:var(--container-sm)}.max-w-xl{max-width:var(--container-xl)}.min-w-0{min-width:0}.min-w-\[180px\]{min-width:180px}.min-w-\[240px\]{min-width:240px}.min-w-\[640px\]{min-width:640px}.flex-1{flex:1}.shrink{flex-shrink:1}.shrink-0{flex-shrink:0}.grow{flex-grow:1}.border-collapse{border-collapse:collapse}.origin-left{transform-origin:0}.-translate-x-full{--tw-translate-x:-100%;translate:var(--tw-translate-x) var(--tw-translate-y)}.translate-x-0{--tw-translate-x:0;translate:var(--tw-translate-x) var(--tw-translate-y)}.-translate-y-1\/2{--tw-translate-y:calc(calc(1 / 2 * 100%) * -1);translate:var(--tw-translate-x) var(--tw-translate-y)}.-translate-y-2{--tw-translate-y:calc(var(--spacing) * -2);translate:var(--tw-translate-x) var(--tw-translate-y)}.translate-y-2{--tw-translate-y:calc(var(--spacing) * 2);translate:var(--tw-translate-x) var(--tw-translate-y)}.-rotate-45{rotate:-45deg}.rotate-45{rotate:45deg}.rotate-180{rotate:180deg}.transform{transform:var(--tw-rotate-x,) var(--tw-rotate-y,) var(--tw-rotate-z,) var(--tw-skew-x,) var(--tw-skew-y,)}.animate-pulse{animation:var(--animate-pulse)}.cursor-default{cursor:default}.cursor-not-allowed{cursor:not-allowed}.cursor-pointer{cursor:pointer}.resize{resize:both}.snap-x{scroll-snap-type:x var(--tw-scroll-snap-strictness)}.snap-mandatory{--tw-scroll-snap-strictness:mandatory}.snap-start{scroll-snap-align:start}.scroll-mt-24{scroll-margin-top:calc(var(--spacing) * 24)}.scroll-mt-28{scroll-margin-top:calc(var(--spacing) * 28)}.scroll-pl-6{scroll-padding-left:calc(var(--spacing) * 6)}.list-decimal{list-style-type:decimal}.list-disc{list-style-type:disc}.list-none{list-style-type:none}.grid-cols-2{grid-template-columns:repeat(2,minmax(0,1fr))}.grid-rows-\[0fr\]{grid-template-rows:0fr}.grid-rows-\[1fr\]{grid-template-rows:1fr}.flex-col{flex-direction:column}.flex-wrap{flex-wrap:wrap}.place-items-center{place-items:center}.items-baseline{align-items:baseline}.items-center{align-items:center}.items-end{align-items:flex-end}.items-start{align-items:flex-start}.justify-between{justify-content:space-between}.justify-center{justify-content:center}.justify-end{justify-content:flex-end}.gap-1{gap:var(--spacing)}.gap-1\.5{gap:calc(var(--spacing) * 1.5)}.gap-2{gap:calc(var(--spacing) * 2)}.gap-2\.5{gap:calc(var(--spacing) * 2.5)}.gap-3{gap:calc(var(--spacing) * 3)}.gap-4{gap:calc(var(--spacing) * 4)}.gap-5{gap:calc(var(--spacing) * 5)}.gap-6{gap:calc(var(--spacing) * 6)}.gap-8{gap:calc(var(--spacing) * 8)}.gap-10{gap:calc(var(--spacing) * 10)}.gap-px{gap:1px}:where(.space-y-1>:not(:last-child)){--tw-space-y-reverse:0;margin-block-start:calc(var(--spacing) * var(--tw-space-y-reverse));margin-block-end:calc(var(--spacing) * calc(1 - var(--tw-space-y-reverse)))}:where(.space-y-2>:not(:last-child)){--tw-space-y-reverse:0;margin-block-start:calc(calc(var(--spacing) * 2) * var(--tw-space-y-reverse));margin-block-end:calc(calc(var(--spacing) * 2) * calc(1 - var(--tw-space-y-reverse)))}:where(.space-y-3>:not(:last-child)){--tw-space-y-reverse:0;margin-block-start:calc(calc(var(--spacing) * 3) * var(--tw-space-y-reverse));margin-block-end:calc(calc(var(--spacing) * 3) * calc(1 - var(--tw-space-y-reverse)))}:where(.space-y-4>:not(:last-child)){--tw-space-y-reverse:0;margin-block-start:calc(calc(var(--spacing) * 4) * var(--tw-space-y-reverse));margin-block-end:calc(calc(var(--spacing) * 4) * calc(1 - var(--tw-space-y-reverse)))}:where(.space-y-5>:not(:last-child)){--tw-space-y-reverse:0;margin-block-start:calc(calc(var(--spacing) * 5) * var(--tw-space-y-reverse));margin-block-end:calc(calc(var(--spacing) * 5) * calc(1 - var(--tw-space-y-reverse)))}:where(.space-y-8>:not(:last-child)){--tw-space-y-reverse:0;margin-block-start:calc(calc(var(--spacing) * 8) * var(--tw-space-y-reverse));margin-block-end:calc(calc(var(--spacing) * 8) * calc(1 - var(--tw-space-y-reverse)))}:where(.space-y-10>:not(:last-child)){--tw-space-y-reverse:0;margin-block-start:calc(calc(var(--spacing) * 10) * var(--tw-space-y-reverse));margin-block-end:calc(calc(var(--spacing) * 10) * calc(1 - var(--tw-space-y-reverse)))}:where(.space-y-14>:not(:last-child)){--tw-space-y-reverse:0;margin-block-start:calc(calc(var(--spacing) * 14) * var(--tw-space-y-reverse));margin-block-end:calc(calc(var(--spacing) * 14) * calc(1 - var(--tw-space-y-reverse)))}:where(.space-y-16>:not(:last-child)){--tw-space-y-reverse:0;margin-block-start:calc(calc(var(--spacing) * 16) * var(--tw-space-y-reverse));margin-block-end:calc(calc(var(--spacing) * 16) * calc(1 - var(--tw-space-y-reverse)))}.gap-x-3{column-gap:calc(var(--spacing) * 3)}.gap-x-4{column-gap:calc(var(--spacing) * 4)}.gap-x-5{column-gap:calc(var(--spacing) * 5)}.gap-x-10{column-gap:calc(var(--spacing) * 10)}.gap-y-1{row-gap:var(--spacing)}.gap-y-2{row-gap:calc(var(--spacing) * 2)}.gap-y-4{row-gap:calc(var(--spacing) * 4)}:where(.divide-y>:not(:last-child)){--tw-divide-y-reverse:0;border-bottom-style:var(--tw-border-style);border-top-style:var(--tw-border-style);border-top-width:calc(1px * var(--tw-divide-y-reverse));border-bottom-width:calc(1px * calc(1 - var(--tw-divide-y-reverse)))}:where(.divide-border>:not(:last-child)){border-color:var(--border)}.truncate{text-overflow:ellipsis;white-space:nowrap;overflow:hidden}.overflow-hidden{overflow:hidden}.overflow-x-auto{overflow-x:auto}.overflow-y-auto{overflow-y:auto}.rounded{border-radius:.25rem}.rounded-2xl{border-radius:var(--radius-2xl)}.rounded-3xl{border-radius:var(--radius-3xl)}.rounded-full{border-radius:3.40282e38px}.rounded-lg{border-radius:var(--radius-lg)}.rounded-sm{border-radius:var(--radius-sm)}.rounded-xl{border-radius:var(--radius-xl)}.rounded-r-\[4px\]{border-top-right-radius:4px;border-bottom-right-radius:4px}.border{border-style:var(--tw-border-style);border-width:1px}.border-0{border-style:var(--tw-border-style);border-width:0}.border-2{border-style:var(--tw-border-style);border-width:2px}.border-x{border-inline-style:var(--tw-border-style);border-inline-width:1px}.border-t{border-top-style:var(--tw-border-style);border-top-width:1px}.border-r{border-right-style:var(--tw-border-style);border-right-width:1px}.border-b{border-bottom-style:var(--tw-border-style);border-bottom-width:1px}.border-b-2{border-bottom-style:var(--tw-border-style);border-bottom-width:2px}.border-l-2{border-left-style:var(--tw-border-style);border-left-width:2px}.border-dashed{--tw-border-style:dashed;border-style:dashed}.border-accent,.border-accent\/60{border-color:var(--accent)}@supports (color:color-mix(in lab, red, red)){.border-accent\/60{border-color:color-mix(in oklab, var(--accent) 60%, transparent)}}.border-border{border-color:var(--border)}.border-border-strong{border-color:var(--border-strong)}.border-transparent{border-color:#0000}.border-white\/35{border-color:#ffffff59}@supports (color:color-mix(in lab, red, red)){.border-white\/35{border-color:color-mix(in oklab, var(--color-white) 35%, transparent)}}.bg-\[\#11172f\]{background-color:#11172f}.bg-accent{background-color:var(--accent)}.bg-background,.bg-background\/30{background-color:var(--background)}@supports (color:color-mix(in lab, red, red)){.bg-background\/30{background-color:color-mix(in oklab, var(--background) 30%, transparent)}}.bg-background\/60{background-color:var(--background)}@supports (color:color-mix(in lab, red, red)){.bg-background\/60{background-color:color-mix(in oklab, var(--background) 60%, transparent)}}.bg-background\/70{background-color:var(--background)}@supports (color:color-mix(in lab, red, red)){.bg-background\/70{background-color:color-mix(in oklab, var(--background) 70%, transparent)}}.bg-background\/80{background-color:var(--background)}@supports (color:color-mix(in lab, red, red)){.bg-background\/80{background-color:color-mix(in oklab, var(--background) 80%, transparent)}}.bg-background\/90{background-color:var(--background)}@supports (color:color-mix(in lab, red, red)){.bg-background\/90{background-color:color-mix(in oklab, var(--background) 90%, transparent)}}.bg-black\/45{background-color:#00000073}@supports (color:color-mix(in lab, red, red)){.bg-black\/45{background-color:color-mix(in oklab, var(--color-black) 45%, transparent)}}.bg-black\/60{background-color:#0009}@supports (color:color-mix(in lab, red, red)){.bg-black\/60{background-color:color-mix(in oklab, var(--color-black) 60%, transparent)}}.bg-border{background-color:var(--border)}.bg-card{background-color:var(--card)}.bg-current{background-color:currentColor}.bg-foreground\/25{background-color:var(--foreground)}@supports (color:color-mix(in lab, red, red)){.bg-foreground\/25{background-color:color-mix(in oklab, var(--foreground) 25%, transparent)}}.bg-secondary{background-color:var(--secondary)}.bg-surface,.bg-surface\/40{background-color:var(--surface)}@supports (color:color-mix(in lab, red, red)){.bg-surface\/40{background-color:color-mix(in oklab, var(--surface) 40%, transparent)}}.bg-surface\/60{background-color:var(--surface)}@supports (color:color-mix(in lab, red, red)){.bg-surface\/60{background-color:color-mix(in oklab, var(--surface) 60%, transparent)}}.bg-white{background-color:var(--color-white)}.bg-gradient-to-r{--tw-gradient-position:to right in oklab;background-image:linear-gradient(var(--tw-gradient-stops))}.bg-gradient-to-t{--tw-gradient-position:to top in oklab;background-image:linear-gradient(var(--tw-gradient-stops))}.from-background{--tw-gradient-from:var(--background);--tw-gradient-stops:var(--tw-gradient-via-stops,var(--tw-gradient-position), var(--tw-gradient-from) var(--tw-gradient-from-position), var(--tw-gradient-to) var(--tw-gradient-to-position))}.from-black\/70{--tw-gradient-from:#000000b3}@supports (color:color-mix(in lab, red, red)){.from-black\/70{--tw-gradient-from:color-mix(in oklab, var(--color-black) 70%, transparent)}}.from-black\/70{--tw-gradient-stops:var(--tw-gradient-via-stops,var(--tw-gradient-position), var(--tw-gradient-from) var(--tw-gradient-from-position), var(--tw-gradient-to) var(--tw-gradient-to-position))}.via-background\/90{--tw-gradient-via:var(--background)}@supports (color:color-mix(in lab, red, red)){.via-background\/90{--tw-gradient-via:color-mix(in oklab, var(--background) 90%, transparent)}}.via-background\/90{--tw-gradient-via-stops:var(--tw-gradient-position), var(--tw-gradient-from) var(--tw-gradient-from-position), var(--tw-gradient-via) var(--tw-gradient-via-position), var(--tw-gradient-to) var(--tw-gradient-to-position);--tw-gradient-stops:var(--tw-gradient-via-stops)}.to-background\/45{--tw-gradient-to:var(--background)}@supports (color:color-mix(in lab, red, red)){.to-background\/45{--tw-gradient-to:color-mix(in oklab, var(--background) 45%, transparent)}}.to-background\/45{--tw-gradient-stops:var(--tw-gradient-via-stops,var(--tw-gradient-position), var(--tw-gradient-from) var(--tw-gradient-from-position), var(--tw-gradient-to) var(--tw-gradient-to-position))}.to-transparent{--tw-gradient-to:transparent;--tw-gradient-stops:var(--tw-gradient-via-stops,var(--tw-gradient-position), var(--tw-gradient-from) var(--tw-gradient-from-position), var(--tw-gradient-to) var(--tw-gradient-to-position))}.bg-cover{background-size:cover}.bg-center{background-position:50%}.bg-top{background-position:top}.object-contain{object-fit:contain}.object-cover{object-fit:cover}.object-\[70\%_35\%\]{object-position:70% 35%}.object-center{object-position:center}.p-1{padding:var(--spacing)}.p-2{padding:calc(var(--spacing) * 2)}.p-4{padding:calc(var(--spacing) * 4)}.p-5{padding:calc(var(--spacing) * 5)}.p-6{padding:calc(var(--spacing) * 6)}.p-7{padding:calc(var(--spacing) * 7)}.p-8{padding:calc(var(--spacing) * 8)}.px-2{padding-inline:calc(var(--spacing) * 2)}.px-2\.5{padding-inline:calc(var(--spacing) * 2.5)}.px-3{padding-inline:calc(var(--spacing) * 3)}.px-4{padding-inline:calc(var(--spacing) * 4)}.px-5{padding-inline:calc(var(--spacing) * 5)}.px-6{padding-inline:calc(var(--spacing) * 6)}.py-0\.5{padding-block:calc(var(--spacing) * .5)}.py-1{padding-block:var(--spacing)}.py-1\.5{padding-block:calc(var(--spacing) * 1.5)}.py-2{padding-block:calc(var(--spacing) * 2)}.py-2\.5{padding-block:calc(var(--spacing) * 2.5)}.py-3{padding-block:calc(var(--spacing) * 3)}.py-4{padding-block:calc(var(--spacing) * 4)}.py-8{padding-block:calc(var(--spacing) * 8)}.py-12{padding-block:calc(var(--spacing) * 12)}.py-16{padding-block:calc(var(--spacing) * 16)}.py-20{padding-block:calc(var(--spacing) * 20)}.py-24{padding-block:calc(var(--spacing) * 24)}.py-28{padding-block:calc(var(--spacing) * 28)}.py-32{padding-block:calc(var(--spacing) * 32)}.pt-2{padding-top:calc(var(--spacing) * 2)}.pt-4{padding-top:calc(var(--spacing) * 4)}.pt-5{padding-top:calc(var(--spacing) * 5)}.pt-6{padding-top:calc(var(--spacing) * 6)}.pt-7{padding-top:calc(var(--spacing) * 7)}.pt-8{padding-top:calc(var(--spacing) * 8)}.pt-10{padding-top:calc(var(--spacing) * 10)}.pt-12{padding-top:calc(var(--spacing) * 12)}.pt-24{padding-top:calc(var(--spacing) * 24)}.pr-4{padding-right:calc(var(--spacing) * 4)}.pb-2{padding-bottom:calc(var(--spacing) * 2)}.pb-3{padding-bottom:calc(var(--spacing) * 3)}.pb-8{padding-bottom:calc(var(--spacing) * 8)}.pb-24{padding-bottom:calc(var(--spacing) * 24)}.pl-1{padding-left:var(--spacing)}.pl-5{padding-left:calc(var(--spacing) * 5)}.pl-6{padding-left:calc(var(--spacing) * 6)}.text-center{text-align:center}.text-left{text-align:left}.align-top{vertical-align:top}.font-heading{font-family:var(--font-bebas)}.font-mono,.font-sans{font-family:var(--font-space-mono)}.font-serif{font-family:var(--font-bebas)}.text-2xl{font-size:var(--text-2xl);line-height:var(--tw-leading,var(--text-2xl--line-height))}.text-3xl{font-size:var(--text-3xl);line-height:var(--tw-leading,var(--text-3xl--line-height))}.text-4xl{font-size:var(--text-4xl);line-height:var(--tw-leading,var(--text-4xl--line-height))}.text-5xl{font-size:var(--text-5xl);line-height:var(--tw-leading,var(--text-5xl--line-height))}.text-7xl{font-size:var(--text-7xl);line-height:var(--tw-leading,var(--text-7xl--line-height))}.text-base{font-size:var(--text-base);line-height:var(--tw-leading,var(--text-base--line-height))}.text-lg{font-size:var(--text-lg);line-height:var(--tw-leading,var(--text-lg--line-height))}.text-sm{font-size:var(--text-sm);line-height:var(--tw-leading,var(--text-sm--line-height))}.text-xl{font-size:var(--text-xl);line-height:var(--tw-leading,var(--text-xl--line-height))}.text-xs{font-size:var(--text-xs);line-height:var(--tw-leading,var(--text-xs--line-height))}.text-\[0\.65rem\]{font-size:.65rem}.text-\[9px\]{font-size:9px}.text-\[10px\]{font-size:10px}.text-\[11px\]{font-size:11px}.text-\[clamp\(4\.25rem\,13vw\,11rem\)\]{font-size:clamp(4.25rem,13vw,11rem)}.leading-\[0\.9\]{--tw-leading:.9;line-height:.9}.leading-\[0\.88\]{--tw-leading:.88;line-height:.88}.leading-\[0\.95\]{--tw-leading:.95;line-height:.95}.leading-\[1\.02\]{--tw-leading:1.02;line-height:1.02}.leading-\[1\.05\]{--tw-leading:1.05;line-height:1.05}.leading-\[1\.12\]{--tw-leading:1.12;line-height:1.12}.leading-none{--tw-leading:1;line-height:1}.leading-relaxed{--tw-leading:var(--leading-relaxed);line-height:var(--leading-relaxed)}.leading-snug{--tw-leading:var(--leading-snug);line-height:var(--leading-snug)}.leading-tight{--tw-leading:var(--leading-tight);line-height:var(--leading-tight)}.font-medium{--tw-font-weight:var(--font-weight-medium);font-weight:var(--font-weight-medium)}.font-normal{--tw-font-weight:var(--font-weight-normal);font-weight:var(--font-weight-normal)}.font-semibold{--tw-font-weight:var(--font-weight-semibold);font-weight:var(--font-weight-semibold)}.tracking-\[0\.02em\]{--tw-tracking:.02em;letter-spacing:.02em}.tracking-\[0\.2em\]{--tw-tracking:.2em;letter-spacing:.2em}.tracking-\[0\.06em\]{--tw-tracking:.06em;letter-spacing:.06em}.tracking-\[0\.12em\]{--tw-tracking:.12em;letter-spacing:.12em}.tracking-\[0\.16em\]{--tw-tracking:.16em;letter-spacing:.16em}.tracking-\[0\.18em\]{--tw-tracking:.18em;letter-spacing:.18em}.tracking-\[0\.25em\]{--tw-tracking:.25em;letter-spacing:.25em}.tracking-\[0\.35em\]{--tw-tracking:.35em;letter-spacing:.35em}.tracking-tight{--tw-tracking:var(--tracking-tight);letter-spacing:var(--tracking-tight)}.tracking-wide{--tw-tracking:var(--tracking-wide);letter-spacing:var(--tracking-wide)}.tracking-wider{--tw-tracking:var(--tracking-wider);letter-spacing:var(--tracking-wider)}.break-words{overflow-wrap:break-word}.whitespace-nowrap{white-space:nowrap}.text-\[\"\]\)\<\/script\>\self\.__next_f\.push\(\[1\,\"11px\]{color:"])404: This page could not be found.NYU Ethical Tech CoLab · Emerging Tech, Human Condition

404

This page could not be found.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file +404: This page could not be found.NYU Ethical Tech CoLab · Emerging Tech, Human Condition

404

This page could not be found.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file diff --git a/static-site/_not-found/index.txt b/static-site/_not-found/index.txt index 9af29dde1..270493479 100644 --- a/static-site/_not-found/index.txt +++ b/static-site/_not-found/index.txt @@ -1,18 +1,18 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -11:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -0:{"P":null,"c":["","_not-found",""],"q":"","i":false,"f":[[["",{"children":["/_not-found",{"children":["__PAGE__",{}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{},null,false,null]},null,false,"$@f"]},null,false,null],"$L10",false]],"m":"$undefined","G":["$11",["$L12"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -13:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -14:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +11:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +0:{"P":null,"c":["","_not-found",""],"q":"","i":false,"f":[[["",{"children":["/_not-found",{"children":["__PAGE__",{}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{},null,false,null]},null,false,"$@f"]},null,false,null],"$L10",false]],"m":"$undefined","G":["$11",["$L12"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +13:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +14:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 15:"$Sreact.suspense" -18:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +18:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L13",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L13",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L13",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L13",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -23,8 +23,8 @@ e:["$","$1","c",{"children":[[["$","title",null,{"children":"404: This page coul 17:[] f:"$W17" 10:["$","$1","h",{"children":[["$","meta",null,{"name":"robots","content":"noindex"}],["$","$L18",null,{"children":"$L19"}],["$","div",null,{"hidden":true,"children":["$","$L1a",null,{"children":["$","$15",null,{"name":"Next.Metadata","children":"$L1b"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -12:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +12:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 19:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -1c:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +1c:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 16:null 1b:[["$","title","0",{"children":"NYU Ethical Tech CoLab · Emerging Tech, Human Condition"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","link","2",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","3",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","4",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L1c","5",{}]] diff --git a/static-site/about/__next._full.txt b/static-site/about/__next._full.txt index 354dde10c..4096b082e 100644 --- a/static-site/about/__next._full.txt +++ b/static-site/about/__next._full.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -11:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +11:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","about",""],"q":"","i":false,"f":[[["",{"children":["about",{"children":["__PAGE__",{}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{},null,false,null]},null,false,"$@f"]},null,false,null],"$L10",false]],"m":"$undefined","G":["$11",["$L12"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -13:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -15:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","about",""],"q":"","i":false,"f":[[["",{"children":["about",{"children":["__PAGE__",{}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{},null,false,null]},null,false,"$@f"]},null,false,null],"$L10",false]],"m":"$undefined","G":["$11",["$L12"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +13:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +15:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 16:"$Sreact.suspense" -19:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1b:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +19:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1b:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L13",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L13",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L13",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L13",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -28,9 +28,9 @@ e:["$","$1","c",{"children":["$L14",null,["$","$L15",null,{"children":["$","$16" 18:[] f:"$W18" 10:["$","$1","h",{"children":[null,["$","$L19",null,{"children":"$L1a"}],["$","div",null,{"hidden":true,"children":["$","$L1b",null,{"children":["$","$16",null,{"name":"Next.Metadata","children":"$L1c"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -12:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +12:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 14:E{"digest":"NEXT_HTTP_ERROR_FALLBACK;404"} 1a:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -1d:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +1d:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 17:null 1c:[["$","title","0",{"children":"NYU Ethical Tech CoLab · Emerging Tech, Human Condition"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","link","2",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","3",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","4",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L1d","5",{}]] diff --git a/static-site/about/__next._head.txt b/static-site/about/__next._head.txt index 1bee38eda..aed0ebfe4 100644 --- a/static-site/about/__next._head.txt +++ b/static-site/about/__next._head.txt @@ -1,6 +1,6 @@ 1:"$Sreact.fragment" -2:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -3:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +2:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +3:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 4:"$Sreact.suspense" -5:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +5:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 0:{"rsc":["$","$1","h",{"children":[null,["$","$L2",null,{"children":[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]]}],["$","div",null,{"hidden":true,"children":["$","$L3",null,{"children":["$","$4",null,{"name":"Next.Metadata","children":[["$","title","0",{"children":"NYU Ethical Tech CoLab · Emerging Tech, Human Condition"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","link","2",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","3",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","4",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L5","5",{}]]}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/about/__next._index.txt b/static-site/about/__next._index.txt index f98a4fd90..18b9d82b6 100644 --- a/static-site/about/__next._index.txt +++ b/static-site/about/__next._index.txt @@ -1,13 +1,13 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -d:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +d:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} 8:["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$Ld",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$Ld",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$Ld",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$Ld",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] diff --git a/static-site/about/__next._tree.txt b/static-site/about/__next._tree.txt index 3974819de..6af1b77ae 100644 --- a/static-site/about/__next._tree.txt +++ b/static-site/about/__next._tree.txt @@ -1,7 +1,7 @@ -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] 0:{"tree":{"name":"","param":null,"prefetchHints":16,"slots":{"children":{"name":"about","param":null,"prefetchHints":0,"slots":{"children":{"name":"__PAGE__","param":null,"prefetchHints":0,"slots":null}}}}},"staleTime":300,"buildId":"static-snapshot"} diff --git a/static-site/about/__next.about.txt b/static-site/about/__next.about.txt index e17ca4ded..e8bfa323c 100644 --- a/static-site/about/__next.about.txt +++ b/static-site/about/__next.about.txt @@ -1,5 +1,5 @@ 1:"$Sreact.fragment" -2:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -3:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +2:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +3:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] 4:[] 0:{"rsc":["$","$1","c",{"children":[null,["$","$L2",null,{"parallelRouterKey":"children","template":["$","$L3",null,{}]}]]}],"isPartial":false,"staleTime":300,"varyParams":"$W4","buildId":"static-snapshot"} diff --git a/static-site/about/index.html b/static-site/about/index.html index 0c36d910f..d37f55008 100644 --- a/static-site/about/index.html +++ b/static-site/about/index.html @@ -1 +1 @@ -NYU Ethical Tech CoLab · Emerging Tech, Human Condition \ No newline at end of file +NYU Ethical Tech CoLab · Emerging Tech, Human Condition \ No newline at end of file diff --git a/static-site/about/index.txt b/static-site/about/index.txt index 354dde10c..4096b082e 100644 --- a/static-site/about/index.txt +++ b/static-site/about/index.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -11:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +11:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","about",""],"q":"","i":false,"f":[[["",{"children":["about",{"children":["__PAGE__",{}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{},null,false,null]},null,false,"$@f"]},null,false,null],"$L10",false]],"m":"$undefined","G":["$11",["$L12"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -13:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -15:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","about",""],"q":"","i":false,"f":[[["",{"children":["about",{"children":["__PAGE__",{}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{},null,false,null]},null,false,"$@f"]},null,false,null],"$L10",false]],"m":"$undefined","G":["$11",["$L12"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +13:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +15:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 16:"$Sreact.suspense" -19:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1b:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +19:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1b:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L13",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L13",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L13",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L13",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -28,9 +28,9 @@ e:["$","$1","c",{"children":["$L14",null,["$","$L15",null,{"children":["$","$16" 18:[] f:"$W18" 10:["$","$1","h",{"children":[null,["$","$L19",null,{"children":"$L1a"}],["$","div",null,{"hidden":true,"children":["$","$L1b",null,{"children":["$","$16",null,{"name":"Next.Metadata","children":"$L1c"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -12:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +12:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 14:E{"digest":"NEXT_HTTP_ERROR_FALLBACK;404"} 1a:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -1d:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +1d:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 17:null 1c:[["$","title","0",{"children":"NYU Ethical Tech CoLab · Emerging Tech, Human Condition"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","link","2",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","3",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","4",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L1d","5",{}]] diff --git a/static-site/contact/__next._full.txt b/static-site/contact/__next._full.txt index 36c280999..778273937 100644 --- a/static-site/contact/__next._full.txt +++ b/static-site/contact/__next._full.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -11:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +11:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","contact",""],"q":"","i":false,"f":[[["",{"children":["contact",{"children":["__PAGE__",{}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{},null,false,null]},null,false,"$@f"]},null,false,null],"$L10",false]],"m":"$undefined","G":["$11",["$L12"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -13:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -14:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","contact",""],"q":"","i":false,"f":[[["",{"children":["contact",{"children":["__PAGE__",{}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{},null,false,null]},null,false,"$@f"]},null,false,null],"$L10",false]],"m":"$undefined","G":["$11",["$L12"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +13:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +14:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 15:"$Sreact.suspense" -18:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +18:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L13",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L13",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L13",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L13",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -28,8 +28,8 @@ e:["$","$1","c",{"children":[[["$","section",null,{"className":"border-b border- 17:[] f:"$W17" 10:["$","$1","h",{"children":[null,["$","$L18",null,{"children":"$L19"}],["$","div",null,{"hidden":true,"children":["$","$L1a",null,{"children":["$","$15",null,{"name":"Next.Metadata","children":"$L1b"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -12:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +12:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 19:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -1c:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +1c:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 16:null 1b:[["$","title","0",{"children":"Contact · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"Partner with the NYU Ethical Tech CoLab. Start a conversation about interventions at the edge of technology and society."}],["$","link","2",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","3",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","4",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L1c","5",{}]] diff --git a/static-site/contact/__next._head.txt b/static-site/contact/__next._head.txt index 225a2625d..363e9dbe3 100644 --- a/static-site/contact/__next._head.txt +++ b/static-site/contact/__next._head.txt @@ -1,6 +1,6 @@ 1:"$Sreact.fragment" -2:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -3:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +2:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +3:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 4:"$Sreact.suspense" -5:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +5:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 0:{"rsc":["$","$1","h",{"children":[null,["$","$L2",null,{"children":[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]]}],["$","div",null,{"hidden":true,"children":["$","$L3",null,{"children":["$","$4",null,{"name":"Next.Metadata","children":[["$","title","0",{"children":"Contact · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"Partner with the NYU Ethical Tech CoLab. Start a conversation about interventions at the edge of technology and society."}],["$","link","2",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","3",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","4",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L5","5",{}]]}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/contact/__next._index.txt b/static-site/contact/__next._index.txt index f98a4fd90..18b9d82b6 100644 --- a/static-site/contact/__next._index.txt +++ b/static-site/contact/__next._index.txt @@ -1,13 +1,13 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -d:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +d:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} 8:["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$Ld",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$Ld",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$Ld",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$Ld",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] diff --git a/static-site/contact/__next._tree.txt b/static-site/contact/__next._tree.txt index f0dad8c98..0288e7c87 100644 --- a/static-site/contact/__next._tree.txt +++ b/static-site/contact/__next._tree.txt @@ -1,7 +1,7 @@ -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] 0:{"tree":{"name":"","param":null,"prefetchHints":16,"slots":{"children":{"name":"contact","param":null,"prefetchHints":0,"slots":{"children":{"name":"__PAGE__","param":null,"prefetchHints":0,"slots":null}}}}},"staleTime":300,"buildId":"static-snapshot"} diff --git a/static-site/contact/__next.contact.txt b/static-site/contact/__next.contact.txt index e17ca4ded..e8bfa323c 100644 --- a/static-site/contact/__next.contact.txt +++ b/static-site/contact/__next.contact.txt @@ -1,5 +1,5 @@ 1:"$Sreact.fragment" -2:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -3:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +2:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +3:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] 4:[] 0:{"rsc":["$","$1","c",{"children":[null,["$","$L2",null,{"parallelRouterKey":"children","template":["$","$L3",null,{}]}]]}],"isPartial":false,"staleTime":300,"varyParams":"$W4","buildId":"static-snapshot"} diff --git a/static-site/contact/index.html b/static-site/contact/index.html index b71aed5c3..599913396 100644 --- a/static-site/contact/index.html +++ b/static-site/contact/index.html @@ -1 +1 @@ -Contact · NYU Ethical Tech CoLab

Contact

Start a conversation.

Have a hard problem at the edge of technology and society? We partner with institutions, agencies, and communities to prototype interventions that hold up outside the lab.

Reach us

Where we are

NYU SPS · CGA · Microsoft · New York

Founding partners

NYU School of Professional StudiesCenter for Global Affairs (CGA)Microsoft

Let's talk

Connect with us on LinkedIn

Send a message or a connection request and tell us about the problem you're working on. We read every note and reply to partnership and collaboration inquiries.

Message us on LinkedIn

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file +Contact · NYU Ethical Tech CoLab

Contact

Start a conversation.

Have a hard problem at the edge of technology and society? We partner with institutions, agencies, and communities to prototype interventions that hold up outside the lab.

Reach us

Where we are

NYU SPS · CGA · Microsoft · New York

Founding partners

NYU School of Professional StudiesCenter for Global Affairs (CGA)Microsoft

Let's talk

Connect with us on LinkedIn

Send a message or a connection request and tell us about the problem you're working on. We read every note and reply to partnership and collaboration inquiries.

Message us on LinkedIn

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file diff --git a/static-site/contact/index.txt b/static-site/contact/index.txt index 36c280999..778273937 100644 --- a/static-site/contact/index.txt +++ b/static-site/contact/index.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -11:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +11:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","contact",""],"q":"","i":false,"f":[[["",{"children":["contact",{"children":["__PAGE__",{}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{},null,false,null]},null,false,"$@f"]},null,false,null],"$L10",false]],"m":"$undefined","G":["$11",["$L12"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -13:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -14:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","contact",""],"q":"","i":false,"f":[[["",{"children":["contact",{"children":["__PAGE__",{}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{},null,false,null]},null,false,"$@f"]},null,false,null],"$L10",false]],"m":"$undefined","G":["$11",["$L12"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +13:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +14:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 15:"$Sreact.suspense" -18:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +18:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L13",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L13",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L13",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L13",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -28,8 +28,8 @@ e:["$","$1","c",{"children":[[["$","section",null,{"className":"border-b border- 17:[] f:"$W17" 10:["$","$1","h",{"children":[null,["$","$L18",null,{"children":"$L19"}],["$","div",null,{"hidden":true,"children":["$","$L1a",null,{"children":["$","$15",null,{"name":"Next.Metadata","children":"$L1b"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -12:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +12:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 19:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -1c:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +1c:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 16:null 1b:[["$","title","0",{"children":"Contact · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"Partner with the NYU Ethical Tech CoLab. Start a conversation about interventions at the edge of technology and society."}],["$","link","2",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","3",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","4",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L1c","5",{}]] diff --git a/static-site/demos/__next._full.txt b/static-site/demos/__next._full.txt index 09fc0e0e1..766438385 100644 --- a/static-site/demos/__next._full.txt +++ b/static-site/demos/__next._full.txt @@ -1,26 +1,26 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -11:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +11:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","demos",""],"q":"","i":false,"f":[[["",{"children":["demos",{"children":["__PAGE__",{}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{},null,false,null]},null,false,"$@f"]},null,false,null],"$L10",false]],"m":"$undefined","G":["$11",["$L12"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -13:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -14:I[82987,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3w90c42nje_vm.js"],"Reveal"] -15:I[18620,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3w90c42nje_vm.js"],"SectionTabs"] -16:I[89852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3w90c42nje_vm.js"],"RepoShowcase"] -17:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","demos",""],"q":"","i":false,"f":[[["",{"children":["demos",{"children":["__PAGE__",{}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{},null,false,null]},null,false,"$@f"]},null,false,null],"$L10",false]],"m":"$undefined","G":["$11",["$L12"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +13:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +14:I[82987,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3w90c42nje_vm.js"],"Reveal"] +15:I[18620,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3w90c42nje_vm.js"],"SectionTabs"] +16:I[89852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3w90c42nje_vm.js"],"RepoShowcase"] +17:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 18:"$Sreact.suspense" -1b:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1d:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1b:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1d:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L13",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L13",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L13",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L13",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -31,8 +31,8 @@ e:["$","$1","c",{"children":[[["$","section",null,{"className":"relative overflo 1a:[] f:"$W1a" 10:["$","$1","h",{"children":[null,["$","$L1b",null,{"children":"$L1c"}],["$","div",null,{"hidden":true,"children":["$","$L1d",null,{"children":["$","$18",null,{"name":"Next.Metadata","children":"$L1e"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -12:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +12:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1c:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -1f:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +1f:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 19:null 1e:[["$","title","0",{"children":"Live Demos · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"Live, in-browser demos and open-source repositories from the Ethical Tech CoLab — evacuation risk platforms, provenance passports, and decision-support tools you can run right now."}],["$","link","2",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","3",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","4",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L1f","5",{}]] diff --git a/static-site/demos/__next._head.txt b/static-site/demos/__next._head.txt index 7776b0d7b..15ca6fd89 100644 --- a/static-site/demos/__next._head.txt +++ b/static-site/demos/__next._head.txt @@ -1,6 +1,6 @@ 1:"$Sreact.fragment" -2:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -3:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +2:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +3:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 4:"$Sreact.suspense" -5:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +5:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 0:{"rsc":["$","$1","h",{"children":[null,["$","$L2",null,{"children":[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]]}],["$","div",null,{"hidden":true,"children":["$","$L3",null,{"children":["$","$4",null,{"name":"Next.Metadata","children":[["$","title","0",{"children":"Live Demos · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"Live, in-browser demos and open-source repositories from the Ethical Tech CoLab — evacuation risk platforms, provenance passports, and decision-support tools you can run right now."}],["$","link","2",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","3",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","4",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L5","5",{}]]}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/demos/__next._index.txt b/static-site/demos/__next._index.txt index f98a4fd90..18b9d82b6 100644 --- a/static-site/demos/__next._index.txt +++ b/static-site/demos/__next._index.txt @@ -1,13 +1,13 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -d:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +d:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} 8:["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$Ld",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$Ld",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$Ld",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$Ld",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] diff --git a/static-site/demos/__next._tree.txt b/static-site/demos/__next._tree.txt index f40dc43c1..402adf7f1 100644 --- a/static-site/demos/__next._tree.txt +++ b/static-site/demos/__next._tree.txt @@ -1,7 +1,7 @@ -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] 0:{"tree":{"name":"","param":null,"prefetchHints":16,"slots":{"children":{"name":"demos","param":null,"prefetchHints":0,"slots":{"children":{"name":"__PAGE__","param":null,"prefetchHints":0,"slots":null}}}}},"staleTime":300,"buildId":"static-snapshot"} diff --git a/static-site/demos/__next.demos.txt b/static-site/demos/__next.demos.txt index e17ca4ded..e8bfa323c 100644 --- a/static-site/demos/__next.demos.txt +++ b/static-site/demos/__next.demos.txt @@ -1,5 +1,5 @@ 1:"$Sreact.fragment" -2:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -3:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +2:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +3:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] 4:[] 0:{"rsc":["$","$1","c",{"children":[null,["$","$L2",null,{"parallelRouterKey":"children","template":["$","$L3",null,{}]}]]}],"isPartial":false,"staleTime":300,"varyParams":"$W4","buildId":"static-snapshot"} diff --git a/static-site/demos/index.html b/static-site/demos/index.html index 00c945678..319d5f382 100644 --- a/static-site/demos/index.html +++ b/static-site/demos/index.html @@ -1 +1 @@ -Live Demos · NYU Ethical Tech CoLab

Live demos · Open source

Run the research.

Every project ships as an open repository — and most run live in your browser. Browse the catalog by theme, pick a title to read what it does, then press play and run it here. This is applied research you can actually use.

Semester19 live · 26 shown
Theme

Evacuation

7 titles

Cultural heritage

4 titles

Traceability

1 title

Early warning

1 title

Diplomacy

2 titles

Research

10 titles

Storytelling

1 title

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file +Live Demos · NYU Ethical Tech CoLab

Live demos · Open source

Run the research.

Every project ships as an open repository — and most run live in your browser. Browse the catalog by theme, pick a title to read what it does, then press play and run it here. This is applied research you can actually use.

Semester22 live · 29 shown
Theme

Evacuation

7 titles

Cultural heritage

4 titles

Traceability

1 title

Early warning

1 title

Diplomacy

2 titles

Research

13 titles

Storytelling

1 title

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file diff --git a/static-site/demos/index.txt b/static-site/demos/index.txt index 09fc0e0e1..766438385 100644 --- a/static-site/demos/index.txt +++ b/static-site/demos/index.txt @@ -1,26 +1,26 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -11:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +11:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","demos",""],"q":"","i":false,"f":[[["",{"children":["demos",{"children":["__PAGE__",{}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{},null,false,null]},null,false,"$@f"]},null,false,null],"$L10",false]],"m":"$undefined","G":["$11",["$L12"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -13:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -14:I[82987,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3w90c42nje_vm.js"],"Reveal"] -15:I[18620,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3w90c42nje_vm.js"],"SectionTabs"] -16:I[89852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3w90c42nje_vm.js"],"RepoShowcase"] -17:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","demos",""],"q":"","i":false,"f":[[["",{"children":["demos",{"children":["__PAGE__",{}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{},null,false,null]},null,false,"$@f"]},null,false,null],"$L10",false]],"m":"$undefined","G":["$11",["$L12"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +13:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +14:I[82987,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3w90c42nje_vm.js"],"Reveal"] +15:I[18620,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3w90c42nje_vm.js"],"SectionTabs"] +16:I[89852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3w90c42nje_vm.js"],"RepoShowcase"] +17:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 18:"$Sreact.suspense" -1b:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1d:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1b:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1d:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L13",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L13",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L13",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L13",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -31,8 +31,8 @@ e:["$","$1","c",{"children":[[["$","section",null,{"className":"relative overflo 1a:[] f:"$W1a" 10:["$","$1","h",{"children":[null,["$","$L1b",null,{"children":"$L1c"}],["$","div",null,{"hidden":true,"children":["$","$L1d",null,{"children":["$","$18",null,{"name":"Next.Metadata","children":"$L1e"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -12:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +12:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1c:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -1f:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +1f:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 19:null 1e:[["$","title","0",{"children":"Live Demos · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"Live, in-browser demos and open-source repositories from the Ethical Tech CoLab — evacuation risk platforms, provenance passports, and decision-support tools you can run right now."}],["$","link","2",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","3",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","4",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L1f","5",{}]] diff --git a/static-site/font-lab/__next._full.txt b/static-site/font-lab/__next._full.txt index a325b22b1..4f88e7c67 100644 --- a/static-site/font-lab/__next._full.txt +++ b/static-site/font-lab/__next._full.txt @@ -1,35 +1,35 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -11:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +11:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] :HL["/website/_next/static/media/01d67e7cc17e7674-s.p.0m75r3fswdwux.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/0acc7fdf55eb3220-s.p.3oprs0vbfre0x.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/0c89a48fa5027cee-s.p.2cyn07wtgehh0.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/0c8c4ded07fff55c-s.p.2tcyrya9o07vu.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/3206eb66b875a5b3-s.p.201eoo3y6c5_i.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/374a88ea0960b3d4-s.p.2fz20jxmyia8o.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/37f34e4fcce0f2d4-s.p.3dosikwwiz7p4.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/8ec783ae88469012-s.p.0twlw20dvfl-2.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/d23f31b94ad01e54-s.p.0teq8rsbqpt0q.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/d6749d348ea0202d-s.p.2avioznhdf_as.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/dc0c65e819e3bb6c-s.p.1d4mg4yt567i3.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/de161955856a921d-s.p.0fxeqrss3ag9h.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fe83cf2ab39e9c57-s.p.2qpii4b6jmzj9.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/chunks/0-y3whugx_hm0.css","style"] -0:{"P":null,"c":["","font-lab",""],"q":"","i":false,"f":[[["",{"children":["font-lab",{"children":["__PAGE__",{}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{},null,false,null]},null,false,"$@f"]},null,false,null],"$L10",false]],"m":"$undefined","G":["$11",["$L12"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -13:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -1d:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1f:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +0:{"P":null,"c":["","font-lab",""],"q":"","i":false,"f":[[["",{"children":["font-lab",{"children":["__PAGE__",{}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{},null,false,null]},null,false,"$@f"]},null,false,null],"$L10",false]],"m":"$undefined","G":["$11",["$L12"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +13:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +1d:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1f:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 20:"$Sreact.suspense" 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L13",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L13",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L13",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] @@ -41,8 +41,8 @@ e:["$","$1","c",{"children":[["$","div",null,{"className":"mx-auto max-w-5xl px- 1c:[] f:"$W1c" 10:["$","$1","h",{"children":[null,["$","$L1d",null,{"children":"$L1e"}],["$","div",null,{"hidden":true,"children":["$","$L1f",null,{"children":["$","$20",null,{"name":"Next.Metadata","children":"$L21"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -12:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] -22:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +12:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +22:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 14:["$","li","6",{"className":"border-t border-border pt-8","children":[["$","p",null,{"className":"text-sm uppercase tracking-[0.18em] text-muted","children":[6," · ","VT323"," ",["$","span",null,{"className":"text-accent","children":["— ","Terminal"]}]]}],["$","p",null,{"className":"mt-2 max-w-2xl text-sm leading-relaxed text-muted","children":"Old CRT terminal. Reads as a screen rather than as a grid of blocks."}],["$","div",null,{"className":"mt-6 uppercase vt323_7a5e21bf-module__LtreqW__className","style":{"fontSize":"clamp(2.8rem,10vw,8rem)","lineHeight":0.95},"children":[["$","span",null,{"className":"block","children":"Ethical Tech"}],["$","span",null,{"className":"block","children":"CoLab"}]]}]]}] 15:["$","li","7",{"className":"border-t border-border pt-8","children":[["$","p",null,{"className":"text-sm uppercase tracking-[0.18em] text-muted","children":[7," · ","Orbitron"," ",["$","span",null,{"className":"text-accent","children":["— ","Wide tech"]}]]}],["$","p",null,{"className":"mt-2 max-w-2xl text-sm leading-relaxed text-muted","children":"Squared geometric. Says 'technology' with no pixel grid at all."}],["$","div",null,{"className":"mt-6 uppercase orbitron_bc263156-module__09bM4q__className","style":{"fontSize":"clamp(1.7rem,6vw,4.4rem)","lineHeight":1.1,"fontWeight":800},"children":[["$","span",null,{"className":"block","children":"Ethical Tech"}],["$","span",null,{"className":"block","children":"CoLab"}]]}]]}] 16:["$","li","8",{"className":"border-t border-border pt-8","children":[["$","p",null,{"className":"text-sm uppercase tracking-[0.18em] text-muted","children":[8," · ","Michroma"," ",["$","span",null,{"className":"text-accent","children":["— ","Wide tech"]}]]}],["$","p",null,{"className":"mt-2 max-w-2xl text-sm leading-relaxed text-muted","children":"Very wide, square, calm. Closest to the banner's proportions."}],["$","div",null,{"className":"mt-6 uppercase michroma_87627b29-module__lO6qWG__className","style":{"fontSize":"clamp(1.3rem,4.6vw,3.3rem)","lineHeight":1.3},"children":[["$","span",null,{"className":"block","children":"Ethical Tech"}],["$","span",null,{"className":"block","children":"CoLab"}]]}]]}] @@ -52,6 +52,6 @@ f:"$W1c" 1a:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/0-y3whugx_hm0.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1b:["$","$L22",null,{"children":["$","$20",null,{"name":"Next.MetadataOutlet","children":"$@23"}]}] 1e:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -24:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +24:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 21:[["$","title","0",{"children":"Wordmark font lab · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L24","6",{}]] 23:null diff --git a/static-site/font-lab/__next._head.txt b/static-site/font-lab/__next._head.txt index d38a3c4f1..0f4c7c5f1 100644 --- a/static-site/font-lab/__next._head.txt +++ b/static-site/font-lab/__next._head.txt @@ -1,6 +1,6 @@ 1:"$Sreact.fragment" -2:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -3:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +2:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +3:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 4:"$Sreact.suspense" -5:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +5:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 0:{"rsc":["$","$1","h",{"children":[null,["$","$L2",null,{"children":[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]]}],["$","div",null,{"hidden":true,"children":["$","$L3",null,{"children":["$","$4",null,{"name":"Next.Metadata","children":[["$","title","0",{"children":"Wordmark font lab · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L5","6",{}]]}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/font-lab/__next._index.txt b/static-site/font-lab/__next._index.txt index f98a4fd90..18b9d82b6 100644 --- a/static-site/font-lab/__next._index.txt +++ b/static-site/font-lab/__next._index.txt @@ -1,13 +1,13 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -d:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +d:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} 8:["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$Ld",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$Ld",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$Ld",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$Ld",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] diff --git a/static-site/font-lab/__next._tree.txt b/static-site/font-lab/__next._tree.txt index 8a46e89a8..a2431ac8c 100644 --- a/static-site/font-lab/__next._tree.txt +++ b/static-site/font-lab/__next._tree.txt @@ -1,19 +1,19 @@ -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] :HL["/website/_next/static/media/01d67e7cc17e7674-s.p.0m75r3fswdwux.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/0acc7fdf55eb3220-s.p.3oprs0vbfre0x.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/0c89a48fa5027cee-s.p.2cyn07wtgehh0.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/0c8c4ded07fff55c-s.p.2tcyrya9o07vu.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/3206eb66b875a5b3-s.p.201eoo3y6c5_i.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/374a88ea0960b3d4-s.p.2fz20jxmyia8o.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/37f34e4fcce0f2d4-s.p.3dosikwwiz7p4.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/8ec783ae88469012-s.p.0twlw20dvfl-2.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/d23f31b94ad01e54-s.p.0teq8rsbqpt0q.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/d6749d348ea0202d-s.p.2avioznhdf_as.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/dc0c65e819e3bb6c-s.p.1d4mg4yt567i3.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/de161955856a921d-s.p.0fxeqrss3ag9h.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fe83cf2ab39e9c57-s.p.2qpii4b6jmzj9.woff2","font",{"crossOrigin":"","type":"font/woff2"}] diff --git a/static-site/font-lab/__next.font-lab.txt b/static-site/font-lab/__next.font-lab.txt index e17ca4ded..e8bfa323c 100644 --- a/static-site/font-lab/__next.font-lab.txt +++ b/static-site/font-lab/__next.font-lab.txt @@ -1,5 +1,5 @@ 1:"$Sreact.fragment" -2:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -3:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +2:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +3:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] 4:[] 0:{"rsc":["$","$1","c",{"children":[null,["$","$L2",null,{"parallelRouterKey":"children","template":["$","$L3",null,{}]}]]}],"isPartial":false,"staleTime":300,"varyParams":"$W4","buildId":"static-snapshot"} diff --git a/static-site/font-lab/index.html b/static-site/font-lab/index.html index 5924b878b..d007bda3b 100644 --- a/static-site/font-lab/index.html +++ b/static-site/font-lab/index.html @@ -1 +1 @@ -Wordmark font lab · NYU Ethical Tech CoLab

Internal · not linked from the site

Wordmark font lab

The lockup at hero size in every candidate, on the real background. Tell me a number and I will set the wordmark in it — or say none of them and we will go the SVG route instead.

  • 1 · Bebas Neue — Current

    What the site uses today, and what every other heading is set in.

    Ethical TechCoLab

  • 2 · Sixtyfour — Dot matrix

    Closest to the banner: separate square tiles. You rejected this one.

    Ethical TechCoLab

  • 3 · Press Start 2P — Pixel

    The arcade classic. Chunky, square, unmistakably 8-bit.

    Ethical TechCoLab

  • 4 · Silkscreen — Pixel

    Small-pixel face. You rejected this one.

    Ethical TechCoLab

  • 5 · Pixelify Sans — Pixel

    Solid strokes with stepped edges. You rejected this one.

    Ethical TechCoLab

  • 6 · VT323 — Terminal

    Old CRT terminal. Reads as a screen rather than as a grid of blocks.

    Ethical TechCoLab

  • 7 · Orbitron — Wide tech

    Squared geometric. Says 'technology' with no pixel grid at all.

    Ethical TechCoLab

  • 8 · Michroma — Wide tech

    Very wide, square, calm. Closest to the banner's proportions.

    Ethical TechCoLab

  • 9 · Chakra Petch — Wide tech

    Squared with clipped corners — a technical look that still reads fast.

    Ethical TechCoLab

  • 10 · Rajdhani — Wide tech

    Condensed and squared: nearest to Bebas's proportions, but technical.

    Ethical TechCoLab

  • 11 · Space Grotesk — Neutral

    A quiet modern sans, sibling to the Space Mono already used for body.

    Ethical TechCoLab

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file +Wordmark font lab · NYU Ethical Tech CoLab

Internal · not linked from the site

Wordmark font lab

The lockup at hero size in every candidate, on the real background. Tell me a number and I will set the wordmark in it — or say none of them and we will go the SVG route instead.

  • 1 · Bebas Neue — Current

    What the site uses today, and what every other heading is set in.

    Ethical TechCoLab

  • 2 · Sixtyfour — Dot matrix

    Closest to the banner: separate square tiles. You rejected this one.

    Ethical TechCoLab

  • 3 · Press Start 2P — Pixel

    The arcade classic. Chunky, square, unmistakably 8-bit.

    Ethical TechCoLab

  • 4 · Silkscreen — Pixel

    Small-pixel face. You rejected this one.

    Ethical TechCoLab

  • 5 · Pixelify Sans — Pixel

    Solid strokes with stepped edges. You rejected this one.

    Ethical TechCoLab

  • 6 · VT323 — Terminal

    Old CRT terminal. Reads as a screen rather than as a grid of blocks.

    Ethical TechCoLab

  • 7 · Orbitron — Wide tech

    Squared geometric. Says 'technology' with no pixel grid at all.

    Ethical TechCoLab

  • 8 · Michroma — Wide tech

    Very wide, square, calm. Closest to the banner's proportions.

    Ethical TechCoLab

  • 9 · Chakra Petch — Wide tech

    Squared with clipped corners — a technical look that still reads fast.

    Ethical TechCoLab

  • 10 · Rajdhani — Wide tech

    Condensed and squared: nearest to Bebas's proportions, but technical.

    Ethical TechCoLab

  • 11 · Space Grotesk — Neutral

    A quiet modern sans, sibling to the Space Mono already used for body.

    Ethical TechCoLab

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file diff --git a/static-site/font-lab/index.txt b/static-site/font-lab/index.txt index a325b22b1..4f88e7c67 100644 --- a/static-site/font-lab/index.txt +++ b/static-site/font-lab/index.txt @@ -1,35 +1,35 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -11:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +11:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] :HL["/website/_next/static/media/01d67e7cc17e7674-s.p.0m75r3fswdwux.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/0acc7fdf55eb3220-s.p.3oprs0vbfre0x.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/0c89a48fa5027cee-s.p.2cyn07wtgehh0.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/0c8c4ded07fff55c-s.p.2tcyrya9o07vu.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/3206eb66b875a5b3-s.p.201eoo3y6c5_i.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/374a88ea0960b3d4-s.p.2fz20jxmyia8o.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/37f34e4fcce0f2d4-s.p.3dosikwwiz7p4.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/8ec783ae88469012-s.p.0twlw20dvfl-2.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/d23f31b94ad01e54-s.p.0teq8rsbqpt0q.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/d6749d348ea0202d-s.p.2avioznhdf_as.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/dc0c65e819e3bb6c-s.p.1d4mg4yt567i3.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/de161955856a921d-s.p.0fxeqrss3ag9h.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fe83cf2ab39e9c57-s.p.2qpii4b6jmzj9.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/chunks/0-y3whugx_hm0.css","style"] -0:{"P":null,"c":["","font-lab",""],"q":"","i":false,"f":[[["",{"children":["font-lab",{"children":["__PAGE__",{}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{},null,false,null]},null,false,"$@f"]},null,false,null],"$L10",false]],"m":"$undefined","G":["$11",["$L12"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -13:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -1d:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1f:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +0:{"P":null,"c":["","font-lab",""],"q":"","i":false,"f":[[["",{"children":["font-lab",{"children":["__PAGE__",{}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{},null,false,null]},null,false,"$@f"]},null,false,null],"$L10",false]],"m":"$undefined","G":["$11",["$L12"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +13:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +1d:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1f:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 20:"$Sreact.suspense" 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L13",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L13",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L13",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] @@ -41,8 +41,8 @@ e:["$","$1","c",{"children":[["$","div",null,{"className":"mx-auto max-w-5xl px- 1c:[] f:"$W1c" 10:["$","$1","h",{"children":[null,["$","$L1d",null,{"children":"$L1e"}],["$","div",null,{"hidden":true,"children":["$","$L1f",null,{"children":["$","$20",null,{"name":"Next.Metadata","children":"$L21"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -12:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] -22:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +12:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +22:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 14:["$","li","6",{"className":"border-t border-border pt-8","children":[["$","p",null,{"className":"text-sm uppercase tracking-[0.18em] text-muted","children":[6," · ","VT323"," ",["$","span",null,{"className":"text-accent","children":["— ","Terminal"]}]]}],["$","p",null,{"className":"mt-2 max-w-2xl text-sm leading-relaxed text-muted","children":"Old CRT terminal. Reads as a screen rather than as a grid of blocks."}],["$","div",null,{"className":"mt-6 uppercase vt323_7a5e21bf-module__LtreqW__className","style":{"fontSize":"clamp(2.8rem,10vw,8rem)","lineHeight":0.95},"children":[["$","span",null,{"className":"block","children":"Ethical Tech"}],["$","span",null,{"className":"block","children":"CoLab"}]]}]]}] 15:["$","li","7",{"className":"border-t border-border pt-8","children":[["$","p",null,{"className":"text-sm uppercase tracking-[0.18em] text-muted","children":[7," · ","Orbitron"," ",["$","span",null,{"className":"text-accent","children":["— ","Wide tech"]}]]}],["$","p",null,{"className":"mt-2 max-w-2xl text-sm leading-relaxed text-muted","children":"Squared geometric. Says 'technology' with no pixel grid at all."}],["$","div",null,{"className":"mt-6 uppercase orbitron_bc263156-module__09bM4q__className","style":{"fontSize":"clamp(1.7rem,6vw,4.4rem)","lineHeight":1.1,"fontWeight":800},"children":[["$","span",null,{"className":"block","children":"Ethical Tech"}],["$","span",null,{"className":"block","children":"CoLab"}]]}]]}] 16:["$","li","8",{"className":"border-t border-border pt-8","children":[["$","p",null,{"className":"text-sm uppercase tracking-[0.18em] text-muted","children":[8," · ","Michroma"," ",["$","span",null,{"className":"text-accent","children":["— ","Wide tech"]}]]}],["$","p",null,{"className":"mt-2 max-w-2xl text-sm leading-relaxed text-muted","children":"Very wide, square, calm. Closest to the banner's proportions."}],["$","div",null,{"className":"mt-6 uppercase michroma_87627b29-module__lO6qWG__className","style":{"fontSize":"clamp(1.3rem,4.6vw,3.3rem)","lineHeight":1.3},"children":[["$","span",null,{"className":"block","children":"Ethical Tech"}],["$","span",null,{"className":"block","children":"CoLab"}]]}]]}] @@ -52,6 +52,6 @@ f:"$W1c" 1a:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/0-y3whugx_hm0.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1b:["$","$L22",null,{"children":["$","$20",null,{"name":"Next.MetadataOutlet","children":"$@23"}]}] 1e:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -24:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +24:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 21:[["$","title","0",{"children":"Wordmark font lab · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L24","6",{}]] 23:null diff --git a/static-site/index.html b/static-site/index.html index 2e1d7ec94..6ac9f985c 100644 --- a/static-site/index.html +++ b/static-site/index.html @@ -1 +1 @@ -NYU Ethical Tech CoLab · Emerging Tech, Human Condition

NYU CGA × Microsoft

Ethical Tech CoLab

Exploring technology to improve the human condition.

A research collaboration between NYU's Center for Global Affairs and Microsoft — changing the conversation on how people are informed, and how emerging technology can be used for good.

Mission

Achieving full human potential through technology — changing how people are informed, and how tech is used for good.

What are the circumstances of the human being, and how can technology improve them? We see technology as a tool — part of the solution.

Current projects

Building at the frontier.

Cohorts · 2025-2026

Cohorts.

Each graduate student cohort of the Ethical Tech CoLab takes on applied research and proof-of-concept projects on intervention opportunities at the intersection of emerging technologies and the human condition.

Past

Fall 2025

Prototyping and partner pilots.

Technical spikes (multi-agent harnesses, verifiable credentials, geospatial pipelines) tested against real partner needs.

  • 8 researchers
  • Forced Labor Structural Risk Index
  • AI Research Question Assistant

Full archive coming soon

Portfolio →Meet the cohort →
Past

Spring 2025

First applied research projects.

The lab's first applied research projects, spanning online safety, sustainability, AI's footprint, and generative storytelling.

  • 8 researchers
  • Online Grooming Prevention
  • ESG Labels & Certificates Transparency
  • AI's Carbon Footprint
  • Generative AI for Good — Avatar Storytelling

Full archive coming soon

Portfolio →Meet the cohort →

Collaborate

Have a hard problem at the edge of technology and society?

We partner with institutions, agencies, and communities to prototype interventions that hold up outside the lab.

Start a conversation ethical-tech-colab@nyu.edu

Stay in the loop

Join the Ethical Tech CoLab newsletter.

Cohort openings, project launches, event invites, and research from the frontier of emerging tech and the human condition — a few times a semester, no noise.

Coming soon

No spam. Unsubscribe anytime.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file +NYU Ethical Tech CoLab · Emerging Tech, Human Condition

NYU CGA × Microsoft

Ethical Tech CoLab

Exploring technology to improve the human condition.

A research collaboration between NYU's Center for Global Affairs and Microsoft — changing the conversation on how people are informed, and how emerging technology can be used for good.

Mission

Achieving full human potential through technology — changing how people are informed, and how tech is used for good.

What are the circumstances of the human being, and how can technology improve them? We see technology as a tool — part of the solution.

Current projects

Building at the frontier.

Cohorts · 2025-2026

Cohorts.

Each graduate student cohort of the Ethical Tech CoLab takes on applied research and proof-of-concept projects on intervention opportunities at the intersection of emerging technologies and the human condition.

Past

Fall 2025

Prototyping and partner pilots.

Technical spikes (multi-agent harnesses, verifiable credentials, geospatial pipelines) tested against real partner needs.

  • 8 researchers
  • Forced Labor Structural Risk Index
  • AI Research Question Assistant

Full archive coming soon

Portfolio →Meet the cohort →
Past

Spring 2025

First applied research projects.

The lab's first applied research projects, spanning online safety, sustainability, AI's footprint, and generative storytelling.

  • 8 researchers
  • Online Grooming Prevention
  • ESG Labels & Certificates Transparency
  • AI's Carbon Footprint
  • Generative AI for Good — Avatar Storytelling

Full archive coming soon

Portfolio →Meet the cohort →

Collaborate

Have a hard problem at the edge of technology and society?

We partner with institutions, agencies, and communities to prototype interventions that hold up outside the lab.

Start a conversation ethical-tech-colab@nyu.edu

Stay in the loop

Join the Ethical Tech CoLab newsletter.

Cohort openings, project launches, event invites, and research from the frontier of emerging tech and the human condition — a few times a semester, no noise.

Coming soon

No spam. Unsubscribe anytime.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file diff --git a/static-site/index.txt b/static-site/index.txt index 390ae5f57..2361f8185 100644 --- a/static-site/index.txt +++ b/static-site/index.txt @@ -1,40 +1,40 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -f:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +f:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["",""],"q":"","i":false,"f":[[["",{"children":["__PAGE__",{}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{},null,false,null]},null,false,null],"$Le",false]],"m":"$undefined","G":["$f",["$L10"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -11:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"Link"] -12:I[85437,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"Image"] -13:I[25150,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"HeroField"] -14:I[82987,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"Reveal"] -15:I[53675,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"StatementCarousel"] -1e:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -20:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +0:{"P":null,"c":["",""],"q":"","i":false,"f":[[["",{"children":["__PAGE__",{}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{},null,false,null]},null,false,null],"$Le",false]],"m":"$undefined","G":["$f",["$L10"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +11:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"Link"] +12:I[85437,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"Image"] +13:I[25150,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"HeroField"] +14:I[82987,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"Reveal"] +15:I[53675,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"StatementCarousel"] +1e:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +20:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 21:"$Sreact.suspense" 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L11",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L11",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L11",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L11",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] b:["$","div",null,{"className":"mt-12 flex flex-col gap-2 border-t border-border pt-6 text-xs text-muted sm:flex-row sm:items-center sm:justify-between","children":[["$","span",null,{"children":["© ",2026," NYU Ethical Tech CoLab"]}],["$","span",null,{"children":"Four cohorts · est. 2024-2026"}]]}] c:["$","div",null,{"className":"mt-6 space-y-3 border-t border-border pt-6 text-[11px] leading-relaxed text-muted/80","children":[["$","p","0",{"children":"The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution."}],["$","p","1",{"children":"Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners."}]]}] -d:["$","$1","c",{"children":[[["$","section",null,{"className":"relative overflow-hidden border-b border-border","children":[["$","$L12",null,{"src":"/website/nyu-subway.jpg","alt":"","aria-hidden":true,"fill":true,"priority":true,"sizes":"100vw","className":"pointer-events-none object-cover object-center opacity-25"}],["$","div",null,{"aria-hidden":true,"className":"pointer-events-none absolute inset-0 bg-background/70"}],["$","span",null,{"className":"aura"}],["$","div",null,{"aria-hidden":true,"className":"pointer-events-none absolute inset-0","style":{"background":"radial-gradient(70% 60% at 20% 0%, color-mix(in oklab, var(--glow) 26%, transparent), transparent 65%)"}}],["$","$L13",null,{}],["$","div",null,{"className":"relative mx-auto max-w-6xl px-6 py-24 text-center sm:py-28","children":[["$","$L14",null,{"children":["$","p",null,{"className":"text-base uppercase tracking-[0.25em] text-accent sm:text-lg","children":"NYU CGA × Microsoft"}]}],["$","$L14",null,{"delay":0.15,"children":["$","$L15",null,{"statements":[{"lead":"Ethical Tech CoLab","headingClass":"text-[clamp(4.25rem,13vw,11rem)] leading-[0.88]","figure":["$","p",null,{"className":"font-serif uppercase leading-[0.95] tracking-tight text-foreground","style":{"fontSize":"clamp(1.9rem, 4vw, 3rem)"},"children":["Exploring technology to improve",["$","br",null,{"className":"hidden sm:block"}]," the"," ",["$","span",null,{"className":"display-em","children":"human condition"}],"."]}],"line":["$","p",null,{"className":"mx-auto mt-7 max-w-2xl leading-relaxed text-foreground/85","children":["A research collaboration between NYU's"," ",["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"link-underline text-accent hover:opacity-80","children":"Center for Global Affairs"}]," ","and Microsoft — changing the conversation on how people are informed, and how emerging technology can be used for good."]}]},{"lead":"Four questions. ","em":"One frontier.","figure":"14 projects in the portfolio","line":"Evacuation, cultural heritage, traceability, diplomacy — each question carried through to a fielded prototype, in the open.","cta":"Explore the portfolio","href":"/portfolio"},{"lead":"Run the ","em":"research","tail":".","figure":"18 demos you can open","line":"Not screenshots: the prototypes themselves, running in the browser with their source alongside.","cta":"Open the live demos","href":"/demos"},{"lead":"The research, ","em":"written up","tail":".","figure":"25 in the catalogue","line":"Every research question the CoLab takes on is written up academically, including what did not hold.","cta":"Read the publications","href":"/publications"},{"lead":"The people ","em":"building","tail":" this.","figure":"23 researchers across 3 cohorts","line":"Graduate researchers at NYU's Center for Global Affairs, with advisors and resident fellows alongside.","cta":"Meet the team","href":"/team"}],"label":"Statement","className":"mt-5"}]}]]}]]}],[["$","section",null,{"className":"border-b border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16 text-center","children":[["$","$L14",null,{"children":["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Mission"}]}],["$","$L14",null,{"delay":0.05,"children":["$","p",null,{"className":"mx-auto mt-4 max-w-4xl fluid-h2 font-heading uppercase leading-[1.02]","children":["Achieving ",["$","span",null,{"className":"display-em","children":"full human potential"}]," ","through technology — changing how people are informed, and how tech is used for good."]}]}],["$","$L14",null,{"delay":0.1,"children":["$","p",null,{"className":"mx-auto mt-6 max-w-2xl text-lg leading-relaxed text-muted","children":"What are the circumstances of the human being, and how can technology improve them? We see technology as a tool — part of the solution."}]}]]}]}],["$","section",null,{"className":"mx-auto max-w-6xl px-6 py-24","children":[["$","$L14",null,{"children":["$","div",null,{"className":"flex items-end justify-between","children":["$L16","$L17"]}]}],"$L18"]}],"$L19","$L1a","$L1b"]],["$L1c"],"$L1d"]}] +d:["$","$1","c",{"children":[[["$","section",null,{"className":"relative overflow-hidden border-b border-border","children":[["$","$L12",null,{"src":"/website/nyu-subway.jpg","alt":"","aria-hidden":true,"fill":true,"priority":true,"sizes":"100vw","className":"pointer-events-none object-cover object-center opacity-25"}],["$","div",null,{"aria-hidden":true,"className":"pointer-events-none absolute inset-0 bg-background/70"}],["$","span",null,{"className":"aura"}],["$","div",null,{"aria-hidden":true,"className":"pointer-events-none absolute inset-0","style":{"background":"radial-gradient(70% 60% at 20% 0%, color-mix(in oklab, var(--glow) 26%, transparent), transparent 65%)"}}],["$","$L13",null,{}],["$","div",null,{"className":"relative mx-auto max-w-6xl px-6 py-24 text-center sm:py-28","children":[["$","$L14",null,{"children":["$","p",null,{"className":"text-base uppercase tracking-[0.25em] text-accent sm:text-lg","children":"NYU CGA × Microsoft"}]}],["$","$L14",null,{"delay":0.15,"children":["$","$L15",null,{"statements":[{"lead":"Ethical Tech CoLab","headingClass":"text-[clamp(4.25rem,13vw,11rem)] leading-[0.88]","figure":["$","p",null,{"className":"font-serif uppercase leading-[0.95] tracking-tight text-foreground","style":{"fontSize":"clamp(1.9rem, 4vw, 3rem)"},"children":["Exploring technology to improve",["$","br",null,{"className":"hidden sm:block"}]," the"," ",["$","span",null,{"className":"display-em","children":"human condition"}],"."]}],"line":["$","p",null,{"className":"mx-auto mt-7 max-w-2xl leading-relaxed text-foreground/85","children":["A research collaboration between NYU's"," ",["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"link-underline text-accent hover:opacity-80","children":"Center for Global Affairs"}]," ","and Microsoft — changing the conversation on how people are informed, and how emerging technology can be used for good."]}]},{"lead":"Four questions. ","em":"One frontier.","figure":"14 projects in the portfolio","line":"Evacuation, cultural heritage, traceability, diplomacy — each question carried through to a fielded prototype, in the open.","cta":"Explore the portfolio","href":"/portfolio"},{"lead":"Run the ","em":"research","tail":".","figure":"21 demos you can open","line":"Not screenshots: the prototypes themselves, running in the browser with their source alongside.","cta":"Open the live demos","href":"/demos"},{"lead":"The research, ","em":"written up","tail":".","figure":"28 in the catalogue","line":"Every research question the CoLab takes on is written up academically, including what did not hold.","cta":"Read the publications","href":"/publications"},{"lead":"The people ","em":"building","tail":" this.","figure":"23 researchers across 3 cohorts","line":"Graduate researchers at NYU's Center for Global Affairs, with advisors and resident fellows alongside.","cta":"Meet the team","href":"/team"}],"label":"Statement","className":"mt-5"}]}]]}]]}],[["$","section",null,{"className":"border-b border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16 text-center","children":[["$","$L14",null,{"children":["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Mission"}]}],["$","$L14",null,{"delay":0.05,"children":["$","p",null,{"className":"mx-auto mt-4 max-w-4xl fluid-h2 font-heading uppercase leading-[1.02]","children":["Achieving ",["$","span",null,{"className":"display-em","children":"full human potential"}]," ","through technology — changing how people are informed, and how tech is used for good."]}]}],["$","$L14",null,{"delay":0.1,"children":["$","p",null,{"className":"mx-auto mt-6 max-w-2xl text-lg leading-relaxed text-muted","children":"What are the circumstances of the human being, and how can technology improve them? We see technology as a tool — part of the solution."}]}]]}]}],["$","section",null,{"className":"mx-auto max-w-6xl px-6 py-24","children":[["$","$L14",null,{"children":["$","div",null,{"className":"flex items-end justify-between","children":["$L16","$L17"]}]}],"$L18"]}],"$L19","$L1a","$L1b"]],["$L1c"],"$L1d"]}] e:["$","$1","h",{"children":[null,["$","$L1e",null,{"children":"$L1f"}],["$","div",null,{"hidden":true,"children":["$","$L20",null,{"children":["$","$21",null,{"name":"Next.Metadata","children":"$L22"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -10:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] -23:I[82987,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"Stagger"] -24:I[82987,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"StaggerItem"] -25:I[35852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"ProjectDiagram"] -26:I[12243,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"Tilt3D"] -2a:I[94376,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"Magnetic"] -2b:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +10:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +23:I[82987,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"Stagger"] +24:I[82987,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"StaggerItem"] +25:I[35852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"ProjectDiagram"] +26:I[12243,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"Tilt3D"] +2a:I[94376,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/3y2ffd9s1k5e5.js"],"Magnetic"] +2b:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 16:["$","div",null,{"children":[["$","$L11",null,{"href":"/portfolio","className":"link-underline text-xs uppercase tracking-wider text-muted transition-colors hover:text-accent","children":"Current projects"}],["$","h2",null,{"className":"mt-3 fluid-h2 font-heading uppercase","children":"Building at the frontier."}]]}] 17:["$","$L11",null,{"href":"/portfolio","className":"link-underline hidden text-sm text-accent sm:block","children":["View all ","four"," →"]}] 18:["$","$L23",null,{"className":"mt-12 grid gap-px overflow-hidden rounded-2xl border border-border bg-border sm:grid-cols-2","children":[[["$","$L24","Evacuation",{"children":["$","$L11",null,{"href":"/portfolio#evacuation","className":"group card-glow flex h-full flex-col gap-3 bg-background p-8 transition-colors hover:bg-surface","children":[["$","$L25",null,{"variant":"Evacuation","className":"diagram-live mb-3 aspect-[16/7] w-full overflow-hidden rounded-xl border border-border bg-surface/60"}],["$","span",null,{"className":"font-mono text-xs text-muted","children":["01"," / ","Evacuation"]}],["$","h3",null,{"className":"font-heading text-2xl uppercase tracking-wide sm:text-3xl group-hover:text-accent","children":"How can AI inform evacuation decisions?"}]]}]}],["$","$L24","Cultural heritage",{"children":["$","$L11",null,{"href":"/portfolio#cultural-heritage","className":"group card-glow flex h-full flex-col gap-3 bg-background p-8 transition-colors hover:bg-surface","children":[["$","$L25",null,{"variant":"Cultural heritage","className":"diagram-live mb-3 aspect-[16/7] w-full overflow-hidden rounded-xl border border-border bg-surface/60"}],["$","span",null,{"className":"font-mono text-xs text-muted","children":["02"," / ","Cultural heritage"]}],["$","h3",null,{"className":"font-heading text-2xl uppercase tracking-wide sm:text-3xl group-hover:text-accent","children":"How can technology support the ethical return of cultural artifacts?"}]]}]}],["$","$L24","Traceability",{"children":["$","$L11",null,{"href":"/portfolio#traceability","className":"group card-glow flex h-full flex-col gap-3 bg-background p-8 transition-colors hover:bg-surface","children":[["$","$L25",null,{"variant":"Traceability","className":"diagram-live mb-3 aspect-[16/7] w-full overflow-hidden rounded-xl border border-border bg-surface/60"}],["$","span",null,{"className":"font-mono text-xs text-muted","children":["03"," / ","Traceability"]}],["$","h3",null,{"className":"font-heading text-2xl uppercase tracking-wide sm:text-3xl group-hover:text-accent","children":"How can ethical claims in supply chains be made verifiable?"}]]}]}],["$","$L24","Diplomacy",{"children":["$","$L11",null,{"href":"/portfolio#diplomacy","className":"group card-glow flex h-full flex-col gap-3 bg-background p-8 transition-colors hover:bg-surface","children":[["$","$L25",null,{"variant":"Diplomacy","className":"diagram-live mb-3 aspect-[16/7] w-full overflow-hidden rounded-xl border border-border bg-surface/60"}],["$","span",null,{"className":"font-mono text-xs text-muted","children":["04"," / ","Diplomacy"]}],["$","h3",null,{"className":"font-heading text-2xl uppercase tracking-wide sm:text-3xl group-hover:text-accent","children":"How can AI help practitioners rehearse high-stakes diplomacy?"}]]}]}]],false]}] @@ -47,6 +47,6 @@ e:["$","$1","h",{"children":[null,["$","$L1e",null,{"children":"$L1f"}],["$","di 28:["$","$L26","02",{"max":7,"children":["$","article",null,{"className":"flex h-full flex-col rounded-2xl border bg-card p-7 transition-colors border-border hover:border-foreground/25","children":[["$","div",null,{"className":"flex items-center justify-end","children":["$","span",null,{"className":"rounded-full border border-border px-2.5 py-0.5 text-xs text-muted","children":"Past"}]}],["$","h3",null,{"className":"mt-5 font-heading text-3xl uppercase leading-none tracking-[0.06em] sm:text-4xl","children":"Fall 2025"}],["$","p",null,{"className":"mt-2 text-sm font-medium text-accent","children":"Prototyping and partner pilots."}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Technical spikes (multi-agent harnesses, verifiable credentials, geospatial pipelines) tested against real partner needs."}],["$","ul",null,{"className":"mt-5 space-y-2 text-sm text-foreground/85","children":[["$","li","8 researchers",{"className":"flex gap-2.5","children":[["$","span",null,{"aria-hidden":true,"className":"mt-1 text-accent","children":"◦"}],["$","span",null,{"children":"8 researchers"}]]}],["$","li","Forced Labor Structural Risk Index",{"className":"flex gap-2.5","children":[["$","span",null,{"aria-hidden":true,"className":"mt-1 text-accent","children":"◦"}],["$","span",null,{"children":"Forced Labor Structural Risk Index"}]]}],["$","li","AI Research Question Assistant",{"className":"flex gap-2.5","children":[["$","span",null,{"aria-hidden":true,"className":"mt-1 text-accent","children":"◦"}],["$","span",null,{"children":"AI Research Question Assistant"}]]}]]}],["$","div",null,{"className":"mt-auto pt-6","children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Full archive coming soon"}],["$","div",null,{"className":"mt-3 flex flex-wrap gap-3","children":[["$","$L11",null,{"href":"/portfolio","className":"inline-flex items-center gap-2 rounded-full bg-accent px-4 py-2 text-sm font-semibold text-background transition-opacity hover:opacity-90","children":"Portfolio →"}],["$","$L11",null,{"href":"/team#alumni-fall-2025","className":"inline-flex items-center gap-2 rounded-full border border-border px-4 py-2 text-sm font-medium transition-colors hover:border-accent hover:text-accent","children":"Meet the cohort →"}]]}]]}]]}]}] 29:["$","$L26","01",{"max":7,"children":["$","article",null,{"className":"flex h-full flex-col rounded-2xl border bg-card p-7 transition-colors border-border hover:border-foreground/25","children":[["$","div",null,{"className":"flex items-center justify-end","children":["$","span",null,{"className":"rounded-full border border-border px-2.5 py-0.5 text-xs text-muted","children":"Past"}]}],["$","h3",null,{"className":"mt-5 font-heading text-3xl uppercase leading-none tracking-[0.06em] sm:text-4xl","children":"Spring 2025"}],["$","p",null,{"className":"mt-2 text-sm font-medium text-accent","children":"First applied research projects."}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"The lab's first applied research projects, spanning online safety, sustainability, AI's footprint, and generative storytelling."}],["$","ul",null,{"className":"mt-5 space-y-2 text-sm text-foreground/85","children":[["$","li","8 researchers",{"className":"flex gap-2.5","children":[["$","span",null,{"aria-hidden":true,"className":"mt-1 text-accent","children":"◦"}],["$","span",null,{"children":"8 researchers"}]]}],["$","li","Online Grooming Prevention",{"className":"flex gap-2.5","children":[["$","span",null,{"aria-hidden":true,"className":"mt-1 text-accent","children":"◦"}],["$","span",null,{"children":"Online Grooming Prevention"}]]}],["$","li","ESG Labels & Certificates Transparency",{"className":"flex gap-2.5","children":[["$","span",null,{"aria-hidden":true,"className":"mt-1 text-accent","children":"◦"}],["$","span",null,{"children":"ESG Labels & Certificates Transparency"}]]}],["$","li","AI's Carbon Footprint",{"className":"flex gap-2.5","children":[["$","span",null,{"aria-hidden":true,"className":"mt-1 text-accent","children":"◦"}],["$","span",null,{"children":"AI's Carbon Footprint"}]]}],["$","li","Generative AI for Good — Avatar Storytelling",{"className":"flex gap-2.5","children":[["$","span",null,{"aria-hidden":true,"className":"mt-1 text-accent","children":"◦"}],["$","span",null,{"children":"Generative AI for Good — Avatar Storytelling"}]]}]]}],["$","div",null,{"className":"mt-auto pt-6","children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Full archive coming soon"}],["$","div",null,{"className":"mt-3 flex flex-wrap gap-3","children":[["$","$L11",null,{"href":"/portfolio","className":"inline-flex items-center gap-2 rounded-full bg-accent px-4 py-2 text-sm font-semibold text-background transition-opacity hover:opacity-90","children":"Portfolio →"}],["$","$L11",null,{"href":"/team#alumni-spring-2025","className":"inline-flex items-center gap-2 rounded-full border border-border px-4 py-2 text-sm font-medium transition-colors hover:border-accent hover:text-accent","children":"Meet the cohort →"}]]}]]}]]}]}] 1f:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -2d:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +2d:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 22:[["$","title","0",{"children":"NYU Ethical Tech CoLab · Emerging Tech, Human Condition"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","link","2",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","3",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","4",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L2d","5",{}]] 2c:null diff --git a/static-site/media/__next._full.txt b/static-site/media/__next._full.txt index 96701cfee..474f5f516 100644 --- a/static-site/media/__next._full.txt +++ b/static-site/media/__next._full.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -11:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +11:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","media",""],"q":"","i":false,"f":[[["",{"children":["media",{"children":["__PAGE__",{}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{},null,false,null]},null,false,"$@f"]},null,false,null],"$L10",false]],"m":"$undefined","G":["$11",["$L12"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -13:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -14:I[82987,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/2qhy58lt3wg92.js"],"Reveal"] -15:I[18620,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/2qhy58lt3wg92.js"],"SectionTabs"] -26:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -28:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +0:{"P":null,"c":["","media",""],"q":"","i":false,"f":[[["",{"children":["media",{"children":["__PAGE__",{}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{},null,false,null]},null,false,"$@f"]},null,false,null],"$L10",false]],"m":"$undefined","G":["$11",["$L12"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +13:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +14:I[82987,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/2qhy58lt3wg92.js"],"Reveal"] +15:I[18620,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/2qhy58lt3wg92.js"],"SectionTabs"] +26:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +28:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 29:"$Sreact.suspense" 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L13",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L13",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L13",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] @@ -29,8 +29,8 @@ e:["$","$1","c",{"children":[[["$","section",null,{"className":"relative overflo 25:[] f:"$W25" 10:["$","$1","h",{"children":[null,["$","$L26",null,{"children":"$L27"}],["$","div",null,{"hidden":true,"children":["$","$L28",null,{"children":["$","$29",null,{"name":"Next.Metadata","children":"$L2a"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -12:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] -33:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +12:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +33:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 16:["$","span",null,{"className":"absolute inset-x-0 bottom-0 bg-gradient-to-t from-black/70 to-transparent p-2 text-[11px] text-white opacity-0 transition-opacity group-hover:opacity-100","children":"Speaker · Deborah Berebichez"}] 17:["$","a","/summit/speaker-ruchira-gupta.jpg",{"href":"/website/summit/speaker-ruchira-gupta.jpg","target":"_blank","rel":"noopener noreferrer","className":"group relative block aspect-[4/3] overflow-hidden rounded-xl border border-border bg-secondary","children":[["$","img",null,{"src":"/website/summit/speaker-ruchira-gupta.jpg","alt":"Speaker · Ruchira Gupta","loading":"lazy","className":"h-full w-full object-cover transition-transform duration-300 group-hover:scale-105"}],["$","span",null,{"className":"absolute inset-x-0 bottom-0 bg-gradient-to-t from-black/70 to-transparent p-2 text-[11px] text-white opacity-0 transition-opacity group-hover:opacity-100","children":"Speaker · Ruchira Gupta"}]]}] 18:["$","a","/summit/ai-researcher.jpg",{"href":"/website/summit/ai-researcher.jpg","target":"_blank","rel":"noopener noreferrer","className":"group relative block aspect-[4/3] overflow-hidden rounded-xl border border-border bg-secondary","children":[["$","img",null,{"src":"/website/summit/ai-researcher.jpg","alt":"AI Researcher track","loading":"lazy","className":"h-full w-full object-cover transition-transform duration-300 group-hover:scale-105"}],["$","span",null,{"className":"absolute inset-x-0 bottom-0 bg-gradient-to-t from-black/70 to-transparent p-2 text-[11px] text-white opacity-0 transition-opacity group-hover:opacity-100","children":"AI Researcher track"}]]}] @@ -55,6 +55,6 @@ f:"$W25" 31:["$","$L14","NY Tech Week Side Event",{"children":["$","article",null,{"className":"card-glow flex h-full flex-col rounded-2xl border border-border bg-card p-7 transition-colors hover:border-border-strong","children":[["$","p",null,{"className":"font-mono text-xs text-accent","children":"June 2024 · NYC"}],["$","h3",null,{"className":"mt-3 font-heading text-xl uppercase leading-snug tracking-wide sm:text-2xl","children":"NY Tech Week Side Event"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Ethical Tech CoLab gathering during NY Tech Week, bringing students and partners together around the CoLab's emerging-technology and human-rights work."}],["$","div",null,{"className":"mt-auto flex flex-wrap gap-2 pt-4","children":[["$","span","NY Tech Week",{"className":"rounded-full border border-border px-2.5 py-0.5 text-[10px] uppercase tracking-wider text-muted","children":"NY Tech Week"}],["$","span","Side Event",{"className":"rounded-full border border-border px-2.5 py-0.5 text-[10px] uppercase tracking-wider text-muted","children":"Side Event"}]]}]]}]}] 32:["$","$L14","Inaugural Summit, Emerging Tech & Ethical Sourcing",{"children":["$","article",null,{"className":"card-glow flex h-full flex-col rounded-2xl border border-border bg-card p-7 transition-colors hover:border-border-strong","children":[["$","p",null,{"className":"font-mono text-xs text-accent","children":"Spring 2024 · NYC"}],["$","h3",null,{"className":"mt-3 font-heading text-xl uppercase leading-snug tracking-wide sm:text-2xl","children":"Inaugural Summit, Emerging Tech & Ethical Sourcing"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"First convening of the Ethical Tech CoLab; framing session on how engineering practice can inform sourcing policy, and how policy can inform engineering boundaries."}],["$","div",null,{"className":"mt-auto flex flex-wrap gap-2 pt-4","children":[["$","span","Inaugural",{"className":"rounded-full border border-border px-2.5 py-0.5 text-[10px] uppercase tracking-wider text-muted","children":"Inaugural"}],["$","span","Curriculum",{"className":"rounded-full border border-border px-2.5 py-0.5 text-[10px] uppercase tracking-wider text-muted","children":"Curriculum"}]]}]]}]}] 27:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -35:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +35:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 2a:[["$","title","0",{"children":"Media · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"The Ethical Tech Summit — a semesterly convening at the intersection of emerging technology, human-rights policy, and global affairs. Past summits, partners, and moments from the room."}],["$","link","2",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","3",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","4",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L35","5",{}]] 34:null diff --git a/static-site/media/__next._head.txt b/static-site/media/__next._head.txt index de193a853..a67ad5029 100644 --- a/static-site/media/__next._head.txt +++ b/static-site/media/__next._head.txt @@ -1,6 +1,6 @@ 1:"$Sreact.fragment" -2:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -3:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +2:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +3:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 4:"$Sreact.suspense" -5:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +5:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 0:{"rsc":["$","$1","h",{"children":[null,["$","$L2",null,{"children":[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]]}],["$","div",null,{"hidden":true,"children":["$","$L3",null,{"children":["$","$4",null,{"name":"Next.Metadata","children":[["$","title","0",{"children":"Media · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"The Ethical Tech Summit — a semesterly convening at the intersection of emerging technology, human-rights policy, and global affairs. Past summits, partners, and moments from the room."}],["$","link","2",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","3",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","4",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L5","5",{}]]}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/media/__next._index.txt b/static-site/media/__next._index.txt index f98a4fd90..18b9d82b6 100644 --- a/static-site/media/__next._index.txt +++ b/static-site/media/__next._index.txt @@ -1,13 +1,13 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -d:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +d:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} 8:["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$Ld",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$Ld",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$Ld",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$Ld",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] diff --git a/static-site/media/__next._tree.txt b/static-site/media/__next._tree.txt index 38ca82631..fa4a04cea 100644 --- a/static-site/media/__next._tree.txt +++ b/static-site/media/__next._tree.txt @@ -1,7 +1,7 @@ -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] 0:{"tree":{"name":"","param":null,"prefetchHints":16,"slots":{"children":{"name":"media","param":null,"prefetchHints":0,"slots":{"children":{"name":"__PAGE__","param":null,"prefetchHints":0,"slots":null}}}}},"staleTime":300,"buildId":"static-snapshot"} diff --git a/static-site/media/__next.media.txt b/static-site/media/__next.media.txt index e17ca4ded..e8bfa323c 100644 --- a/static-site/media/__next.media.txt +++ b/static-site/media/__next.media.txt @@ -1,5 +1,5 @@ 1:"$Sreact.fragment" -2:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -3:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +2:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +3:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] 4:[] 0:{"rsc":["$","$1","c",{"children":[null,["$","$L2",null,{"parallelRouterKey":"children","template":["$","$L3",null,{}]}]]}],"isPartial":false,"staleTime":300,"varyParams":"$W4","buildId":"static-snapshot"} diff --git a/static-site/media/index.html b/static-site/media/index.html index ddf16d0e6..1b788b8d4 100644 --- a/static-site/media/index.html +++ b/static-site/media/index.html @@ -1 +1 @@ -Media · NYU Ethical Tech CoLab

Media · Ethical Tech Summit

The Summit, in the room.

Each semester the Ethical Tech CoLab convenes students, technologists, and policy practitioners for a Summit and Hackathon on the live boundary between emerging technology and global affairs.

Recurring threads: agentic systems and the future of work; verifiable supply chains under UFLPA / NDAA; online grooming and platform safety; forced labor and human trafficking; and micropayment rails for vulnerable populations and creator economies.

Media & moments

A visual sampler from past Summits and Hackathons — panels, student demos, partner roundtables, and the convening floor. Click any image to open the full-size version or linked demo.

Avatar Storytelling demo dashboardAvatar Storytelling demo dashboardSpeaker · Amanda LindseySpeaker · Amanda LindseySpeaker · Deborah BerebichezSpeaker · Deborah BerebichezSpeaker · Ruchira GuptaSpeaker · Ruchira GuptaAI Researcher trackAI Researcher trackSummit floorSummit floorWorking sessionWorking sessionPanel discussionPanel discussionStudent demoStudent demoHackathonHackathonConveningConveningSummit momentSummit momentSummit momentSummit moment

Past events

Spring 2026 · NYC

Agentic Commerce & the Human Condition

Student-led applied work on x402, agent-to-agent micropayments, and downstream effects on vulnerable populations and creator economies. Industry roundtable on responsible deployment.

x402Agentic Commerce

Fall 2025 · NYC

Verifiable Supply Chains: From Compliance to Trust

UFLPA / NDAA practitioner panel; demos of attestation patterns over the Supplychain Graph; student briefs on traceability gaps in critical-mineral pathways.

UFLPATraceability

Sept 2025 · NYC

NY UN Climate Week Side Event

Ethical Tech CoLab side event during NY UN Climate Week, connecting the CoLab's emerging-technology and human-rights work to the climate agenda.

Climate WeekSide Event

June 2025 · NYC

NY Tech Week Side Event

Ethical Tech CoLab gathering during NY Tech Week, convening technologists, students, and partners across the CoLab's emerging-technology and ethical-sourcing work.

NY Tech WeekSide Event

Spring 2025 · NYC

Online Grooming, Platform Safety & Policy

Joint session with multilateral and civil-society partners; student demos on detection patterns and trauma-informed design; closing panel on platform accountability.

Platform SafetyOSCE/ODIHR

Fall 2024 · NYC

Forced Labor, Structural Risk & Applied Research

Launch of student-led applied-research projects on forced-labor structural risk; introduction of the Forced Labor Structural Risk Index methodology.

Human RightsFLSRI

Sept 2024 · NYC

NY UN Climate Week Side Event

Ethical Tech CoLab side event during NY UN Climate Week, bringing the CoLab's technologists and policy practitioners into the climate conversation.

Climate WeekSide Event

June 2024 · NYC

NY Tech Week Side Event

Ethical Tech CoLab gathering during NY Tech Week, bringing students and partners together around the CoLab's emerging-technology and human-rights work.

NY Tech WeekSide Event

Spring 2024 · NYC

Inaugural Summit, Emerging Tech & Ethical Sourcing

First convening of the Ethical Tech CoLab; framing session on how engineering practice can inform sourcing policy, and how policy can inform engineering boundaries.

InauguralCurriculum

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file +Media · NYU Ethical Tech CoLab

Media · Ethical Tech Summit

The Summit, in the room.

Each semester the Ethical Tech CoLab convenes students, technologists, and policy practitioners for a Summit and Hackathon on the live boundary between emerging technology and global affairs.

Recurring threads: agentic systems and the future of work; verifiable supply chains under UFLPA / NDAA; online grooming and platform safety; forced labor and human trafficking; and micropayment rails for vulnerable populations and creator economies.

Media & moments

A visual sampler from past Summits and Hackathons — panels, student demos, partner roundtables, and the convening floor. Click any image to open the full-size version or linked demo.

Avatar Storytelling demo dashboardAvatar Storytelling demo dashboardSpeaker · Amanda LindseySpeaker · Amanda LindseySpeaker · Deborah BerebichezSpeaker · Deborah BerebichezSpeaker · Ruchira GuptaSpeaker · Ruchira GuptaAI Researcher trackAI Researcher trackSummit floorSummit floorWorking sessionWorking sessionPanel discussionPanel discussionStudent demoStudent demoHackathonHackathonConveningConveningSummit momentSummit momentSummit momentSummit moment

Past events

Spring 2026 · NYC

Agentic Commerce & the Human Condition

Student-led applied work on x402, agent-to-agent micropayments, and downstream effects on vulnerable populations and creator economies. Industry roundtable on responsible deployment.

x402Agentic Commerce

Fall 2025 · NYC

Verifiable Supply Chains: From Compliance to Trust

UFLPA / NDAA practitioner panel; demos of attestation patterns over the Supplychain Graph; student briefs on traceability gaps in critical-mineral pathways.

UFLPATraceability

Sept 2025 · NYC

NY UN Climate Week Side Event

Ethical Tech CoLab side event during NY UN Climate Week, connecting the CoLab's emerging-technology and human-rights work to the climate agenda.

Climate WeekSide Event

June 2025 · NYC

NY Tech Week Side Event

Ethical Tech CoLab gathering during NY Tech Week, convening technologists, students, and partners across the CoLab's emerging-technology and ethical-sourcing work.

NY Tech WeekSide Event

Spring 2025 · NYC

Online Grooming, Platform Safety & Policy

Joint session with multilateral and civil-society partners; student demos on detection patterns and trauma-informed design; closing panel on platform accountability.

Platform SafetyOSCE/ODIHR

Fall 2024 · NYC

Forced Labor, Structural Risk & Applied Research

Launch of student-led applied-research projects on forced-labor structural risk; introduction of the Forced Labor Structural Risk Index methodology.

Human RightsFLSRI

Sept 2024 · NYC

NY UN Climate Week Side Event

Ethical Tech CoLab side event during NY UN Climate Week, bringing the CoLab's technologists and policy practitioners into the climate conversation.

Climate WeekSide Event

June 2024 · NYC

NY Tech Week Side Event

Ethical Tech CoLab gathering during NY Tech Week, bringing students and partners together around the CoLab's emerging-technology and human-rights work.

NY Tech WeekSide Event

Spring 2024 · NYC

Inaugural Summit, Emerging Tech & Ethical Sourcing

First convening of the Ethical Tech CoLab; framing session on how engineering practice can inform sourcing policy, and how policy can inform engineering boundaries.

InauguralCurriculum

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file diff --git a/static-site/media/index.txt b/static-site/media/index.txt index 96701cfee..474f5f516 100644 --- a/static-site/media/index.txt +++ b/static-site/media/index.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -11:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +11:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","media",""],"q":"","i":false,"f":[[["",{"children":["media",{"children":["__PAGE__",{}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{},null,false,null]},null,false,"$@f"]},null,false,null],"$L10",false]],"m":"$undefined","G":["$11",["$L12"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -13:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -14:I[82987,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/2qhy58lt3wg92.js"],"Reveal"] -15:I[18620,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/2qhy58lt3wg92.js"],"SectionTabs"] -26:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -28:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +0:{"P":null,"c":["","media",""],"q":"","i":false,"f":[[["",{"children":["media",{"children":["__PAGE__",{}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{},null,false,null]},null,false,"$@f"]},null,false,null],"$L10",false]],"m":"$undefined","G":["$11",["$L12"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +13:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +14:I[82987,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/2qhy58lt3wg92.js"],"Reveal"] +15:I[18620,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/2qhy58lt3wg92.js"],"SectionTabs"] +26:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +28:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 29:"$Sreact.suspense" 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L13",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L13",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L13",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] @@ -29,8 +29,8 @@ e:["$","$1","c",{"children":[[["$","section",null,{"className":"relative overflo 25:[] f:"$W25" 10:["$","$1","h",{"children":[null,["$","$L26",null,{"children":"$L27"}],["$","div",null,{"hidden":true,"children":["$","$L28",null,{"children":["$","$29",null,{"name":"Next.Metadata","children":"$L2a"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -12:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] -33:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +12:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +33:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 16:["$","span",null,{"className":"absolute inset-x-0 bottom-0 bg-gradient-to-t from-black/70 to-transparent p-2 text-[11px] text-white opacity-0 transition-opacity group-hover:opacity-100","children":"Speaker · Deborah Berebichez"}] 17:["$","a","/summit/speaker-ruchira-gupta.jpg",{"href":"/website/summit/speaker-ruchira-gupta.jpg","target":"_blank","rel":"noopener noreferrer","className":"group relative block aspect-[4/3] overflow-hidden rounded-xl border border-border bg-secondary","children":[["$","img",null,{"src":"/website/summit/speaker-ruchira-gupta.jpg","alt":"Speaker · Ruchira Gupta","loading":"lazy","className":"h-full w-full object-cover transition-transform duration-300 group-hover:scale-105"}],["$","span",null,{"className":"absolute inset-x-0 bottom-0 bg-gradient-to-t from-black/70 to-transparent p-2 text-[11px] text-white opacity-0 transition-opacity group-hover:opacity-100","children":"Speaker · Ruchira Gupta"}]]}] 18:["$","a","/summit/ai-researcher.jpg",{"href":"/website/summit/ai-researcher.jpg","target":"_blank","rel":"noopener noreferrer","className":"group relative block aspect-[4/3] overflow-hidden rounded-xl border border-border bg-secondary","children":[["$","img",null,{"src":"/website/summit/ai-researcher.jpg","alt":"AI Researcher track","loading":"lazy","className":"h-full w-full object-cover transition-transform duration-300 group-hover:scale-105"}],["$","span",null,{"className":"absolute inset-x-0 bottom-0 bg-gradient-to-t from-black/70 to-transparent p-2 text-[11px] text-white opacity-0 transition-opacity group-hover:opacity-100","children":"AI Researcher track"}]]}] @@ -55,6 +55,6 @@ f:"$W25" 31:["$","$L14","NY Tech Week Side Event",{"children":["$","article",null,{"className":"card-glow flex h-full flex-col rounded-2xl border border-border bg-card p-7 transition-colors hover:border-border-strong","children":[["$","p",null,{"className":"font-mono text-xs text-accent","children":"June 2024 · NYC"}],["$","h3",null,{"className":"mt-3 font-heading text-xl uppercase leading-snug tracking-wide sm:text-2xl","children":"NY Tech Week Side Event"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Ethical Tech CoLab gathering during NY Tech Week, bringing students and partners together around the CoLab's emerging-technology and human-rights work."}],["$","div",null,{"className":"mt-auto flex flex-wrap gap-2 pt-4","children":[["$","span","NY Tech Week",{"className":"rounded-full border border-border px-2.5 py-0.5 text-[10px] uppercase tracking-wider text-muted","children":"NY Tech Week"}],["$","span","Side Event",{"className":"rounded-full border border-border px-2.5 py-0.5 text-[10px] uppercase tracking-wider text-muted","children":"Side Event"}]]}]]}]}] 32:["$","$L14","Inaugural Summit, Emerging Tech & Ethical Sourcing",{"children":["$","article",null,{"className":"card-glow flex h-full flex-col rounded-2xl border border-border bg-card p-7 transition-colors hover:border-border-strong","children":[["$","p",null,{"className":"font-mono text-xs text-accent","children":"Spring 2024 · NYC"}],["$","h3",null,{"className":"mt-3 font-heading text-xl uppercase leading-snug tracking-wide sm:text-2xl","children":"Inaugural Summit, Emerging Tech & Ethical Sourcing"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"First convening of the Ethical Tech CoLab; framing session on how engineering practice can inform sourcing policy, and how policy can inform engineering boundaries."}],["$","div",null,{"className":"mt-auto flex flex-wrap gap-2 pt-4","children":[["$","span","Inaugural",{"className":"rounded-full border border-border px-2.5 py-0.5 text-[10px] uppercase tracking-wider text-muted","children":"Inaugural"}],["$","span","Curriculum",{"className":"rounded-full border border-border px-2.5 py-0.5 text-[10px] uppercase tracking-wider text-muted","children":"Curriculum"}]]}]]}]}] 27:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -35:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +35:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 2a:[["$","title","0",{"children":"Media · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"The Ethical Tech Summit — a semesterly convening at the intersection of emerging technology, human-rights policy, and global affairs. Past summits, partners, and moments from the room."}],["$","link","2",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","3",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","4",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L35","5",{}]] 34:null diff --git a/static-site/newsletter/2026-07/__next._full.txt b/static-site/newsletter/2026-07/__next._full.txt index 8e3647f1d..635eaa486 100644 --- a/static-site/newsletter/2026-07/__next._full.txt +++ b/static-site/newsletter/2026-07/__next._full.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/1300ctii0r7_l.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -12:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/1300ctii0r7_l.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","newsletter","2026-07",""],"q":"","i":false,"f":[[["",{"children":["newsletter",{"children":[["slug","2026-07","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,"$@10"]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -14:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/1300ctii0r7_l.js"],"Link"] -16:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","newsletter","2026-07",""],"q":"","i":false,"f":[[["",{"children":["newsletter",{"children":[["slug","2026-07","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,"$@10"]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/1300ctii0r7_l.js"],"Link"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:"$Sreact.suspense" -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1c:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -29,12 +29,12 @@ f:["$","$1","c",{"children":["$L15",[["$","script","script-0",{"src":"/website/_ 19:[] 10:"$W19" 11:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] -1e:I[82987,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/1300ctii0r7_l.js"],"Reveal"] -1f:I[18620,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/1300ctii0r7_l.js"],"SectionTabs"] -20:I[92234,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/1300ctii0r7_l.js"],"IssueFrame"] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +1e:I[82987,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/1300ctii0r7_l.js"],"Reveal"] +1f:I[18620,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/1300ctii0r7_l.js"],"SectionTabs"] +20:I[92234,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/1300ctii0r7_l.js"],"IssueFrame"] 15:[["$","section",null,{"className":"relative overflow-hidden border-b border-border","children":[["$","span",null,{"className":"aura"}],["$","div",null,{"className":"relative mx-auto max-w-6xl px-6 py-16","children":[["$","$L1e",null,{"children":["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Newsletter · Monthly Intelligence Brief"}]}],["$","$L1e",null,{"delay":0.05,"children":["$","h1",null,{"className":"mt-4 font-heading text-4xl uppercase leading-[0.95] sm:text-5xl","children":"July 2026 · Edition 01"}]}]]}]]}],["$","$L1f",null,{}],["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-12","children":[["$","$L1e",null,{"children":["$","$L14",null,{"href":"/newsletter","className":"text-sm font-medium text-muted transition-colors hover:text-foreground","children":"← All issues"}]}],["$","div",null,{"className":"mt-6","children":["$","$L20",null,{"src":"/website/newsletter/2026-07.html","title":"Monthly Intelligence Brief — July 2026"}]}]]}]] 1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -21:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +21:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 18:null 1d:[["$","title","0",{"children":"Newsletter · July 2026 · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"HASTE open-sourced for disaster response, the EU AI Act transparency deadline, the Opportunity Board, Tool of the Month, and more."}],["$","link","2",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","3",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","4",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L21","5",{}]] diff --git a/static-site/newsletter/2026-07/__next._head.txt b/static-site/newsletter/2026-07/__next._head.txt index fbc2ba2fa..2d1506f63 100644 --- a/static-site/newsletter/2026-07/__next._head.txt +++ b/static-site/newsletter/2026-07/__next._head.txt @@ -1,6 +1,6 @@ 1:"$Sreact.fragment" -2:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -3:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +2:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +3:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 4:"$Sreact.suspense" -5:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +5:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 0:{"rsc":["$","$1","h",{"children":[null,["$","$L2",null,{"children":[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]]}],["$","div",null,{"hidden":true,"children":["$","$L3",null,{"children":["$","$4",null,{"name":"Next.Metadata","children":[["$","title","0",{"children":"Newsletter · July 2026 · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"HASTE open-sourced for disaster response, the EU AI Act transparency deadline, the Opportunity Board, Tool of the Month, and more."}],["$","link","2",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","3",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","4",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L5","5",{}]]}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/newsletter/2026-07/__next._index.txt b/static-site/newsletter/2026-07/__next._index.txt index c5a39fecb..50ee4b923 100644 --- a/static-site/newsletter/2026-07/__next._index.txt +++ b/static-site/newsletter/2026-07/__next._index.txt @@ -1,13 +1,13 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/1300ctii0r7_l.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -d:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/1300ctii0r7_l.js"],"Link"] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/1300ctii0r7_l.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +d:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/1300ctii0r7_l.js"],"Link"] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} 8:["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$Ld",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$Ld",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$Ld",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$Ld",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] diff --git a/static-site/newsletter/2026-07/__next._tree.txt b/static-site/newsletter/2026-07/__next._tree.txt index 0050ad7d7..8fcac21ba 100644 --- a/static-site/newsletter/2026-07/__next._tree.txt +++ b/static-site/newsletter/2026-07/__next._tree.txt @@ -1,7 +1,7 @@ -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] 0:{"tree":{"name":"","param":null,"prefetchHints":16,"slots":{"children":{"name":"newsletter","param":null,"prefetchHints":0,"slots":{"children":{"name":"slug","param":{"type":"d","key":"2026-07","siblings":null},"prefetchHints":0,"slots":{"children":{"name":"__PAGE__","param":null,"prefetchHints":0,"slots":null}}}}}}},"staleTime":300,"buildId":"static-snapshot"} diff --git a/static-site/newsletter/2026-07/__next.newsletter.txt b/static-site/newsletter/2026-07/__next.newsletter.txt index e17ca4ded..e8bfa323c 100644 --- a/static-site/newsletter/2026-07/__next.newsletter.txt +++ b/static-site/newsletter/2026-07/__next.newsletter.txt @@ -1,5 +1,5 @@ 1:"$Sreact.fragment" -2:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -3:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +2:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +3:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] 4:[] 0:{"rsc":["$","$1","c",{"children":[null,["$","$L2",null,{"parallelRouterKey":"children","template":["$","$L3",null,{}]}]]}],"isPartial":false,"staleTime":300,"varyParams":"$W4","buildId":"static-snapshot"} diff --git a/static-site/newsletter/2026-07/index.html b/static-site/newsletter/2026-07/index.html index 2a60776b2..e0c9c20de 100644 --- a/static-site/newsletter/2026-07/index.html +++ b/static-site/newsletter/2026-07/index.html @@ -1 +1 @@ -Newsletter · July 2026 · NYU Ethical Tech CoLab

Newsletter · Monthly Intelligence Brief

July 2026 · Edition 01

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file +Newsletter · July 2026 · NYU Ethical Tech CoLab

Newsletter · Monthly Intelligence Brief

July 2026 · Edition 01

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file diff --git a/static-site/newsletter/2026-07/index.txt b/static-site/newsletter/2026-07/index.txt index 8e3647f1d..635eaa486 100644 --- a/static-site/newsletter/2026-07/index.txt +++ b/static-site/newsletter/2026-07/index.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/1300ctii0r7_l.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -12:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/1300ctii0r7_l.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","newsletter","2026-07",""],"q":"","i":false,"f":[[["",{"children":["newsletter",{"children":[["slug","2026-07","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,"$@10"]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -14:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/1300ctii0r7_l.js"],"Link"] -16:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","newsletter","2026-07",""],"q":"","i":false,"f":[[["",{"children":["newsletter",{"children":[["slug","2026-07","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,"$@10"]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/1300ctii0r7_l.js"],"Link"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:"$Sreact.suspense" -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1c:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -29,12 +29,12 @@ f:["$","$1","c",{"children":["$L15",[["$","script","script-0",{"src":"/website/_ 19:[] 10:"$W19" 11:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] -1e:I[82987,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/1300ctii0r7_l.js"],"Reveal"] -1f:I[18620,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/1300ctii0r7_l.js"],"SectionTabs"] -20:I[92234,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/1300ctii0r7_l.js"],"IssueFrame"] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +1e:I[82987,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/1300ctii0r7_l.js"],"Reveal"] +1f:I[18620,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/1300ctii0r7_l.js"],"SectionTabs"] +20:I[92234,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/1300ctii0r7_l.js"],"IssueFrame"] 15:[["$","section",null,{"className":"relative overflow-hidden border-b border-border","children":[["$","span",null,{"className":"aura"}],["$","div",null,{"className":"relative mx-auto max-w-6xl px-6 py-16","children":[["$","$L1e",null,{"children":["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Newsletter · Monthly Intelligence Brief"}]}],["$","$L1e",null,{"delay":0.05,"children":["$","h1",null,{"className":"mt-4 font-heading text-4xl uppercase leading-[0.95] sm:text-5xl","children":"July 2026 · Edition 01"}]}]]}]]}],["$","$L1f",null,{}],["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-12","children":[["$","$L1e",null,{"children":["$","$L14",null,{"href":"/newsletter","className":"text-sm font-medium text-muted transition-colors hover:text-foreground","children":"← All issues"}]}],["$","div",null,{"className":"mt-6","children":["$","$L20",null,{"src":"/website/newsletter/2026-07.html","title":"Monthly Intelligence Brief — July 2026"}]}]]}]] 1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -21:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +21:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 18:null 1d:[["$","title","0",{"children":"Newsletter · July 2026 · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"HASTE open-sourced for disaster response, the EU AI Act transparency deadline, the Opportunity Board, Tool of the Month, and more."}],["$","link","2",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","3",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","4",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L21","5",{}]] diff --git a/static-site/newsletter/__next._full.txt b/static-site/newsletter/__next._full.txt index cda34dbed..f5224e159 100644 --- a/static-site/newsletter/__next._full.txt +++ b/static-site/newsletter/__next._full.txt @@ -1,25 +1,25 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/2qhy58lt3wg92.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -11:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/2qhy58lt3wg92.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +11:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","newsletter",""],"q":"","i":false,"f":[[["",{"children":["newsletter",{"children":["__PAGE__",{}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{},null,false,null]},null,false,"$@f"]},null,false,null],"$L10",false]],"m":"$undefined","G":["$11",["$L12"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -13:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/2qhy58lt3wg92.js"],"Link"] -14:I[82987,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/2qhy58lt3wg92.js"],"Reveal"] -15:I[18620,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/2qhy58lt3wg92.js"],"SectionTabs"] -16:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","newsletter",""],"q":"","i":false,"f":[[["",{"children":["newsletter",{"children":["__PAGE__",{}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{},null,false,null]},null,false,"$@f"]},null,false,null],"$L10",false]],"m":"$undefined","G":["$11",["$L12"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +13:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/2qhy58lt3wg92.js"],"Link"] +14:I[82987,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/2qhy58lt3wg92.js"],"Reveal"] +15:I[18620,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/2qhy58lt3wg92.js"],"SectionTabs"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:"$Sreact.suspense" -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1c:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L13",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L13",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L13",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L13",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -30,8 +30,8 @@ e:["$","$1","c",{"children":[[["$","section",null,{"className":"relative overflo 19:[] f:"$W19" 10:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -12:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +12:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -1e:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +1e:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 18:null 1d:[["$","title","0",{"children":"Newsletter · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"The Monthly Intelligence Brief — one email a month on technology creating measurable impact for people, communities, and the planet. AI for good, humanitarian innovation, startups, opportunities, and tools."}],["$","link","2",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","3",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","4",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L1e","5",{}]] diff --git a/static-site/newsletter/__next._head.txt b/static-site/newsletter/__next._head.txt index 59fd4e39f..ca3c1f4ec 100644 --- a/static-site/newsletter/__next._head.txt +++ b/static-site/newsletter/__next._head.txt @@ -1,6 +1,6 @@ 1:"$Sreact.fragment" -2:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -3:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +2:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +3:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 4:"$Sreact.suspense" -5:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +5:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 0:{"rsc":["$","$1","h",{"children":[null,["$","$L2",null,{"children":[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]]}],["$","div",null,{"hidden":true,"children":["$","$L3",null,{"children":["$","$4",null,{"name":"Next.Metadata","children":[["$","title","0",{"children":"Newsletter · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"The Monthly Intelligence Brief — one email a month on technology creating measurable impact for people, communities, and the planet. AI for good, humanitarian innovation, startups, opportunities, and tools."}],["$","link","2",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","3",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","4",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L5","5",{}]]}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/newsletter/__next._index.txt b/static-site/newsletter/__next._index.txt index 4f0dfc676..4c740b519 100644 --- a/static-site/newsletter/__next._index.txt +++ b/static-site/newsletter/__next._index.txt @@ -1,13 +1,13 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/2qhy58lt3wg92.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -d:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/2qhy58lt3wg92.js"],"Link"] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/2qhy58lt3wg92.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +d:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/2qhy58lt3wg92.js"],"Link"] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} 8:["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$Ld",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$Ld",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$Ld",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$Ld",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] diff --git a/static-site/newsletter/__next._tree.txt b/static-site/newsletter/__next._tree.txt index 3c9c39489..39804fb12 100644 --- a/static-site/newsletter/__next._tree.txt +++ b/static-site/newsletter/__next._tree.txt @@ -1,7 +1,7 @@ -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] 0:{"tree":{"name":"","param":null,"prefetchHints":16,"slots":{"children":{"name":"newsletter","param":null,"prefetchHints":0,"slots":{"children":{"name":"__PAGE__","param":null,"prefetchHints":0,"slots":null}}}}},"staleTime":300,"buildId":"static-snapshot"} diff --git a/static-site/newsletter/__next.newsletter.txt b/static-site/newsletter/__next.newsletter.txt index e17ca4ded..e8bfa323c 100644 --- a/static-site/newsletter/__next.newsletter.txt +++ b/static-site/newsletter/__next.newsletter.txt @@ -1,5 +1,5 @@ 1:"$Sreact.fragment" -2:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -3:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +2:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +3:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] 4:[] 0:{"rsc":["$","$1","c",{"children":[null,["$","$L2",null,{"parallelRouterKey":"children","template":["$","$L3",null,{}]}]]}],"isPartial":false,"staleTime":300,"varyParams":"$W4","buildId":"static-snapshot"} diff --git a/static-site/newsletter/index.html b/static-site/newsletter/index.html index 0861e1ad5..ccf39095d 100644 --- a/static-site/newsletter/index.html +++ b/static-site/newsletter/index.html @@ -1 +1 @@ -Newsletter · NYU Ethical Tech CoLab

Newsletter · Monthly Intelligence Brief

Technology that creates impact.

One email a month: breakthrough AI for good, humanitarian innovation, exciting startups, research in plain English, opportunities for students and early-career professionals, and tools that make us better researchers. No model-release hype.

Subscribe

Join the list to get each edition in your inbox. Free, one email a month, unsubscribe anytime.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file +Newsletter · NYU Ethical Tech CoLab

Newsletter · Monthly Intelligence Brief

Technology that creates impact.

One email a month: breakthrough AI for good, humanitarian innovation, exciting startups, research in plain English, opportunities for students and early-career professionals, and tools that make us better researchers. No model-release hype.

Subscribe

Join the list to get each edition in your inbox. Free, one email a month, unsubscribe anytime.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file diff --git a/static-site/newsletter/index.txt b/static-site/newsletter/index.txt index cda34dbed..f5224e159 100644 --- a/static-site/newsletter/index.txt +++ b/static-site/newsletter/index.txt @@ -1,25 +1,25 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/2qhy58lt3wg92.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -11:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/2qhy58lt3wg92.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +11:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","newsletter",""],"q":"","i":false,"f":[[["",{"children":["newsletter",{"children":["__PAGE__",{}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{},null,false,null]},null,false,"$@f"]},null,false,null],"$L10",false]],"m":"$undefined","G":["$11",["$L12"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -13:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/2qhy58lt3wg92.js"],"Link"] -14:I[82987,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/2qhy58lt3wg92.js"],"Reveal"] -15:I[18620,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/2qhy58lt3wg92.js"],"SectionTabs"] -16:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","newsletter",""],"q":"","i":false,"f":[[["",{"children":["newsletter",{"children":["__PAGE__",{}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{},null,false,null]},null,false,"$@f"]},null,false,null],"$L10",false]],"m":"$undefined","G":["$11",["$L12"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +13:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/2qhy58lt3wg92.js"],"Link"] +14:I[82987,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/2qhy58lt3wg92.js"],"Reveal"] +15:I[18620,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/2qhy58lt3wg92.js"],"SectionTabs"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:"$Sreact.suspense" -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1c:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L13",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L13",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L13",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L13",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -30,8 +30,8 @@ e:["$","$1","c",{"children":[[["$","section",null,{"className":"relative overflo 19:[] f:"$W19" 10:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -12:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +12:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -1e:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +1e:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 18:null 1d:[["$","title","0",{"children":"Newsletter · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"The Monthly Intelligence Brief — one email a month on technology creating measurable impact for people, communities, and the planet. AI for good, humanitarian innovation, startups, opportunities, and tools."}],["$","link","2",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","3",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","4",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L1e","5",{}]] diff --git a/static-site/portfolio/__next._full.txt b/static-site/portfolio/__next._full.txt index c688273ea..3cc90d92b 100644 --- a/static-site/portfolio/__next._full.txt +++ b/static-site/portfolio/__next._full.txt @@ -1,25 +1,25 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/369wdzv6no3y1.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -11:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/369wdzv6no3y1.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +11:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","portfolio",""],"q":"","i":false,"f":[[["",{"children":["portfolio",{"children":["__PAGE__",{}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{},null,false,null]},null,false,"$@f"]},null,false,null],"$L10",false]],"m":"$undefined","G":["$11",["$L12"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -13:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/369wdzv6no3y1.js"],"Link"] -14:I[85437,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/369wdzv6no3y1.js"],"Image"] -15:I[82987,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/369wdzv6no3y1.js"],"Reveal"] -16:I[18620,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/369wdzv6no3y1.js"],"SectionTabs"] -17:I[67120,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/369wdzv6no3y1.js"],"PortfolioExplorer"] -1f:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -21:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +0:{"P":null,"c":["","portfolio",""],"q":"","i":false,"f":[[["",{"children":["portfolio",{"children":["__PAGE__",{}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{},null,false,null]},null,false,"$@f"]},null,false,null],"$L10",false]],"m":"$undefined","G":["$11",["$L12"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +13:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/369wdzv6no3y1.js"],"Link"] +14:I[85437,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/369wdzv6no3y1.js"],"Image"] +15:I[82987,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/369wdzv6no3y1.js"],"Reveal"] +16:I[18620,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/369wdzv6no3y1.js"],"SectionTabs"] +17:I[67120,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/369wdzv6no3y1.js"],"PortfolioExplorer"] +1f:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +21:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 22:"$Sreact.suspense" 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L13",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L13",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L13",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] @@ -31,8 +31,8 @@ e:["$","$1","c",{"children":[[["$","section",null,{"className":"relative overflo 1e:[] f:"$W1e" 10:["$","$1","h",{"children":[null,["$","$L1f",null,{"children":"$L20"}],["$","div",null,{"hidden":true,"children":["$","$L21",null,{"children":["$","$22",null,{"name":"Next.Metadata","children":"$L23"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -12:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] -27:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +12:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +27:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 18:["$","span",null,{"aria-hidden":true,"children":"→"}] 19:["$","a",null,{"href":"https://github.com/Ethical-Tech-CoLab/forced-labor-structural-risk-index","target":"_blank","rel":"noopener noreferrer","className":"inline-flex items-center gap-1 text-sm font-medium text-accent transition-opacity hover:opacity-80","children":["View code ",["$","span",null,{"aria-hidden":true,"children":"↗"}]]}] 1a:["$","article","AI Research Question Assistant",{"className":"flex flex-col rounded-2xl border border-border bg-background p-7","children":[["$","h4",null,{"className":"font-heading text-2xl uppercase leading-[1.12] tracking-[0.02em] sm:text-3xl","children":"AI Research Question Assistant"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"A survey and toolkit for how AI helps researchers formulate rigorous, original research questions — finding gaps, generating candidate questions, summarizing the state of the art, and flagging contradictions, with guardrails against hallucinated citations. Ships with a reusable Copilot 'Researcher' prompt and a journal-credibility rubric."}],["$","div",null,{"className":"mb-6 mt-5 flex flex-wrap gap-2","children":[["$","span","LLM",{"className":"rounded-full border border-border px-2.5 py-0.5 text-xs text-muted","children":["#","LLM"]}],["$","span","Research",{"className":"rounded-full border border-border px-2.5 py-0.5 text-xs text-muted","children":["#","Research"]}],["$","span","Tooling",{"className":"rounded-full border border-border px-2.5 py-0.5 text-xs text-muted","children":["#","Tooling"]}]]}],["$","div",null,{"className":"mt-auto flex flex-wrap items-center gap-x-4 gap-y-2 border-t border-border pt-5","children":[["$","a",null,{"href":"https://ethical-tech-colab.github.io/ai-research-question-assistant/","target":"_blank","rel":"noopener noreferrer","className":"btn-sweep inline-flex items-center gap-2 rounded-full bg-accent px-4 py-2 text-sm font-semibold text-accent-ink transition-transform hover:scale-[1.03]","children":"▶ Launch live demo"}],"$undefined",["$","$L13",null,{"href":"/publications/ai-research-assistant","className":"inline-flex items-center gap-1 text-sm font-medium text-accent transition-opacity hover:opacity-80","children":["Read the report ",["$","span",null,{"aria-hidden":true,"children":"→"}]]}],["$","a",null,{"href":"https://github.com/Ethical-Tech-CoLab/ai-research-question-assistant","target":"_blank","rel":"noopener noreferrer","className":"inline-flex items-center gap-1 text-sm font-medium text-accent transition-opacity hover:opacity-80","children":["View code ",["$","span",null,{"aria-hidden":true,"children":"↗"}]]}]]}]]}] @@ -43,6 +43,6 @@ f:"$W1e" 25:["$","a",null,{"href":"https://github.com/Ethical-Tech-CoLab/ai-carbon-footprint","target":"_blank","rel":"noopener noreferrer","className":"inline-flex items-center gap-1 text-sm font-medium text-accent transition-opacity hover:opacity-80","children":["View code ",["$","span",null,{"aria-hidden":true,"children":"↗"}]]}] 26:["$","article","Generative AI for Good — Avatar Storytelling",{"className":"flex flex-col rounded-2xl border border-border bg-background p-7","children":[["$","h4",null,{"className":"font-heading text-2xl uppercase leading-[1.12] tracking-[0.02em] sm:text-3xl","children":"Generative AI for Good — Avatar Storytelling"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Human-condition storytelling with culturally grounded digital-human avatars — short pieces produced with generative media."}],["$","div",null,{"className":"mb-6 mt-5 flex flex-wrap gap-2","children":[["$","span","Generative AI",{"className":"rounded-full border border-border px-2.5 py-0.5 text-xs text-muted","children":["#","GenerativeAI"]}],["$","span","Storytelling",{"className":"rounded-full border border-border px-2.5 py-0.5 text-xs text-muted","children":["#","Storytelling"]}],["$","span","Media",{"className":"rounded-full border border-border px-2.5 py-0.5 text-xs text-muted","children":["#","Media"]}]]}],["$","div",null,{"className":"mt-auto flex flex-wrap items-center gap-x-4 gap-y-2 border-t border-border pt-5","children":[["$","a",null,{"href":"https://ethical-tech-colab.github.io/Avatar-Impact-Stories/","target":"_blank","rel":"noopener noreferrer","className":"btn-sweep inline-flex items-center gap-2 rounded-full bg-accent px-4 py-2 text-sm font-semibold text-accent-ink transition-transform hover:scale-[1.03]","children":"▶ Launch live demo"}],"$undefined","$undefined",["$","a",null,{"href":"https://github.com/Ethical-Tech-CoLab/Avatar-Impact-Stories","target":"_blank","rel":"noopener noreferrer","className":"inline-flex items-center gap-1 text-sm font-medium text-accent transition-opacity hover:opacity-80","children":["View code ",["$","span",null,{"aria-hidden":true,"children":"↗"}]]}]]}]]}] 20:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -29:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +29:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 23:[["$","title","0",{"children":"Portfolio · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"Research questions the current cohort is exploring across disaster response, cultural heritage, supply chains, and diplomacy."}],["$","link","2",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","3",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","4",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L29","5",{}]] 28:null diff --git a/static-site/portfolio/__next._head.txt b/static-site/portfolio/__next._head.txt index e9a59c02b..e02a68b37 100644 --- a/static-site/portfolio/__next._head.txt +++ b/static-site/portfolio/__next._head.txt @@ -1,6 +1,6 @@ 1:"$Sreact.fragment" -2:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -3:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +2:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +3:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 4:"$Sreact.suspense" -5:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +5:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 0:{"rsc":["$","$1","h",{"children":[null,["$","$L2",null,{"children":[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]]}],["$","div",null,{"hidden":true,"children":["$","$L3",null,{"children":["$","$4",null,{"name":"Next.Metadata","children":[["$","title","0",{"children":"Portfolio · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"Research questions the current cohort is exploring across disaster response, cultural heritage, supply chains, and diplomacy."}],["$","link","2",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","3",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","4",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L5","5",{}]]}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/portfolio/__next._index.txt b/static-site/portfolio/__next._index.txt index b8da2b48d..ffad54bc8 100644 --- a/static-site/portfolio/__next._index.txt +++ b/static-site/portfolio/__next._index.txt @@ -1,13 +1,13 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/369wdzv6no3y1.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -d:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/369wdzv6no3y1.js"],"Link"] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/369wdzv6no3y1.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +d:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/369wdzv6no3y1.js"],"Link"] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} 8:["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$Ld",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$Ld",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$Ld",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$Ld",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] diff --git a/static-site/portfolio/__next._tree.txt b/static-site/portfolio/__next._tree.txt index c219f312a..bec7aecf7 100644 --- a/static-site/portfolio/__next._tree.txt +++ b/static-site/portfolio/__next._tree.txt @@ -1,7 +1,7 @@ -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] 0:{"tree":{"name":"","param":null,"prefetchHints":16,"slots":{"children":{"name":"portfolio","param":null,"prefetchHints":0,"slots":{"children":{"name":"__PAGE__","param":null,"prefetchHints":0,"slots":null}}}}},"staleTime":300,"buildId":"static-snapshot"} diff --git a/static-site/portfolio/__next.portfolio.txt b/static-site/portfolio/__next.portfolio.txt index e17ca4ded..e8bfa323c 100644 --- a/static-site/portfolio/__next.portfolio.txt +++ b/static-site/portfolio/__next.portfolio.txt @@ -1,5 +1,5 @@ 1:"$Sreact.fragment" -2:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -3:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +2:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +3:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] 4:[] 0:{"rsc":["$","$1","c",{"children":[null,["$","$L2",null,{"parallelRouterKey":"children","template":["$","$L3",null,{}]}]]}],"isPartial":false,"staleTime":300,"varyParams":"$W4","buildId":"static-snapshot"} diff --git a/static-site/portfolio/index.html b/static-site/portfolio/index.html index ff1a9187c..008cc0281 100644 --- a/static-site/portfolio/index.html +++ b/static-site/portfolio/index.html @@ -1 +1 @@ -Portfolio · NYU Ethical Tech CoLab

Portfolio · Current cohort

Four questions. One frontier.

The current cohort is exploring research questions across disaster response, cultural heritage, supply chains, and diplomacy. Open a question to see the projects exploring it, or filter by topic.

Filter
Information Quality · By ZoneINDEXEIIQUALITY · ACCESS · EQUITY
Provenance · Verified ChainORIGINTRANSFERCUSTODYRETURN
Auditable Traceability · Origin → ShelfMATERIALSLABORIMPACTSOURCEPROCESSTRANSPORTSHELF
Scenario · Multi-Agent RehearsalAGENT AAGENT BMEDIATOROUTCOMES

Archive

Previous portfolios.

Projects from earlier cohorts, grouped by the year they were worked on, with their live demos, code, and reports alongside.

Fall 2025

2 projects

Forced Labor Structural Risk Index

An interactive index mapping the structural conditions that enable forced labor across 184 countries on a 0–1 risk scale, with national and sub-national layers.

#Traceability#Humanrights#Data
▶ Launch live demoRead the report View code

AI Research Question Assistant

A survey and toolkit for how AI helps researchers formulate rigorous, original research questions — finding gaps, generating candidate questions, summarizing the state of the art, and flagging contradictions, with guardrails against hallucinated citations. Ships with a reusable Copilot 'Researcher' prompt and a journal-credibility rubric.

#LLM#Research#Tooling
▶ Launch live demoRead the report View code

Spring 2025

4 projects

Online Grooming Prevention

Repurposing existing detection technology to identify and interrupt grooming behavior in online spaces, protecting minors on platforms where predators operate.

#Onlinesafety#Childprotection#NLP

ESG Labels & Certificates Transparency

Standardizing ESG certificates and labels into a comparable, trustworthy signal so consumers can steer their behavior toward genuinely sustainable products.

#Sustainability#Consumertrust#Standardization

Generative AI for Good — Avatar Storytelling

Human-condition storytelling with culturally grounded digital-human avatars — short pieces produced with generative media.

#GenerativeAI#Storytelling#Media
▶ Launch live demoView code

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file +Portfolio · NYU Ethical Tech CoLab

Portfolio · Current cohort

Four questions. One frontier.

The current cohort is exploring research questions across disaster response, cultural heritage, supply chains, and diplomacy. Open a question to see the projects exploring it, or filter by topic.

Filter
Information Quality · By ZoneINDEXEIIQUALITY · ACCESS · EQUITY
Provenance · Verified ChainORIGINTRANSFERCUSTODYRETURN
Auditable Traceability · Origin → ShelfMATERIALSLABORIMPACTSOURCEPROCESSTRANSPORTSHELF
Scenario · Multi-Agent RehearsalAGENT AAGENT BMEDIATOROUTCOMES

Archive

Previous portfolios.

Projects from earlier cohorts, grouped by the year they were worked on, with their live demos, code, and reports alongside.

Fall 2025

2 projects

Forced Labor Structural Risk Index

An interactive index mapping the structural conditions that enable forced labor across 184 countries on a 0–1 risk scale, with national and sub-national layers.

#Traceability#Humanrights#Data
▶ Launch live demoRead the report View code

AI Research Question Assistant

A survey and toolkit for how AI helps researchers formulate rigorous, original research questions — finding gaps, generating candidate questions, summarizing the state of the art, and flagging contradictions, with guardrails against hallucinated citations. Ships with a reusable Copilot 'Researcher' prompt and a journal-credibility rubric.

#LLM#Research#Tooling
▶ Launch live demoRead the report View code

Spring 2025

4 projects

Online Grooming Prevention

Repurposing existing detection technology to identify and interrupt grooming behavior in online spaces, protecting minors on platforms where predators operate.

#Onlinesafety#Childprotection#NLP

ESG Labels & Certificates Transparency

Standardizing ESG certificates and labels into a comparable, trustworthy signal so consumers can steer their behavior toward genuinely sustainable products.

#Sustainability#Consumertrust#Standardization

Generative AI for Good — Avatar Storytelling

Human-condition storytelling with culturally grounded digital-human avatars — short pieces produced with generative media.

#GenerativeAI#Storytelling#Media
▶ Launch live demoView code

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file diff --git a/static-site/portfolio/index.txt b/static-site/portfolio/index.txt index c688273ea..3cc90d92b 100644 --- a/static-site/portfolio/index.txt +++ b/static-site/portfolio/index.txt @@ -1,25 +1,25 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/369wdzv6no3y1.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -11:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/369wdzv6no3y1.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +11:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","portfolio",""],"q":"","i":false,"f":[[["",{"children":["portfolio",{"children":["__PAGE__",{}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{},null,false,null]},null,false,"$@f"]},null,false,null],"$L10",false]],"m":"$undefined","G":["$11",["$L12"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -13:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/369wdzv6no3y1.js"],"Link"] -14:I[85437,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/369wdzv6no3y1.js"],"Image"] -15:I[82987,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/369wdzv6no3y1.js"],"Reveal"] -16:I[18620,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/369wdzv6no3y1.js"],"SectionTabs"] -17:I[67120,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/369wdzv6no3y1.js"],"PortfolioExplorer"] -1f:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -21:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +0:{"P":null,"c":["","portfolio",""],"q":"","i":false,"f":[[["",{"children":["portfolio",{"children":["__PAGE__",{}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{},null,false,null]},null,false,"$@f"]},null,false,null],"$L10",false]],"m":"$undefined","G":["$11",["$L12"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +13:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/369wdzv6no3y1.js"],"Link"] +14:I[85437,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/369wdzv6no3y1.js"],"Image"] +15:I[82987,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/369wdzv6no3y1.js"],"Reveal"] +16:I[18620,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/369wdzv6no3y1.js"],"SectionTabs"] +17:I[67120,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/369wdzv6no3y1.js"],"PortfolioExplorer"] +1f:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +21:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 22:"$Sreact.suspense" 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L13",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L13",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L13",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] @@ -31,8 +31,8 @@ e:["$","$1","c",{"children":[[["$","section",null,{"className":"relative overflo 1e:[] f:"$W1e" 10:["$","$1","h",{"children":[null,["$","$L1f",null,{"children":"$L20"}],["$","div",null,{"hidden":true,"children":["$","$L21",null,{"children":["$","$22",null,{"name":"Next.Metadata","children":"$L23"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -12:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] -27:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +12:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +27:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 18:["$","span",null,{"aria-hidden":true,"children":"→"}] 19:["$","a",null,{"href":"https://github.com/Ethical-Tech-CoLab/forced-labor-structural-risk-index","target":"_blank","rel":"noopener noreferrer","className":"inline-flex items-center gap-1 text-sm font-medium text-accent transition-opacity hover:opacity-80","children":["View code ",["$","span",null,{"aria-hidden":true,"children":"↗"}]]}] 1a:["$","article","AI Research Question Assistant",{"className":"flex flex-col rounded-2xl border border-border bg-background p-7","children":[["$","h4",null,{"className":"font-heading text-2xl uppercase leading-[1.12] tracking-[0.02em] sm:text-3xl","children":"AI Research Question Assistant"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"A survey and toolkit for how AI helps researchers formulate rigorous, original research questions — finding gaps, generating candidate questions, summarizing the state of the art, and flagging contradictions, with guardrails against hallucinated citations. Ships with a reusable Copilot 'Researcher' prompt and a journal-credibility rubric."}],["$","div",null,{"className":"mb-6 mt-5 flex flex-wrap gap-2","children":[["$","span","LLM",{"className":"rounded-full border border-border px-2.5 py-0.5 text-xs text-muted","children":["#","LLM"]}],["$","span","Research",{"className":"rounded-full border border-border px-2.5 py-0.5 text-xs text-muted","children":["#","Research"]}],["$","span","Tooling",{"className":"rounded-full border border-border px-2.5 py-0.5 text-xs text-muted","children":["#","Tooling"]}]]}],["$","div",null,{"className":"mt-auto flex flex-wrap items-center gap-x-4 gap-y-2 border-t border-border pt-5","children":[["$","a",null,{"href":"https://ethical-tech-colab.github.io/ai-research-question-assistant/","target":"_blank","rel":"noopener noreferrer","className":"btn-sweep inline-flex items-center gap-2 rounded-full bg-accent px-4 py-2 text-sm font-semibold text-accent-ink transition-transform hover:scale-[1.03]","children":"▶ Launch live demo"}],"$undefined",["$","$L13",null,{"href":"/publications/ai-research-assistant","className":"inline-flex items-center gap-1 text-sm font-medium text-accent transition-opacity hover:opacity-80","children":["Read the report ",["$","span",null,{"aria-hidden":true,"children":"→"}]]}],["$","a",null,{"href":"https://github.com/Ethical-Tech-CoLab/ai-research-question-assistant","target":"_blank","rel":"noopener noreferrer","className":"inline-flex items-center gap-1 text-sm font-medium text-accent transition-opacity hover:opacity-80","children":["View code ",["$","span",null,{"aria-hidden":true,"children":"↗"}]]}]]}]]}] @@ -43,6 +43,6 @@ f:"$W1e" 25:["$","a",null,{"href":"https://github.com/Ethical-Tech-CoLab/ai-carbon-footprint","target":"_blank","rel":"noopener noreferrer","className":"inline-flex items-center gap-1 text-sm font-medium text-accent transition-opacity hover:opacity-80","children":["View code ",["$","span",null,{"aria-hidden":true,"children":"↗"}]]}] 26:["$","article","Generative AI for Good — Avatar Storytelling",{"className":"flex flex-col rounded-2xl border border-border bg-background p-7","children":[["$","h4",null,{"className":"font-heading text-2xl uppercase leading-[1.12] tracking-[0.02em] sm:text-3xl","children":"Generative AI for Good — Avatar Storytelling"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Human-condition storytelling with culturally grounded digital-human avatars — short pieces produced with generative media."}],["$","div",null,{"className":"mb-6 mt-5 flex flex-wrap gap-2","children":[["$","span","Generative AI",{"className":"rounded-full border border-border px-2.5 py-0.5 text-xs text-muted","children":["#","GenerativeAI"]}],["$","span","Storytelling",{"className":"rounded-full border border-border px-2.5 py-0.5 text-xs text-muted","children":["#","Storytelling"]}],["$","span","Media",{"className":"rounded-full border border-border px-2.5 py-0.5 text-xs text-muted","children":["#","Media"]}]]}],["$","div",null,{"className":"mt-auto flex flex-wrap items-center gap-x-4 gap-y-2 border-t border-border pt-5","children":[["$","a",null,{"href":"https://ethical-tech-colab.github.io/Avatar-Impact-Stories/","target":"_blank","rel":"noopener noreferrer","className":"btn-sweep inline-flex items-center gap-2 rounded-full bg-accent px-4 py-2 text-sm font-semibold text-accent-ink transition-transform hover:scale-[1.03]","children":"▶ Launch live demo"}],"$undefined","$undefined",["$","a",null,{"href":"https://github.com/Ethical-Tech-CoLab/Avatar-Impact-Stories","target":"_blank","rel":"noopener noreferrer","className":"inline-flex items-center gap-1 text-sm font-medium text-accent transition-opacity hover:opacity-80","children":["View code ",["$","span",null,{"aria-hidden":true,"children":"↗"}]]}]]}]]}] 20:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -29:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +29:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 23:[["$","title","0",{"children":"Portfolio · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"Research questions the current cohort is exploring across disaster response, cultural heritage, supply chains, and diplomacy."}],["$","link","2",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","3",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","4",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L29","5",{}]] 28:null diff --git a/static-site/print/after-the-corridor/__next._full.txt b/static-site/print/after-the-corridor/__next._full.txt index f15d580a2..ecd2a5788 100644 --- a/static-site/print/after-the-corridor/__next._full.txt +++ b/static-site/print/after-the-corridor/__next._full.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -12:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","print","after-the-corridor",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","after-the-corridor","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -14:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -16:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","print","after-the-corridor",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","after-the-corridor","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:"$Sreact.suspense" -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1c:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -29,7 +29,7 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 19:[] 10:"$W19" 11:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1e:Ta14, @page { size: 210mm 297mm; margin: 20mm 18mm; } @@ -197,6 +197,6 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 72:["$","li","36",{"children":[["$","span",null,{"className":"print-ref-num","children":"37"}],["$","span",null,{"children":"Innovations for Poverty Action, Displaced Livelihoods Initiative: Call for Proposals, Round V (IPA, 2026); IPA, DLI Eligibility Self-Assessment and Budget Template, Round V (2026)."}]]}] 73:["$","p",null,{"className":"print-note","children":"The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings are those of the researchers and do not represent the official positions of New York University, Microsoft, UNHCR, WFP, or any partner institution. External programs cited are referenced as evidence, not as CoLab partnerships."}] 1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -74:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +74:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 18:null 1d:[["$","title","0",{"children":"After the Corridor (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L74","6",{}]] diff --git a/static-site/print/after-the-corridor/__next._head.txt b/static-site/print/after-the-corridor/__next._head.txt index 2c10658e9..a9667e18b 100644 --- a/static-site/print/after-the-corridor/__next._head.txt +++ b/static-site/print/after-the-corridor/__next._head.txt @@ -1,6 +1,6 @@ 1:"$Sreact.fragment" -2:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -3:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +2:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +3:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 4:"$Sreact.suspense" -5:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +5:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 0:{"rsc":["$","$1","h",{"children":[null,["$","$L2",null,{"children":[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]]}],["$","div",null,{"hidden":true,"children":["$","$L3",null,{"children":["$","$4",null,{"name":"Next.Metadata","children":[["$","title","0",{"children":"After the Corridor (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L5","6",{}]]}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/print/after-the-corridor/__next._index.txt b/static-site/print/after-the-corridor/__next._index.txt index f98a4fd90..18b9d82b6 100644 --- a/static-site/print/after-the-corridor/__next._index.txt +++ b/static-site/print/after-the-corridor/__next._index.txt @@ -1,13 +1,13 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -d:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +d:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} 8:["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$Ld",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$Ld",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$Ld",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$Ld",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] diff --git a/static-site/print/after-the-corridor/__next._tree.txt b/static-site/print/after-the-corridor/__next._tree.txt index 9e05aab03..c8741a6bf 100644 --- a/static-site/print/after-the-corridor/__next._tree.txt +++ b/static-site/print/after-the-corridor/__next._tree.txt @@ -1,7 +1,7 @@ -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] 0:{"tree":{"name":"","param":null,"prefetchHints":16,"slots":{"children":{"name":"print","param":null,"prefetchHints":0,"slots":{"children":{"name":"slug","param":{"type":"d","key":"after-the-corridor","siblings":null},"prefetchHints":0,"slots":{"children":{"name":"__PAGE__","param":null,"prefetchHints":0,"slots":null}}}}}}},"staleTime":300,"buildId":"static-snapshot"} diff --git a/static-site/print/after-the-corridor/__next.print.txt b/static-site/print/after-the-corridor/__next.print.txt index 460afe242..fa617e997 100644 --- a/static-site/print/after-the-corridor/__next.print.txt +++ b/static-site/print/after-the-corridor/__next.print.txt @@ -1,4 +1,4 @@ 1:"$Sreact.fragment" -2:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -3:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +2:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +3:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] 0:{"rsc":["$","$1","c",{"children":[null,[["$","style",null,{"children":"\n /* The site is dark-first, and a dark color-scheme makes Chrome paint\n the printed sheet's margin area dark too. Paper is light. */\n html { color-scheme: light !important; }\n html, body { background: #ffffff !important; }\n body::after { content: none !important; }\n header, footer, .site-bg, .aura { display: none !important; }\n "}],["$","$L2",null,{"parallelRouterKey":"children","template":["$","$L3",null,{}]}]]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/print/after-the-corridor/index.html b/static-site/print/after-the-corridor/index.html index 926408fbb..ef553a7be 100644 --- a/static-site/print/after-the-corridor/index.html +++ b/static-site/print/after-the-corridor/index.html @@ -1,4 +1,4 @@ -After the Corridor (print edition) · NYU Ethical Tech CoLab

Publications · Academic report

After the Corridor

From AI-Informed Evacuation to Digital Public Goods for Refugee Economic Inclusion

Ethical Tech CoLab

NYU Center for Global Affairs

July 2026

Carolina de Almeida Pernambuco Moron, Christine Lumen, Alana Robertson, Melanie MacKew, Carlos D. Ruiz, India Clarke, and Yago Rocha, with Teresa Cantero (PhD Candidate, Universidad Carlos III de Madrid; Adjunct Professor, IE University; Visiting Scholar, NYU CGA). Faculty directors: Prof. Yorke E. Rhodes III and Sylvia Maier.

How can AI inform evacuation decisions, and what happens after? This report examines the CoLab's five fielded evacuation prototypes, their limitations and research frontiers, and extends the arc to the camp: the measurement, financial-modeling, and rights infrastructure that can shorten protracted displacement and improve lives. Dzaleka Refugee Camp in Malawi is the case study ground.

Key figures

6

built instruments: five fielded evacuation prototypes and a completed forced-labour risk index, all open source

53k to 60k

people now at Dzaleka, a camp built in 1994 for 10,000 to 12,000

90%

cut to UNHCR's Malawi budget, from about US$8 million to about US$1 million

0

biometric identity systems: the category the CoLab refuses to build, recorded as a versioned decision pattern

01

Executive Summary

The Ethical Tech CoLab's evacuation portfolio was built to answer a protection question: how can and should AI tools be used to enhance the protection and evacuation of civilians during armed conflict, and in what ways do they support or challenge existing obligations under International Humanitarian Law?

Across five fielded prototypes, a crisis-information index, a unified risk platform, a calibrated cost-and-risk decision framework, an agent-based behavior model, and a readiness-under-uncertainty simulator, one finding recurred: the binding constraint on protection is rarely the absence of a route or a resource, but the quality, accessibility, and equity of the information people and responders act on. IHL frames evacuation as a last-resort duty; the technology's job is to make that duty implementable under uncertainty.

What happens after the evacuation? A successful evacuation typically ends in a camp, where the modal outcome is years of protracted displacement. Using Dzaleka Refugee Camp, Malawi, where humanitarian financing has collapsed while an encampment policy bars residents from work, we show that the same measurement discipline, financial modeling, and uncertainty methods that served evacuation convert directly into durable-solutions infrastructure: an open Livelihoods Information and Access Index, a resettlement-and-inclusion financial model built on the CoLab's operational evacuation cost engine, privacy-preserving cash and credential rails, and a civil-society legal-assistance map. We also state what we will not build, biometric identity for displaced people, and why that refusal is itself a public good.

Most of the world's refugees live in protracted situations, and the modal outcome of reaching safety is years, sometimes decades, inside a camp or settlement. The legal architecture that governs what comes next is the Geneva refugee-protection consensus: the 1951 Convention relating to the Status of Refugees and its 1967 Protocol, the doctrine of durable solutions that UNHCR administers, and the 2018 Global Compact on Refugees. That consensus recognizes three ways a displacement can end well, voluntary repatriation, local integration, and resettlement, and each of them is gated far more by information and legal status than by physical capacity.

The CoLab's built work already reaches beyond the evacuation corridor. The same measurement discipline produced the Forced Labor Structural Risk Index, a completed instrument that maps the structural risk of forced labour across countries and sub-national regions, the exploitation that displacement, encampment, and the loss of the right to work make more likely. Read together, the evacuation models and the risk index share one purpose: to improve the information on which protection depends, and through it to reduce forced labour and the vulnerabilities that displaced people carry.

This report keeps three registers distinct throughout: what is built and validated, what is built but not yet validated, and what is proposed. No claim in these pages exceeds the register of the tool it describes.

The ask, in one paragraph. The Ethical Tech CoLab seeks grant funding to develop this technology infrastructure work as open, portable public goods, to improve the lives of displaced communities. It would be validated first at Dzaleka with the field partner Fraternidade Sem Fronteiras, and we are open to other partners with the same mission-driven motivations. Our intent is developing tools and releasing them to the public so that any research or practitioner team can adopt them. One possibility is the Displaced Livelihoods Initiative run by Innovations for Poverty Action and J-PAL, in its Infrastructure and Public Goods track. The report is written to align with that submission and to stand on its own for any peer funder or academic reviewer.

02

The Research Question, and Why It Runs Past the Corridor

The protection question and the accounting gap. The portfolio's organizing question comes from Teresa Cantero's doctoral research at the Instituto de Derechos Humanos Gregorio Peces-Barba, Universidad Carlos III de Madrid: how can and should artificial intelligence be used to enhance the protection and evacuation of civilians during armed conflict, and in what ways does it support or challenge the implementation of existing obligations under international humanitarian law.

Under that law, evacuation is not discretionary. When the security of the population or imperative military reasons demand it, moving civilians out of danger is a binding duty under Article 49 of the Fourth Geneva Convention, Articles 57 and 58 of Additional Protocol I, and Article 17 of Additional Protocol II. A movement into which people are coerced is forcible transfer, a war crime regardless of motive. The duty runs wider than the battlefield: the doctrine of the responsibility to protect holds that where a state cannot or will not shield its population from mass atrocity, that responsibility passes to the international community, which frames protection as an obligation rather than a matter of goodwill.

Two structural facts motivate the program. The first is a protection gap. States move quickly to adopt AI for targeting and operational advantage, yet rarely operationalize computational models to protect civilians or to make an evacuation decision legible under the incomplete, fast-changing information those decisions face. The second is an accounting gap. Displacement is tracked in detail, on the order of 117 million people forcibly displaced worldwide by 2025, yet there is no global count of evacuated civilians, because evacuation is fragmented across militaries, governments, humanitarian agencies, volunteers, and civilians acting on their own.

The legal and institutional architecture. The 1951 Convention and its 1967 Protocol define who is a refugee and fix the cornerstone duty of non-refoulement, that no one may be returned to a place where they face persecution. The 2018 Global Compact on Refugees adds a cooperative framework organized around four objectives: easing pressure on host countries, enhancing refugee self-reliance, expanding access to third-country solutions, and supporting conditions for return in safety and dignity. International humanitarian law governs conduct in armed conflict, and the responsibility to protect frames the state's duty toward its own population.

The institutions matter as much as the instruments. UNHCR leads protection and durable solutions for refugees; the International Organization for Migration is the United Nations migration agency; the Office for the Coordination of Humanitarian Affairs coordinates emergency response through the cluster approach; and the Office of the High Commissioner for Human Rights anchors the broader human-rights instruments that apply regardless of a person's status. One structural limitation shapes everything downstream: many host states are not parties to the 1951 Convention or maintain significant reservations, so the protection a displaced person actually receives depends heavily on where they land.

The three durable solutions. The arc does not end when the shooting stops. An evacuation succeeds when a person reaches relative safety, and safety, in practice, usually means a camp. There the analytical problem inverts. The question is no longer whether to move and how, but how to leave well now that one has arrived.

The three durable solutions recognized by the refugee-protection regime, voluntary repatriation to the country of origin, local integration in the country of asylum, and resettlement to a third country, are the recognized routes out. Each is gated by information and legal status as much as by physical or economic capacity, a point the regime's own scholarship makes plainly. Resettlement reaches only a small fraction of those who need it, and complementary pathways such as labour or study routes remain modest in scale, which is why local integration and the conditions for safe return carry most of the weight in practice. This report treats the durable-solutions phase as the natural continuation of the evacuation research question, and asks what open measurement and modeling infrastructure would help displaced people and the organizations serving them find the fastest defensible route out.

03

The Built Portfolio

This section describes what the CoLab has already built, and is candid about how mature each piece is, because a reader deciding whether to fund the work is owed that precision. Three registers are kept distinct. The Evacuation Risk and Cost Framework is built and validated in its evacuation form. The Evacuation Information Index, the Exodus platform, and the two simulators are built and operational but not yet expert-validated or calibrated. The Forced Labor Structural Risk Index is built and completed, documented in a methodology page and an associated publication. The livelihoods instruments the report proposes, the Livelihoods Information and Access Index and the Resettlement and Inclusion Capacity Simulator, together with the value-transfer and legal-map builds, are proposed, and their validation is future work. All are open source.

PrototypeFunctionCore method
Evacuation Information Index (EII)Scores quality, accessibility, and equity of crisis information by geographyComposite index on INFORM, ACLED, and FSI logic; live news and conflict feeds
Exodus, Civilian Evacuation Risk PlatformUnifies EII, the cost model, and endangerment and feasibility assessmentFastAPI backend; shared calculators and data layer
ERCF, Evacuation Risk and Cost FrameworkEstimates human and financial cost of evacuating against stayingSeven-dimension scoring; calibrated cost model; break-even analysis
Evacuation Behavior SimulatorModels how information and demographics shape who leaves, and whenAgent-based model; probabilistic status lifecycle
Readiness and Uncertainty SimulatorModels how intelligence uncertainty degrades destination choicesMonte Carlo; gatekeeper factors; uncertainty slider
The five fielded evacuation prototypes. All are open source at github.com/Ethical-Tech-CoLab.

The Evacuation Information Index. The index treats crisis information itself as the object of measurement. It scores three families of factors, impact, conditions, and complexity, on a scale of 1 to 5, using weights synthesized from the INFORM Severity Index, the Armed Conflict Location and Event Data project, and the Fragile States Index. A live backend adds current news and a conflict timeline for each crisis, and a map layer can overlay AI-derived building-damage and route-damage tiles from Microsoft HASTE and Planet imagery. The design is deliberately transparent: weights are labeled provisional pending expert validation, and where the hosted site cannot run the live backend, responses are baked into a clearly labeled static snapshot so that nothing is presented as live when it is not.

The index has real limits, and the report states them rather than hiding them.

  • Weights are borrowed by analogy, not yet expert-validated or sensitivity-tested; a reviewer will read pre-validation weights as pre-evidence.
  • Source and language bias: feeds skew to English-language, internationally indexed reporting, so the populations with the worst information access are least visible to the instrument that scores information access. The index must correct for this, not inherit it.
  • The damage overlay is currently self-hosted with no public interface, so for most users the combined information-and-damage view is empty until it is connected.

Future research on the index. Run the three-phase validation the CoLab's methodology work already specifies, budget-allocation and Delphi elicitation with 8 to 12 domain experts, Fuzzy-AHP on contested weights, then historical case calibration, and add an equity layer measuring who receives information, disaggregated by gender, age, nationality, and residence inside or outside a camp. That equity layer is the move that turns the index into the livelihoods instrument described later in this report.

Exodus, the integration layer. Exodus unifies the crisis map, the cost-and-scenario model, and an endangerment-and-feasibility assessment behind a single backend, design language, and interface, including calculators for the cost of staying, the cost of evacuating, resource needs, and scenario management. Integration is not validation. Unifying three research prototypes behind a more authoritative interface inherits their uncalibrated assumptions, which raises rather than lowers the duty to label confidence honestly. The same interface is the seam where the durable-solutions module plugs in. Its stated limitation is that it is not automatically updated; its future research is to validate the approaches, weights, and measurements.

The Evacuation Risk and Cost Framework. Now at version 7.2, the framework is the CoLab's calibrated decision-support tool for the human and financial cost of civilian evacuation, and it is the one component of the portfolio that is validated rather than provisional, in its evacuation form. It scores scenarios across seven dimensions, kinetic threat, mobility, authorization, logistics, destination, urgency, and information, estimates evacuation resources and in-situ assistance costs, and produces a break-even analysis: the day after which a one-time evacuation becomes cheaper than the mounting daily cost of sustaining a population in place. Its cost parameters are sourced to humanitarian standards and individually tagged as Validated, Estimated, or Unvalidated, drawing on UNHCR water and sanitation guidance, the Sphere standards, and World Food Programme figures.

The mortality model is disclosed as indicative rather than predictive, with a reported log-log fit of 0.855, a leave-one-out cross-validation of 0.807, and a documented case where the structure fails. The financial estimates are substantially more reliable than the mortality figures, which is why the durable-solutions adaptation leads with cost. The framework's ethical architecture is explicit: it does not place a monetary value on a human life. Its figures are logistics estimates, its labels are descriptive rather than prescriptive, and it presents data to support a decision rather than issuing one. That framing carries verbatim into every extension in this report.

The framework's limitations are equally explicit.

  • It models the immediate field operation only, on the order of US$134 per person at moderate risk over 50 kilometers by ground. Real operations, Kosovo 1999 at about US$562 per person and Lebanon 2006 at about US$4,500 to US$5,700, add international transport, reception, and hosting. Extending the cost boundary to the full displacement journey, including protracted encampment, is the frontier the durable-solutions model takes up.
  • Ground-truthing against OCHA's Financial Tracking Service is on the backlog and belongs in the funded work plan.

The Evacuation Behavior Simulator. An agent-based model of how information spread and demographics shape who leaves and when. Agents move through a lifecycle from unaware to seeking, milling, evacuating, and done, with family clusters that wait for all members, confirmation loops whose speed depends on how clear the information is, neighbor influence, and calibrated delays for elders, young children, pregnant individuals, and unaccompanied minors. It is pedagogical rather than predictive, and its household-composition logic transfers directly to modeling who is left behind by information failures inside a camp. Its behavior may vary by society, culture, location, personality, income, and education, and the future research is to build intersectionality between behavior simulations.

The Readiness and Uncertainty Simulator. A seeded, fully reproducible instrument that models how uncertainty in field intelligence degrades an evacuation decision. Destinations are scored across seven factors, 500 Monte Carlo trials per group-and-destination pair convert readiness into a distribution of success probabilities, and an uncertainty control degrades every factor's confidence at once, so a user can watch predicted outcomes collapse while nothing on the ground has changed.

Two design choices carry into the rest of this report. First, three factors are non-substitutable gatekeepers, security, the consent of the local authority, and the willingness of the host community, any of which caps readiness: a destination that is dangerous, legally closed, or unwelcoming cannot be redeemed by good logistics. Second, the model insists on the difference between a factor that is unknown, a gap that better information can close, and a factor that is a confirmed refusal, a settled political exclusion. Treating the two as the same would make an unresolvable refusal look like a solvable data problem. Structurally, assigning groups to destinations under consent, willingness, security, and capacity constraints, in the presence of uncertainty, is resettlement matching, which is the transfer the durable-solutions model makes. The simulator has not yet been stress tested with different AI models.

The Forced Labor Structural Risk Index. The index applies the same composite-index method to a different harm. It is a completed, interactive index that maps the structural conditions enabling forced labour across 184 countries on a scale from 0 to 1, with national and sub-national layers, so that the risk can be seen and compared rather than assumed. It belongs in this report because displacement is a forced-labour risk multiplier: the loss of legal status and the right to work, the collapse of assistance, and weakened oversight are precisely the conditions under which exploitation grows, and forced labour is a defined harm in international law with an estimated 27.6 million people subject to it worldwide.

Its significance for the durable-solutions argument is conceptual: measuring structural forced-labour risk and measuring livelihoods-information access are two halves of one protective picture, one showing where exploitation is likely, the other showing whether the information that could prevent it is reaching people. As a composite index its weights remain open to expert re-elicitation, it is not automatically updated, and sponsorship mechanisms remain unsourced at country scale while an eight-country hand-coded kafala signal feeds two domains in the shipped build. Writing an ILO-endorsed kafala passage on top of that unresolved discrepancy would have overstated the index's grounding.

Cross-cutting synthesis. Read together, the built prototypes yield a short set of lessons. Information access is the binding constraint, and it is the thread that carries from the corridor into the camp. Uncertainty has to be modeled explicitly, with the unknown kept separate from the confirmed, or the analysis misleads by omission. We address gaps in data by generating synthetic data that mimics real-life data, and address the residual uncertainty with Monte Carlo simulation. A composite index built from the sequence of borrowing precedents, eliciting expert weights, and calibrating against cases is portable infrastructure: the same method that scores crisis information also scores structural forced-labour risk. And cost modeling in a context of human lives has to stay descriptive rather than prescriptive, following the framework's refusal to monetize life.

04

After Arrival: Dzaleka and the Second Emergency

A successful evacuation deposits people into a slower emergency. The concept of a protracted refugee situation, conventionally a population of at least 25,000 of the same nationality in exile for five years or more, describes where most refugees actually live, and care-and-maintenance encampment becomes the default by inertia rather than by anyone's plan.

The duration is not a rounding error in a person's life. By UNHCR's own accounting, the average duration of a protracted refugee situation rose from about 9 years in 1993 to roughly 17 years by 2003, and while later analysis rightly cautions that a single average conceals wide variation, the order of magnitude is not in dispute: protracted exile routinely lasts a decade or more, and often a generation. The literature on encampment and self-settlement has long argued that the camp is often less a humanitarian necessity than an administrative and political choice, and that its costs to refugees, in mobility, work, and self-reliance, are systematically understated.

Dzaleka and the funding collapse. Dzaleka Refugee Camp in Dowa District, Malawi, is the sharp end of that pattern. Built in 1994 for 10,000 to 12,000 people, it now holds on the order of 53,000 to 60,000 people, predominantly from the Democratic Republic of the Congo, Burundi, and Rwanda, and it is caught in the funding collapse of 2025 and 2026.

The numbers describe a system in retreat.

  • UNHCR's Malawi budget fell from about US$8 million to about US$1 million, a cut of roughly 90 percent, ending general non-food distributions and laying off protection staff, including paralegals.
  • World Food Programme assistance is secured only through the middle of 2026, with about US$11 million still needed to restore full rations through December.
  • Monthly cash support has fallen toward the range of US$6 to US$9 per person, down from roughly US$100, forcing skipped meals and other negative coping.

Malawi's encampment policy, enforced since March 2023, bars most residents from working outside the camp, so the population cannot substitute self-reliance for collapsing aid. Local civil society, including Inua Advocacy, argues the crisis is inseparable from that policy and is pressing for the right to work; weakened oversight is meanwhile reported to embolden trafficking networks.

This is not a context for piloting speculative technology on vulnerable people. It is a context where the highest-value contribution is measurement, evidence, and legibility: tools that help residents and the organizations serving them see and argue for the fastest route out, and that document the true cost of the status quo. This is a global humanitarian emergency that needs collective collaborative effort, and building infrastructure, digital goods and services, information, training, and capacity might transform generational poverty and increase access to resources.

Encampment as a driver of forced labour. These are the textbook conditions for forced labour. When legal work is barred and assistance collapses, the informal and coerced economy fills the gap. Reducing that exposure is a direct objective of the CoLab's work: the Forced Labor Structural Risk Index names where the risk concentrates, and the livelihoods-information instrument measures whether the information that could protect people, about their rights, their options, and legitimate work, is actually reaching them. That link has no official validation yet, and the future research is to separate recruitment, exploitation, and monetization by intersectionality using social demography, and to deepen the demand side.

05

The Financial Model: From Evacuation Break-Even to Inclusion Break-Even

The financial-modeling engine already exists, validated for one problem and now re-pointed to another, and the distinction matters. The Evacuation Risk and Cost Framework is validated for evacuation costing; its encampment-versus-inclusion parameterization is new and unvalidated, and most of its durable-solutions cost parameters are tagged Estimated or Unvalidated. What transfers cleanly is the architecture and the break-even mathematics, a one-time investment set against a mounting recurring cost, not a set of validated cost figures for the new use.

The engine, re-pointed.

  • Trajectory A prices status-quo encampment as a recurring per-person, per-year assistance cost, covering rations, water and sanitation, health, protection, and administration, using the same Sphere, UNHCR, and WFP parameter families the framework already carries, with the same Validated, Estimated, or Unvalidated tags.
  • Trajectory B prices an inclusion or resettlement pathway as a front-loaded investment, covering documentation, recognition of qualifications, transport, start-up capital, and host co-investment.
  • The output is the break-even year after which the status quo costs more, which is the framework's break-even chart with its axes relabeled.

The Resettlement and Inclusion Capacity Simulator is that engine re-pointed, and re-earning its validity in the durable-solutions domain is the work, not a formality. The capacity-and-willingness layer transfers from the readiness simulator: it maps destinations, whether resettlement states or local-integration pathways, against the non-substitutable gatekeepers of legal consent, host willingness, and security, alongside the substitutable capacities of housing, jobs, and services, while preserving the distinction between an option that is unknown and one that is refused, so that an unlobbied state is never mistaken for a closed door. The uncertainty layer carries over as well, so that thin data on host capacity or a state's posture visibly widens the confidence band and prevents false precision in an advocacy number.

What the evidence predicts. The World Bank's recent work on refugee self-reliance in Sub-Saharan Africa indicates that if refugees had the right to move and work, they would contribute to the hosting country's economy, with real fiscal and economic gains. The honest caveat is built into that same literature. The World Bank finds that self-reliance remains elusive across much of the region, because camps often sit in marginal, high-poverty areas where even full local integration would lift incomes only a little, and the scholarship on self-reliance programming warns that the concept is often invoked to justify the withdrawal of assistance rather than to expand rights.

The simulator has to be able to return the answer that inclusion does not pay here, where host markets are too thin, or it is advocacy rather than analysis. Its output is a defensible, reproducible sentence of the form: sustaining this population at Dzaleka costs approximately this much per year; this pathway costs approximately that much up front and breaks even in a given year, subject to these gatekeepers and this uncertainty band. The simulator could also provide country comparisons across different legal models and measure the economic impact of different refugee policies, and listing the policies of each hosting country would better inform humanitarian organizations and populations.

What the literature says the model will find. The engine formalizes an argument the development-economics evidence already supports.

  • UNHCR estimates that if freedom of movement and the right to work were relaxed so refugees could earn on par with hosts, complementary assistance costs would fall from about US$3.2 billion to roughly US$900 million a year.
  • A joint World Bank and UNHCR benchmark puts the annual cost of bringing every refugee in low and middle income host countries up to the global poverty line at almost US$62 billion in a scenario where refugees earn nothing; refugees' own earnings already cover roughly US$40 billion of that, leaving about US$22 billion.
  • World Bank analysis estimates that in Uganda's inclusive regime, economic participation saves about US$150 per refugee per year, roughly US$225 million across 1.5 million refugees.
  • The same World Bank work finds that around Kakuma, refugee presence raised the host county's gross regional product by about 3.4 percent.
  • It also reports that Colombia's regularization of Venezuelan migrants raised beneficiaries' incomes by about 30 percent and their formal employment by around 10 percentage points, with little displacement of host workers.

06

Digital Public Goods for Economic Inclusion

The toolkit is ordered from the safest build to the one that demands the most caution. The organizing principle is data minimization: build the thinnest system that solves the problem, and refuse the parts whose risks outweigh their benefit for this population. A digital public good, in the sense set out by the Digital Public Goods Alliance and endorsed within the United Nations system, is open-source software, open data, an open standard, or an open model that is relevant to the Sustainable Development Goals, does no harm by design, and, importantly for this population, minimizes the collection of personally identifiable information. The Principles for Digital Development point the same way, toward privacy, reuse, and designing with the user. Measured against those criteria, the CoLab's contributions are intended to qualify as digital public goods rather than to resemble them loosely.

Measurement: the Livelihoods Information and Access Index. The measurement instrument is the Evacuation Information Index re-specified for displaced livelihoods and legal rights: an open, versioned instrument measuring the quality, accessibility, and equity of information about work rights, employment pathways, financial services, and entrepreneurship support reaching displaced populations. The architecture transfers from the evacuation index, but the construct does not, and saying so is part of the method. Measuring information access to work rights is a different thing from scoring crisis severity, so the index has to establish its own construct validity from the ground up rather than inherit it.

In draft form the index is a matrix. Its three dimensions, quality, meaning whether information is accurate, current, and actionable; accessibility, meaning whether people can reach it in a language and channel they use; and equity, meaning whether it reaches sub-populations evenly, are scored across four domains: work rights and legal status, employment pathways, financial services, and entrepreneurship support. Each cell draws on a mix of sources: an audit of what official and organizational information exists, a structured record from field partners of what actually circulates, and, in the future primary study, what residents themselves report receiving.

The equity dimension takes seriously the finding in the forced-migration literature that displacement is gendered, that women, men, and sexual and gender minorities face different risks and different information barriers, so disaggregation is a design requirement rather than an afterthought. Because information inside a camp travels largely by word of mouth, community leaders, and channels that are not in English, the instrument has to weigh informal and in-language sources deliberately, or it will reproduce the very bias it is built to detect. The framework is provisional and is itself a deliverable, with the weights subject to the expert-elicitation pipeline before any score is treated as evidence.

The index does not enter an empty field, and the report says so plainly rather than claiming to be first. The humanitarian sector already assesses information reaching affected populations through accountability-to-affected-populations frameworks, communication-with-communities practice, and information-ecosystem assessments. What those approaches largely lack, and what the index adds, is a composite, versioned, openly standardized measure that renders information access as a comparable variable, disaggregated by sub-population, which a later impact evaluation can treat as an outcome or a mechanism. The contribution is the measure and its portability, not the discovery that information matters. Its deliverables are the index, an open data standard, a reference implementation, and methodology documentation, validated first at Dzaleka with Fraternidade Sem Fronteiras.

The open questions the index carries.

  • One size does not fit all: the geopolitical impact and the limits of building an index and adapting it live, when everything can change.
  • How can local communities own their data governance?
  • What would be the most beneficial digital goods a displaced population needs?
  • How can the information be made accessible in low-connectivity environments?
  • How can marginalized communities and people who do not read be included in the design of digital goods?

Future research on the index. Developing a work card and a system that would allow refugees to record their expertise and be matched with regions or organizations that need those services.

Value transfer: privacy-preserving cash, decoupled from biometrics. The nearest-term practical contribution is moving value more cheaply and reliably as traditional financing collapses. The reference case is the World Food Programme's Building Blocks, the largest blockchain-based humanitarian platform, which settled entitlements on a private, permissioned ledger and paid vendors in consolidated sums, saving roughly 98 percent of transaction costs on an early deployment, improving reconciliation across agencies, and establishing a co-governed network that UN Women later extended to cash-for-skills in refugee camps.

The same case is the cautionary tale. Building Blocks ran on biometric registration, and independent analysis warns that many such pilots functioned as donor showpieces under-invested in privacy, and that even a permissioned ledger can leak sensitive metadata when on-chain identifiers are correlated with off-chain records. The lesson is to decouple the payment rail from biometric identity.

A wallet designed for Dzaleka would therefore:

  • Authenticate with the thinnest possible credential, a personal identification number together with a locally issued token or code, never a biometric.
  • Keep identifying data off the ledger, with only pseudonymous entitlement state recorded on it.
  • Integrate an existing audited, co-governed network rather than build a bespoke ledger, so that the CoLab's contribution is the threat model and the measurement wrapper.
  • Run a data-protection impact assessment with community consultation before any collection.

Portability: skills credentials, never identity. Portability of skills is the third contribution, and it must stop short of identity. The CoLab operates an adjacent architecture that demonstrates the cryptographic pattern, an agent that issues cryptographically signed, tamper-evident verifiable credentials for cultural objects, and Professor Rhodes's prior work with the ID2020 effort concerned decentralized, portable humanitarian credentials with UNHCR and the World Food Programme. The pattern is shared; the risk surface of issuing credentials to displaced people is not, and the report treats that gap as real rather than solved.

The responsible application is narrow: portable, self-held credentials for skills, qualifications, and completed training, so that a nurse or an electrician does not lose a decade of accreditation to displacement, issued to and controlled by the individual and presented selectively. The bright line is firm. A skills credential records what a person can do. It must never become a de facto identity card that gatekeeps aid or is legible to a persecuting authority. The moment it records who someone is, it inherits the risks discussed in the next section and should be stopped.

Rights: mapping civil-society legal assistance. Donor cuts stripped UNHCR's own paralegals out of Dzaleka, which makes a maintained, open directory of who actually provides legal help the safest and highest-leverage build. Organized by legal need, refugee-status determination, work authorization, documentation, protection from gender-based violence, protection from trafficking, and resettlement referral, the map records which civil-society and refugee-led organizations provide assistance, with contact details, language, and eligibility, under the same allowlist-and-citation discipline the provenance agent enforces, so that no unsourced claim ships.

It connects to the index directly. The index shows where rights information fails; the map routes a person to the organization that can close the gap. Local actors such as Inua Advocacy are the anchor nodes, rather than the government. This legal-and-jurisdiction layer, covering work rights, documentation rules, encampment policy, recognition of qualifications, and access to financial services, is a first-class component of the work, coded from public legal sources, and it reflects the long-standing finding in the refugee literature that legal status and the right to work are among the decisive determinants of whether displaced people can build a livelihood at all.

07

Vulnerable Populations, Digital Identity Risk, and Mitigation

The Rohingya precedent. Refugees are an exceptionally vulnerable group whose personal data demands unusual rigor, and the technology that most tempts humanitarian builders, a comprehensive digital identity for displaced people, is the one this team refuses to build. The evidence is documented rather than hypothetical. Human Rights Watch found that biometric and biographic data collected from Rohingya refugees in Bangladesh was shared onward, ultimately reaching Myanmar, the government they had fled, for the purpose of assessing repatriation eligibility, covering some 830,000 people between 2018 and 2021, without the full data-protection impact assessment UNHCR's own policy requires and without meaningful informed consent.

Consent in that system was structurally compromised. Refugees were told that aid depended on their registration, which makes agreement coerced rather than free. Consent materials were commonly presented in English, a language many of those required to register did not read, and reporting indicates that families who refused enrollment were cut off from food and services. The defining property of the harm is its irreversibility. Biometrics are sticky; an iris cannot be reissued once it is compromised, and a database built to facilitate solutions suffered precisely the function creep that critics had predicted, as data gathered for one stated purpose was repurposed for another.

Mitigation is mostly refusal, then minimization, and it can be stated as a small number of firm commitments.

  • Do not build biometric identity. A category refusal, encoded as a versioned decision pattern in the CoLab's open schema, DDC-0001, biometric registration of displaced populations.
  • Decouple entitlement from identity: access gated by the thinnest credential, never by who you are proven biometrically.
  • Data minimization and purpose limitation: least data, single stated purpose, no secondary use or third-party transfer, as a matter of design rather than policy alone.
  • Individual control and local governance: the person holds and selectively presents any credential; refugee-led and local actors govern the data.
  • Real consent or none: if consent cannot be free and informed, whether because aid is conditioned on it or because a language or power gap cannot be bridged, treat it as absent and do not proceed.
  • Mandatory pre-collection data-protection impact assessment and community consultation: the omitted step whose absence produced the Rohingya harms.

For a funder, this refusal is an asset. A responsible-AI lab that can name the one thing it will not build, and cite exactly why, is more credible than one that promises everything.

08

Aggregating Evidence So Displaced People Can Argue Their Case Sooner

Taken together, these instruments become an evidence engine that attacks the core information asymmetry facing displaced people, the fact that the rights-and-options case that could change an individual's situation today requires specialized legal and economic knowledge that the individual does not have. The index shows, with disaggregation, which sub-populations are systematically failed by rights-and-pathways information, which is evidence rather than anecdote. The financial model reframes the ask from a plea for help into a costed comparison, that prolonged encampment costs a given amount per year while a given pathway costs less by a given year, which is the argument that donors and states respond to. The legal map converts the evidence into action, routing people to organizations that can file, refer, and litigate.

One limit has to be stated plainly, because the evidence engine cannot dissolve it. Durable solutions are rationed by state consent, resettlement quotas, and political will, and the readiness model's own logic holds that a confirmed refusal is a settled exclusion that no measurement moves. Better information and a sharper cost argument can widen the set of options that are seriously considered and can equip advocates who would otherwise argue from anecdote, but where the binding constraint is political rather than informational, rigorous evidence may change nothing. The impact claims in this report are scoped to that reality: the instruments improve the quality and legibility of the case, which is necessary but not sufficient for a different outcome.

The outputs are designed to be wielded by refugee-led organizations and civil-society advocates, and, where they will use them, by durable-solutions units within UNHCR, rather than to sit in a dashboard. This design choice reflects a clear direction in the field, the growing insistence, visible across the refugee-studies and practitioner literature on knowledge, voice, and power and on refugee self-representation, that displaced people should hold and deploy the evidence about their own situations rather than be the objects of it.

Two guardrails hold throughout. Every advocacy artifact carries its uncertainty band honestly, and the aggregation layer never becomes a registry of identifiable individuals. It argues about populations and gaps, not named people.

09

What Infrastructure Is Actually Needed

Two things are absent by design: any biometric or identity system, and any bespoke ledger. Everything proposed is either a measurement or knowledge artifact that the CoLab is suited to own, or a thin integration on audited external rails. That is the correct risk posture for a small lab working with a highly vulnerable population, and it follows directly from the digital-public-goods and data-minimization commitments set out above.

LayerComponentReusesPosture
MeasurementLIAI index, open data standard, reference implementation, methodologyEII and the composite-index methodBuild, the funding ask
EconomicsRICS: encampment-versus-inclusion break-even; capacity and willingness mapERCF engine (operational) and readiness gatekeepersAdapt, engine in place
Value transferPrivacy-preserving cash wallet, biometric-freeExternal audited network; CoLab adds threat model and measurementIntegrate, do not rebuild
PortabilitySelf-held skills and qualification credentialsProvenance verifiable-credential architecture; ID2020 lineageBuild narrowly, skills only
RightsOpen civil-society legal-assistance mapProvenance sourcing disciplineBuild, safest and high leverage
GovernanceDDC decision-pattern schema; DPIA and ethics layerDDC-0001Build, the credibility spine
The proposed stack, and what each layer reuses from the built portfolio.

10

How Success Will Be Measured

A funder should not have to take this program on faith, so success is defined in advance in three tiers: outputs, whether the instruments shipped and are rigorous; outcomes, whether they work and are used; and impact, whether they changed decisions. The measurement instruments themselves double as the evidence base. One line governs the whole framework and resolves an ambiguity a careful reader would otherwise catch: under this grant, every activity involving residents is consultation and co-design of the instruments, together with a desk audit and the field partner's structured knowledge, and no personal data is collected from residents. The measurement of residents' own information access, which would require primary data collection, is a separate future study.

Outputs: did we ship, and is it rigorous? Over the first year, success means a first open-source release of the index, comprising the index itself, an open data standard, a reference implementation, and methodology documentation, each versioned and citable. It means completed and published validation of the instrument's design, with expert elicitation across 8 to 12 experts over two rounds, reported weights, consistency ratios within accepted thresholds, and divergence analysis, including negative results. It means the reference-context validation executed with Fraternidade Sem Fronteiras, drawing on an audit of existing information and the partner's structured knowledge of what circulates, using open and synthetic data. It means the financial model parameterized for Dzaleka on the operational cost engine, with every parameter tagged and the full parameter table public. And it means the civil-society legal map live with the large majority of entries verified against primary sources, and the biometric-refusal decision pattern published with its governance schema.

Full psychometric validation of the instrument on responses from the displaced population itself, covering test-retest reliability, construct validity, and measurement invariance across subgroups, requires primary data collection from that population and is not part of this phase. It is written up here as a published protocol and pre-analysis plan for a future primary study, so that the instrument and its validation plan are themselves a public good, and so that a later study can execute the plan without re-deriving it. The longitudinal read on whether information-access scores for the worst-served sub-populations improve over time belongs to that same future study. This is the honest boundary of the present work, and stating it plainly is part of the method.

Outcomes: does it work, and is it used? Over the first two years, success means concordance between the instrument and the field partner's operational knowledge of where information gaps sit at Dzaleka, and equity disaggregation that reproduces independently observed disparities. It means adoption, with at least one external research team taking up the index as an instrument and at least one practitioner organization using its scores to retarget outreach. It means the financial model cited in at least two advocacy or policy documents by refugee-led or civil-society organizations, with the uncertainty bands reproduced intact, since fidelity of use matters and not only volume. It means community legitimacy, with refugee community members engaged as co-designers and a community review session signing off on how findings are represented before publication. And it means open-source health, with public contribution metrics, quarterly releases, and a documented path to a second site.

Impact: did decisions change? Over one to three years, success means that a livelihoods or durable-solutions actor at Dzaleka changes a concrete allocation, outreach, or referral decision that can be attributed to the evidence, documented through decision memos or partner attestation. It means that the measurement gap itself closes, with information access appearing as a specified variable in at least one subsequent funded evaluation design that cites the standard, which is the definition of success for infrastructure. These impact claims are deliberately modest about attribution, for the gatekeeper reason set out above: the instruments can change what is known and argued, while the political constraints on durable solutions lie outside their reach.

Failure and accountability. Failure deserves equal candor, stated now: weights that fail expert consistency checks but ship anyway; adoption claims that rest on download counts rather than documented decisions; a map that decays unverified; or any drift from measurement toward identity. The project commits to reporting against both lists, to publishing a pre-registered analysis plan for the validation study, and to releasing all instruments, de-identified aggregate data, and code under open access at the end of the grant. A standing do-no-harm review screens any proposed feature that touches personal data against the commitments above before development, and the screening record is public.

11

Alignment with the Displaced Livelihoods Initiative and Other Funders

This report backs the CoLab's submission to the Displaced Livelihoods Initiative in its Infrastructure and Public Goods track: the Livelihoods Information and Access Index as an open measurement public good, validated at Dzaleka. The fit is structural. That track invites a new measure that enables future rigorous evaluations rather than the evaluation of a single program, and the initiative's own call names the lack of information on legal status as a barrier to displaced livelihoods and invites AI-supported measurement in hard-to-reach contexts. The forced-labour framing is not a detour from that agenda but its sharpest edge: strengthening information about work rights and legitimate pathways is a protective factor against the exploitation the Forced Labor Structural Risk Index measures.

The proposal is scoped to research and measurement costs only, with no intervention-delivery lines, and it treats the operational cost engine, the provenance-credential architecture, and the field partner's community access as in-kind contributions rather than billed items. The request sits within the initiative's cap for this track, and the design honors the initiative's guidance that researcher-only designs tend to underperform on policy influence, which is why the field partnership with Fraternidade Sem Fronteiras is written into the plan rather than added to it.

Team and partnership. Sylvia Maier, of the NYU SPS Center for Global Affairs, serves as Principal Investigator and eligibility anchor, since the initiative requires a Principal Investigator with a primary university affiliation and demonstrated experience with rigorous evaluation. Teresa Cantero, a doctoral candidate at Universidad Carlos III de Madrid and a visiting scholar at the NYU Center for Global Affairs, is Primary Researcher; the Summer 2026 cohort carries implementation; and Fraternidade Sem Fronteiras is the field partner.

The partnership is established rather than cold: Yago Rocha volunteered with Fraternidade Sem Fronteiras at Dzaleka on refugee livelihoods and financial inclusion, and Carolina Moron visited Dzaleka while living in Malawi and building a multi-sector data coalition, which is first-hand familiarity that lowers the risk around community access. The same architecture serves adjacent funders, because the measurement-and-governance framing travels. The honest assessment is that the initiative's funded portfolio favors teams with strong evaluation track records, so competitiveness rests on the rigor of the measurement work and the depth of the team's field experience rather than on the breadth of the claims.

12

Programme Design and Outlook

The work is designed for a term of roughly twenty-four months, beginning within the funder's required window after an award. The sequence moves from re-specifying the indicator framework for livelihoods and legal-rights information, through the expert-elicitation and calibration pipeline for the index, to the reference-context validation at Dzaleka with the field partner, and then to open release of the index, the data standard, the reference implementation, and the methodology, followed by dissemination to the research and practitioner community, including remote and partner-led presentation of results in Malawi.

Because this is an infrastructure-and-public-goods effort, the technology is the deliverable rather than overhead, and a substantial share of the budget maps to the reference implementation and the model adaptation. The cost engine itself is an in-kind contribution, already operational and validated for evacuation, which is why its re-pointing to durable solutions is an adaptation rather than a new build.

13

Consolidated Limitations and Future Research

The limitations are stated together so that they are hard to miss.

  • Maturity varies by tool, and the report holds the line between them: the cost engine is validated for evacuation, the Evacuation Information Index is built but unvalidated, the Forced Labor Structural Risk Index is built and documented with a published methodology, and the livelihoods instruments are proposed.
  • Measurement comes before deployment, so the weights and thresholds of the index and the financial model are provisional until they are expert-elicited and calibrated, and nothing here is operational decision-support before that validation.
  • Costs are more reliable than modeled outcomes, so the work leads with cost and break-even and treats any modeled mortality or lives-saved figure as indicative only.
  • The economics of inclusion are context-dependent, so the model has to be able to find against inclusion where host markets are too thin, or it is advocacy rather than analysis.
  • Evidence has political limits, so the impact claims are scoped to what better information can plausibly move.
  • Do-no-harm dominates, so the documented refusal to build biometric identity and the privacy-preserving patterns are public goods in their own right.
  • The honest scope boundary holds: full psychometric validation on the displaced population is future primary work, published here as a protocol rather than claimed as done.

The next steps follow from those limits.

  1. Run the expert-elicitation and calibration pipeline for the index against the Dzaleka reference context.
  2. Parameterize the financial model on Dzaleka cost inputs.
  3. Stand up the civil-society legal map with the field partner and local advocates.
  4. Publish the biometric-refusal decision pattern and stress-test the governance schema against automated benefits-eligibility systems.
  5. Carry the design into a second displacement context along a documented adoption path.

The through-line from the evacuation corridor to the camp is a single claim, tested in more than one domain: that the binding constraint on protection, on durable solutions, and on freedom from forced labour alike is the quality, accessibility, and equity of the information that people can act on, and that open, honestly labeled measurement infrastructure is the public good most worth building for it.

References

Sources

  1. 01Geneva Convention Relative to the Protection of Civilian Persons in Time of War (Fourth Geneva Convention), art. 49; Protocol Additional to the Geneva Conventions (Protocol I), arts. 57 to 58; Protocol II, art. 17; Rome Statute of the International Criminal Court, art. 8. See also “Evacuation,” How Does Law Protect in War? ICRC Casebook.
  2. 02Convention Relating to the Status of Refugees (1951) and its 1967 Protocol, esp. arts. 1 and 33 (definition and non-refoulement).
  3. 03UN General Assembly, Global Compact on Refugees, affirmed 17 Dec. 2018, A/73/12 (Part II), res. 73/151.
  4. 04Alexander Betts, “The Normative Terrain of the Global Refugee Regime,” Ethics and International Affairs 29, no. 4 (2015): pp. 363 to 375; and Roger Zetter, Protecting Forced Migrants: A State of the Art Report (Swiss Federal Commission on Migration, 2014).
  5. 05Ethical Tech CoLab, “Forced Labor Structural Risk Index,” Fall 2025. An interactive index mapping the structural conditions that enable forced labour across 184 countries on a 0 to 1 risk scale, with national and sub-national layers.
  6. 06ILO Forced Labour Convention, 1930 (No. 29) and its 2014 Protocol; the UN Protocol to Prevent, Suppress and Punish Trafficking in Persons (Palermo Protocol, 2000); and ILO, Walk Free, and IOM, Global Estimates of Modern Slavery: Forced Labour and Forced Marriage (2022), estimating 27.6 million people in forced labour worldwide.
  7. 07Cantero, Teresa. “Artificial Intelligence and Civilian Protection in Evacuation during Armed Conflict under International Humanitarian Law.” Presentation, NYU Center for Global Affairs, 2026.
  8. 08UN Office on Genocide Prevention and the Responsibility to Protect; Roberta Cohen, “Reconciling the Responsibility to Protect with IDP Protection,” Brookings Institution (2010).
  9. 09Volker Türk and Rebecca Dowd, “Protection Gaps,” in The Oxford Handbook of Refugee and Forced Migration Studies (Oxford UP, 2014), pp. 278 to 285.
  10. 10UNHCR, Global Trends: Forced Displacement (UNHCR, 2025 to 2026), reporting the global figure declining for the first time in a decade to roughly 117 million. The count of evacuated civilians has no equivalent global series.
  11. 11International Organization for Migration; OCHA, “What Is the Cluster Approach?”; OHCHR, core international human-rights instruments. On the protection consequences where host states are not parties to the 1951 Convention, see “The Protection of Refugees in Non-Signatory States,” International Journal of Refugee Law 33, no. 2 (2021): p. 188.
  12. 12Haine Beirens and Susan Fratzke, “Taking Stock of Refugee Resettlement,” Migration Policy Institute (2017); Susan Fratzke and María Belén Zanzuchi, “Complementary Pathways,” Migration Policy Institute (2024); UNHCR, Safe Pathways for Refugees (2018).
  13. 13Ethical Tech CoLab, “Evacuation Information Index (EII),” GitHub, 2026. Weighting precedents: ACAPS and the EU Joint Research Centre, INFORM Severity Index; ACLED, Conflict Index; Fund for Peace, Fragile States Index.
  14. 14Thomas L. Saaty, “How to Make a Decision: The Analytic Hierarchy Process,” European Journal of Operational Research 48, no. 1 (1990): pp. 9 to 26; Mohammed Al Fozaie, “A Guide to Integrating Expert Opinion and Fuzzy AHP,” Advances in Fuzzy Systems (2022); OECD and JRC, Handbook on Constructing Composite Indicators (OECD Publishing, 2008).
  15. 15Ethical Tech CoLab, “Exodus: Civilian Evacuation Risk Platform,” GitHub, 2026.
  16. 16Cost parameters compiled in the ERCF v7.2 documentation from humanitarian-standard sources: UNHCR WASH Manual; Sphere Association, The Sphere Handbook, 2018 ed.; World Food Programme, Annual Performance Report 2023 and Executive Board figures 2025. Primary sources should be cited directly in operational use.
  17. 17Ethical Tech CoLab and Yago Rocha, “ERCF: Evacuation Risk and Cost Framework, v7.2,” GitHub, 2026.
  18. 18Ethical Tech CoLab and Melanie MacKew, “Evacuation Simulation,” GitHub, 2026.
  19. 19Ethical Tech CoLab and India Clarke, “Evacuation Readiness and Uncertainty Simulator,” GitHub, 2026.
  20. 20UNHCR, Policy on Alternatives to Camps (2014); Bram J. Jansen, “Digging Aid: The Camp as an Option in East and the Horn of Africa,” Journal of Refugee Studies 29, no. 2 (2016): pp. 149 to 165; Oliver Bakewell, “Encampment and Self-Settlement,” and Loren B. Landau, “Urban Refugees and IDPs,” both in The Oxford Handbook of Refugee and Forced Migration Studies (Oxford UP, 2014), pp. 127 to 148; Lewis Turner, “Explaining the (Non-)Encampment of Syrian Refugees,” Mediterranean Politics (2015).
  21. 21UNHCR, Protracted Refugee Situations, EC/54/SC/CRP.14 (2004), reporting the average duration of major protracted refugee situations rising from about 9 years in 1993 to about 17 years by 2003. See also World Bank, “How Long Do Refugees Stay in Exile?” (2019).
  22. 22IOM, World Migration Report 2024 (IOM, 2024), Chapter 1 overview.
  23. 23“Dzaleka Refugees at Risk as WFP Faces K19bn Shortfall,” Nation Online, 8 Feb. 2026; “Funding Cuts Push Malawi's Dzaleka Refugee Camp to the Brink,” FairPlanet, 20 Aug. 2025; UNHCR, “Refugees in Dzaleka Struggle to Make Ends Meet amid Funding Cuts,” UNHCR Africa, 5 May 2025; “We Still Need People to Hold Our Hands,” Dialogue Earth, June 2026.
  24. 24World Bank, Making Refugee Self-Reliance Work: From Aid to Employment in Sub-Saharan Africa (World Bank, 2025); World Bank, “The Costs Come Before the Benefits” (2024). On the politics of self-reliance, see Suzan Ilcan et al., “Humanitarian Assistance and the Politics of Self-Reliance: Uganda's Nakivale Refugee Settlement,” CIGI Papers no. 86 (2015); Merle Kreibaum, “Build Towns Instead of Camps” (German Development Institute, 2016); Amelia Kuch, “Naturalization of Burundi Refugees in Tanzania,” Journal of Refugee Studies (2016); and World Bank, An Assessment of Uganda's Progressive Approach to Refugee Management (2016).
  25. 25UNHCR, “Investing in Refugees' Self-Reliance: A More Cost-Effective and Sustainable Response.”
  26. 26World Bank, “The Costs Come Before the Benefits” (World Bank Group, 2024); UNHCR and World Bank, “Yes” in My Backyard? The Economics of Refugees and Their Social Dynamics in Kakuma, Kenya (World Bank, 2016), for the 3.4 per cent Kakuma figure; on Colombia's PEP regularization, Ana María Ibáñez et al., as summarized in “Refugees Mean Business: This Is Why Investing in Them Pays,” World Economic Forum, 20 Jan. 2025.
  27. 27Digital Public Goods Alliance, DPG Standard (nine indicators, including SDG relevance, open licensing, non-collection of personally identifiable information, and privacy and applicable laws); Principles for Digital Development.
  28. 28E. Fiddian-Qasmiyeh, “Gender and Forced Migration,” in The Oxford Handbook of Refugee and Forced Migration Studies (Oxford UP, 2014), pp. 395 to 408; Volker Türk, “Ensuring Protection for LGBTI Persons of Concern,” Forced Migration Review.
  29. 29On existing approaches to measuring information reaching affected populations, see the IASC framework on accountability to affected populations; the CDAC Network and Ground Truth Solutions on communication with communities; and Internews information-ecosystem assessments.
  30. 30World Food Programme, “Building Blocks,” WFP Innovation; UN Women and WFP, Blockchain Pilot Project for Cash Transfers in Refugee Camps: Jordan Case Study (UN Women, 2021).
  31. 31“Privacy in Peril: Safeguarding Digital Data in Humanitarian Blockchain Initiatives,” Humanitarian Practice Network and ODI, 27 Aug. 2025.
  32. 32Ethical Tech CoLab, “Arts Provenance Agent,” GitHub, 2026; on the ID2020 lineage, see public documentation of the decentralized-identity work with UNHCR and WFP.
  33. 33Alice Edwards and Laura van Waas, “Statelessness,” in The Oxford Handbook of Refugee and Forced Migration Studies (Oxford UP, 2014), pp. 290 to 299; Sarah Bidinger, “Syrian Refugees and the Right to Work,” Boston University International Law Journal 33 (2015): pp. 223 to 249.
  34. 34Human Rights Watch, “UN Shared Rohingya Data Without Informed Consent,” 15 June 2021.
  35. 35“Although Shocking, the Rohingya Biometrics Scandal Is Not Surprising,” ODI Insights, 2021; “Rohingya Data Protection and the UN's Betrayal,” The New Humanitarian, 21 June 2021.
  36. 36“Knowledge, Voice and Power,” Forced Migration Review 70 (2022); Mauricio Vitoria et al., “The Global Summit of Refugees and the Importance of Refugee Self-Representation,” Forced Migration Review (2018).
  37. 37Innovations for Poverty Action, Displaced Livelihoods Initiative: Call for Proposals, Round V (IPA, 2026); IPA, DLI Eligibility Self-Assessment and Budget Template, Round V (2026).

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings are those of the researchers and do not represent the official positions of New York University, Microsoft, UNHCR, WFP, or any partner institution. External programs cited are referenced as evidence, not as CoLab partnerships.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file +

Publications · Academic report

After the Corridor

From AI-Informed Evacuation to Digital Public Goods for Refugee Economic Inclusion

Ethical Tech CoLab

NYU Center for Global Affairs

July 2026

Carolina de Almeida Pernambuco Moron, Christine Lumen, Alana Robertson, Melanie MacKew, Carlos D. Ruiz, India Clarke, and Yago Rocha, with Teresa Cantero (PhD Candidate, Universidad Carlos III de Madrid; Adjunct Professor, IE University; Visiting Scholar, NYU CGA). Faculty directors: Prof. Yorke E. Rhodes III and Sylvia Maier.

How can AI inform evacuation decisions, and what happens after? This report examines the CoLab's five fielded evacuation prototypes, their limitations and research frontiers, and extends the arc to the camp: the measurement, financial-modeling, and rights infrastructure that can shorten protracted displacement and improve lives. Dzaleka Refugee Camp in Malawi is the case study ground.

Key figures

6

built instruments: five fielded evacuation prototypes and a completed forced-labour risk index, all open source

53k to 60k

people now at Dzaleka, a camp built in 1994 for 10,000 to 12,000

90%

cut to UNHCR's Malawi budget, from about US$8 million to about US$1 million

0

biometric identity systems: the category the CoLab refuses to build, recorded as a versioned decision pattern

01

Executive Summary

The Ethical Tech CoLab's evacuation portfolio was built to answer a protection question: how can and should AI tools be used to enhance the protection and evacuation of civilians during armed conflict, and in what ways do they support or challenge existing obligations under International Humanitarian Law?

Across five fielded prototypes, a crisis-information index, a unified risk platform, a calibrated cost-and-risk decision framework, an agent-based behavior model, and a readiness-under-uncertainty simulator, one finding recurred: the binding constraint on protection is rarely the absence of a route or a resource, but the quality, accessibility, and equity of the information people and responders act on. IHL frames evacuation as a last-resort duty; the technology's job is to make that duty implementable under uncertainty.

What happens after the evacuation? A successful evacuation typically ends in a camp, where the modal outcome is years of protracted displacement. Using Dzaleka Refugee Camp, Malawi, where humanitarian financing has collapsed while an encampment policy bars residents from work, we show that the same measurement discipline, financial modeling, and uncertainty methods that served evacuation convert directly into durable-solutions infrastructure: an open Livelihoods Information and Access Index, a resettlement-and-inclusion financial model built on the CoLab's operational evacuation cost engine, privacy-preserving cash and credential rails, and a civil-society legal-assistance map. We also state what we will not build, biometric identity for displaced people, and why that refusal is itself a public good.

Most of the world's refugees live in protracted situations, and the modal outcome of reaching safety is years, sometimes decades, inside a camp or settlement. The legal architecture that governs what comes next is the Geneva refugee-protection consensus: the 1951 Convention relating to the Status of Refugees and its 1967 Protocol, the doctrine of durable solutions that UNHCR administers, and the 2018 Global Compact on Refugees. That consensus recognizes three ways a displacement can end well, voluntary repatriation, local integration, and resettlement, and each of them is gated far more by information and legal status than by physical capacity.

The CoLab's built work already reaches beyond the evacuation corridor. The same measurement discipline produced the Forced Labor Structural Risk Index, a completed instrument that maps the structural risk of forced labour across countries and sub-national regions, the exploitation that displacement, encampment, and the loss of the right to work make more likely. Read together, the evacuation models and the risk index share one purpose: to improve the information on which protection depends, and through it to reduce forced labour and the vulnerabilities that displaced people carry.

This report keeps three registers distinct throughout: what is built and validated, what is built but not yet validated, and what is proposed. No claim in these pages exceeds the register of the tool it describes.

The ask, in one paragraph. The Ethical Tech CoLab seeks grant funding to develop this technology infrastructure work as open, portable public goods, to improve the lives of displaced communities. It would be validated first at Dzaleka with the field partner Fraternidade Sem Fronteiras, and we are open to other partners with the same mission-driven motivations. Our intent is developing tools and releasing them to the public so that any research or practitioner team can adopt them. One possibility is the Displaced Livelihoods Initiative run by Innovations for Poverty Action and J-PAL, in its Infrastructure and Public Goods track. The report is written to align with that submission and to stand on its own for any peer funder or academic reviewer.

02

The Research Question, and Why It Runs Past the Corridor

The protection question and the accounting gap. The portfolio's organizing question comes from Teresa Cantero's doctoral research at the Instituto de Derechos Humanos Gregorio Peces-Barba, Universidad Carlos III de Madrid: how can and should artificial intelligence be used to enhance the protection and evacuation of civilians during armed conflict, and in what ways does it support or challenge the implementation of existing obligations under international humanitarian law.

Under that law, evacuation is not discretionary. When the security of the population or imperative military reasons demand it, moving civilians out of danger is a binding duty under Article 49 of the Fourth Geneva Convention, Articles 57 and 58 of Additional Protocol I, and Article 17 of Additional Protocol II. A movement into which people are coerced is forcible transfer, a war crime regardless of motive. The duty runs wider than the battlefield: the doctrine of the responsibility to protect holds that where a state cannot or will not shield its population from mass atrocity, that responsibility passes to the international community, which frames protection as an obligation rather than a matter of goodwill.

Two structural facts motivate the program. The first is a protection gap. States move quickly to adopt AI for targeting and operational advantage, yet rarely operationalize computational models to protect civilians or to make an evacuation decision legible under the incomplete, fast-changing information those decisions face. The second is an accounting gap. Displacement is tracked in detail, on the order of 117 million people forcibly displaced worldwide by 2025, yet there is no global count of evacuated civilians, because evacuation is fragmented across militaries, governments, humanitarian agencies, volunteers, and civilians acting on their own.

The legal and institutional architecture. The 1951 Convention and its 1967 Protocol define who is a refugee and fix the cornerstone duty of non-refoulement, that no one may be returned to a place where they face persecution. The 2018 Global Compact on Refugees adds a cooperative framework organized around four objectives: easing pressure on host countries, enhancing refugee self-reliance, expanding access to third-country solutions, and supporting conditions for return in safety and dignity. International humanitarian law governs conduct in armed conflict, and the responsibility to protect frames the state's duty toward its own population.

The institutions matter as much as the instruments. UNHCR leads protection and durable solutions for refugees; the International Organization for Migration is the United Nations migration agency; the Office for the Coordination of Humanitarian Affairs coordinates emergency response through the cluster approach; and the Office of the High Commissioner for Human Rights anchors the broader human-rights instruments that apply regardless of a person's status. One structural limitation shapes everything downstream: many host states are not parties to the 1951 Convention or maintain significant reservations, so the protection a displaced person actually receives depends heavily on where they land.

The three durable solutions. The arc does not end when the shooting stops. An evacuation succeeds when a person reaches relative safety, and safety, in practice, usually means a camp. There the analytical problem inverts. The question is no longer whether to move and how, but how to leave well now that one has arrived.

The three durable solutions recognized by the refugee-protection regime, voluntary repatriation to the country of origin, local integration in the country of asylum, and resettlement to a third country, are the recognized routes out. Each is gated by information and legal status as much as by physical or economic capacity, a point the regime's own scholarship makes plainly. Resettlement reaches only a small fraction of those who need it, and complementary pathways such as labour or study routes remain modest in scale, which is why local integration and the conditions for safe return carry most of the weight in practice. This report treats the durable-solutions phase as the natural continuation of the evacuation research question, and asks what open measurement and modeling infrastructure would help displaced people and the organizations serving them find the fastest defensible route out.

03

The Built Portfolio

This section describes what the CoLab has already built, and is candid about how mature each piece is, because a reader deciding whether to fund the work is owed that precision. Three registers are kept distinct. The Evacuation Risk and Cost Framework is built and validated in its evacuation form. The Evacuation Information Index, the Exodus platform, and the two simulators are built and operational but not yet expert-validated or calibrated. The Forced Labor Structural Risk Index is built and completed, documented in a methodology page and an associated publication. The livelihoods instruments the report proposes, the Livelihoods Information and Access Index and the Resettlement and Inclusion Capacity Simulator, together with the value-transfer and legal-map builds, are proposed, and their validation is future work. All are open source.

PrototypeFunctionCore method
Evacuation Information Index (EII)Scores quality, accessibility, and equity of crisis information by geographyComposite index on INFORM, ACLED, and FSI logic; live news and conflict feeds
Exodus, Civilian Evacuation Risk PlatformUnifies EII, the cost model, and endangerment and feasibility assessmentFastAPI backend; shared calculators and data layer
ERCF, Evacuation Risk and Cost FrameworkEstimates human and financial cost of evacuating against stayingSeven-dimension scoring; calibrated cost model; break-even analysis
Evacuation Behavior SimulatorModels how information and demographics shape who leaves, and whenAgent-based model; probabilistic status lifecycle
Readiness and Uncertainty SimulatorModels how intelligence uncertainty degrades destination choicesMonte Carlo; gatekeeper factors; uncertainty slider
The five fielded evacuation prototypes. All are open source at github.com/Ethical-Tech-CoLab.

The Evacuation Information Index. The index treats crisis information itself as the object of measurement. It scores three families of factors, impact, conditions, and complexity, on a scale of 1 to 5, using weights synthesized from the INFORM Severity Index, the Armed Conflict Location and Event Data project, and the Fragile States Index. A live backend adds current news and a conflict timeline for each crisis, and a map layer can overlay AI-derived building-damage and route-damage tiles from Microsoft HASTE and Planet imagery. The design is deliberately transparent: weights are labeled provisional pending expert validation, and where the hosted site cannot run the live backend, responses are baked into a clearly labeled static snapshot so that nothing is presented as live when it is not.

The index has real limits, and the report states them rather than hiding them.

  • Weights are borrowed by analogy, not yet expert-validated or sensitivity-tested; a reviewer will read pre-validation weights as pre-evidence.
  • Source and language bias: feeds skew to English-language, internationally indexed reporting, so the populations with the worst information access are least visible to the instrument that scores information access. The index must correct for this, not inherit it.
  • The damage overlay is currently self-hosted with no public interface, so for most users the combined information-and-damage view is empty until it is connected.

Future research on the index. Run the three-phase validation the CoLab's methodology work already specifies, budget-allocation and Delphi elicitation with 8 to 12 domain experts, Fuzzy-AHP on contested weights, then historical case calibration, and add an equity layer measuring who receives information, disaggregated by gender, age, nationality, and residence inside or outside a camp. That equity layer is the move that turns the index into the livelihoods instrument described later in this report.

Exodus, the integration layer. Exodus unifies the crisis map, the cost-and-scenario model, and an endangerment-and-feasibility assessment behind a single backend, design language, and interface, including calculators for the cost of staying, the cost of evacuating, resource needs, and scenario management. Integration is not validation. Unifying three research prototypes behind a more authoritative interface inherits their uncalibrated assumptions, which raises rather than lowers the duty to label confidence honestly. The same interface is the seam where the durable-solutions module plugs in. Its stated limitation is that it is not automatically updated; its future research is to validate the approaches, weights, and measurements.

The Evacuation Risk and Cost Framework. Now at version 7.2, the framework is the CoLab's calibrated decision-support tool for the human and financial cost of civilian evacuation, and it is the one component of the portfolio that is validated rather than provisional, in its evacuation form. It scores scenarios across seven dimensions, kinetic threat, mobility, authorization, logistics, destination, urgency, and information, estimates evacuation resources and in-situ assistance costs, and produces a break-even analysis: the day after which a one-time evacuation becomes cheaper than the mounting daily cost of sustaining a population in place. Its cost parameters are sourced to humanitarian standards and individually tagged as Validated, Estimated, or Unvalidated, drawing on UNHCR water and sanitation guidance, the Sphere standards, and World Food Programme figures.

The mortality model is disclosed as indicative rather than predictive, with a reported log-log fit of 0.855, a leave-one-out cross-validation of 0.807, and a documented case where the structure fails. The financial estimates are substantially more reliable than the mortality figures, which is why the durable-solutions adaptation leads with cost. The framework's ethical architecture is explicit: it does not place a monetary value on a human life. Its figures are logistics estimates, its labels are descriptive rather than prescriptive, and it presents data to support a decision rather than issuing one. That framing carries verbatim into every extension in this report.

The framework's limitations are equally explicit.

  • It models the immediate field operation only, on the order of US$134 per person at moderate risk over 50 kilometers by ground. Real operations, Kosovo 1999 at about US$562 per person and Lebanon 2006 at about US$4,500 to US$5,700, add international transport, reception, and hosting. Extending the cost boundary to the full displacement journey, including protracted encampment, is the frontier the durable-solutions model takes up.
  • Ground-truthing against OCHA's Financial Tracking Service is on the backlog and belongs in the funded work plan.

The Evacuation Behavior Simulator. An agent-based model of how information spread and demographics shape who leaves and when. Agents move through a lifecycle from unaware to seeking, milling, evacuating, and done, with family clusters that wait for all members, confirmation loops whose speed depends on how clear the information is, neighbor influence, and calibrated delays for elders, young children, pregnant individuals, and unaccompanied minors. It is pedagogical rather than predictive, and its household-composition logic transfers directly to modeling who is left behind by information failures inside a camp. Its behavior may vary by society, culture, location, personality, income, and education, and the future research is to build intersectionality between behavior simulations.

The Readiness and Uncertainty Simulator. A seeded, fully reproducible instrument that models how uncertainty in field intelligence degrades an evacuation decision. Destinations are scored across seven factors, 500 Monte Carlo trials per group-and-destination pair convert readiness into a distribution of success probabilities, and an uncertainty control degrades every factor's confidence at once, so a user can watch predicted outcomes collapse while nothing on the ground has changed.

Two design choices carry into the rest of this report. First, three factors are non-substitutable gatekeepers, security, the consent of the local authority, and the willingness of the host community, any of which caps readiness: a destination that is dangerous, legally closed, or unwelcoming cannot be redeemed by good logistics. Second, the model insists on the difference between a factor that is unknown, a gap that better information can close, and a factor that is a confirmed refusal, a settled political exclusion. Treating the two as the same would make an unresolvable refusal look like a solvable data problem. Structurally, assigning groups to destinations under consent, willingness, security, and capacity constraints, in the presence of uncertainty, is resettlement matching, which is the transfer the durable-solutions model makes. The simulator has not yet been stress tested with different AI models.

The Forced Labor Structural Risk Index. The index applies the same composite-index method to a different harm. It is a completed, interactive index that maps the structural conditions enabling forced labour across 184 countries on a scale from 0 to 1, with national and sub-national layers, so that the risk can be seen and compared rather than assumed. It belongs in this report because displacement is a forced-labour risk multiplier: the loss of legal status and the right to work, the collapse of assistance, and weakened oversight are precisely the conditions under which exploitation grows, and forced labour is a defined harm in international law with an estimated 27.6 million people subject to it worldwide.

Its significance for the durable-solutions argument is conceptual: measuring structural forced-labour risk and measuring livelihoods-information access are two halves of one protective picture, one showing where exploitation is likely, the other showing whether the information that could prevent it is reaching people. As a composite index its weights remain open to expert re-elicitation, it is not automatically updated, and sponsorship mechanisms remain unsourced at country scale while an eight-country hand-coded kafala signal feeds two domains in the shipped build. Writing an ILO-endorsed kafala passage on top of that unresolved discrepancy would have overstated the index's grounding.

Cross-cutting synthesis. Read together, the built prototypes yield a short set of lessons. Information access is the binding constraint, and it is the thread that carries from the corridor into the camp. Uncertainty has to be modeled explicitly, with the unknown kept separate from the confirmed, or the analysis misleads by omission. We address gaps in data by generating synthetic data that mimics real-life data, and address the residual uncertainty with Monte Carlo simulation. A composite index built from the sequence of borrowing precedents, eliciting expert weights, and calibrating against cases is portable infrastructure: the same method that scores crisis information also scores structural forced-labour risk. And cost modeling in a context of human lives has to stay descriptive rather than prescriptive, following the framework's refusal to monetize life.

04

After Arrival: Dzaleka and the Second Emergency

A successful evacuation deposits people into a slower emergency. The concept of a protracted refugee situation, conventionally a population of at least 25,000 of the same nationality in exile for five years or more, describes where most refugees actually live, and care-and-maintenance encampment becomes the default by inertia rather than by anyone's plan.

The duration is not a rounding error in a person's life. By UNHCR's own accounting, the average duration of a protracted refugee situation rose from about 9 years in 1993 to roughly 17 years by 2003, and while later analysis rightly cautions that a single average conceals wide variation, the order of magnitude is not in dispute: protracted exile routinely lasts a decade or more, and often a generation. The literature on encampment and self-settlement has long argued that the camp is often less a humanitarian necessity than an administrative and political choice, and that its costs to refugees, in mobility, work, and self-reliance, are systematically understated.

Dzaleka and the funding collapse. Dzaleka Refugee Camp in Dowa District, Malawi, is the sharp end of that pattern. Built in 1994 for 10,000 to 12,000 people, it now holds on the order of 53,000 to 60,000 people, predominantly from the Democratic Republic of the Congo, Burundi, and Rwanda, and it is caught in the funding collapse of 2025 and 2026.

The numbers describe a system in retreat.

  • UNHCR's Malawi budget fell from about US$8 million to about US$1 million, a cut of roughly 90 percent, ending general non-food distributions and laying off protection staff, including paralegals.
  • World Food Programme assistance is secured only through the middle of 2026, with about US$11 million still needed to restore full rations through December.
  • Monthly cash support has fallen toward the range of US$6 to US$9 per person, down from roughly US$100, forcing skipped meals and other negative coping.

Malawi's encampment policy, enforced since March 2023, bars most residents from working outside the camp, so the population cannot substitute self-reliance for collapsing aid. Local civil society, including Inua Advocacy, argues the crisis is inseparable from that policy and is pressing for the right to work; weakened oversight is meanwhile reported to embolden trafficking networks.

This is not a context for piloting speculative technology on vulnerable people. It is a context where the highest-value contribution is measurement, evidence, and legibility: tools that help residents and the organizations serving them see and argue for the fastest route out, and that document the true cost of the status quo. This is a global humanitarian emergency that needs collective collaborative effort, and building infrastructure, digital goods and services, information, training, and capacity might transform generational poverty and increase access to resources.

Encampment as a driver of forced labour. These are the textbook conditions for forced labour. When legal work is barred and assistance collapses, the informal and coerced economy fills the gap. Reducing that exposure is a direct objective of the CoLab's work: the Forced Labor Structural Risk Index names where the risk concentrates, and the livelihoods-information instrument measures whether the information that could protect people, about their rights, their options, and legitimate work, is actually reaching them. That link has no official validation yet, and the future research is to separate recruitment, exploitation, and monetization by intersectionality using social demography, and to deepen the demand side.

05

The Financial Model: From Evacuation Break-Even to Inclusion Break-Even

The financial-modeling engine already exists, validated for one problem and now re-pointed to another, and the distinction matters. The Evacuation Risk and Cost Framework is validated for evacuation costing; its encampment-versus-inclusion parameterization is new and unvalidated, and most of its durable-solutions cost parameters are tagged Estimated or Unvalidated. What transfers cleanly is the architecture and the break-even mathematics, a one-time investment set against a mounting recurring cost, not a set of validated cost figures for the new use.

The engine, re-pointed.

  • Trajectory A prices status-quo encampment as a recurring per-person, per-year assistance cost, covering rations, water and sanitation, health, protection, and administration, using the same Sphere, UNHCR, and WFP parameter families the framework already carries, with the same Validated, Estimated, or Unvalidated tags.
  • Trajectory B prices an inclusion or resettlement pathway as a front-loaded investment, covering documentation, recognition of qualifications, transport, start-up capital, and host co-investment.
  • The output is the break-even year after which the status quo costs more, which is the framework's break-even chart with its axes relabeled.

The Resettlement and Inclusion Capacity Simulator is that engine re-pointed, and re-earning its validity in the durable-solutions domain is the work, not a formality. The capacity-and-willingness layer transfers from the readiness simulator: it maps destinations, whether resettlement states or local-integration pathways, against the non-substitutable gatekeepers of legal consent, host willingness, and security, alongside the substitutable capacities of housing, jobs, and services, while preserving the distinction between an option that is unknown and one that is refused, so that an unlobbied state is never mistaken for a closed door. The uncertainty layer carries over as well, so that thin data on host capacity or a state's posture visibly widens the confidence band and prevents false precision in an advocacy number.

What the evidence predicts. The World Bank's recent work on refugee self-reliance in Sub-Saharan Africa indicates that if refugees had the right to move and work, they would contribute to the hosting country's economy, with real fiscal and economic gains. The honest caveat is built into that same literature. The World Bank finds that self-reliance remains elusive across much of the region, because camps often sit in marginal, high-poverty areas where even full local integration would lift incomes only a little, and the scholarship on self-reliance programming warns that the concept is often invoked to justify the withdrawal of assistance rather than to expand rights.

The simulator has to be able to return the answer that inclusion does not pay here, where host markets are too thin, or it is advocacy rather than analysis. Its output is a defensible, reproducible sentence of the form: sustaining this population at Dzaleka costs approximately this much per year; this pathway costs approximately that much up front and breaks even in a given year, subject to these gatekeepers and this uncertainty band. The simulator could also provide country comparisons across different legal models and measure the economic impact of different refugee policies, and listing the policies of each hosting country would better inform humanitarian organizations and populations.

What the literature says the model will find. The engine formalizes an argument the development-economics evidence already supports.

  • UNHCR estimates that if freedom of movement and the right to work were relaxed so refugees could earn on par with hosts, complementary assistance costs would fall from about US$3.2 billion to roughly US$900 million a year.
  • A joint World Bank and UNHCR benchmark puts the annual cost of bringing every refugee in low and middle income host countries up to the global poverty line at almost US$62 billion in a scenario where refugees earn nothing; refugees' own earnings already cover roughly US$40 billion of that, leaving about US$22 billion.
  • World Bank analysis estimates that in Uganda's inclusive regime, economic participation saves about US$150 per refugee per year, roughly US$225 million across 1.5 million refugees.
  • The same World Bank work finds that around Kakuma, refugee presence raised the host county's gross regional product by about 3.4 percent.
  • It also reports that Colombia's regularization of Venezuelan migrants raised beneficiaries' incomes by about 30 percent and their formal employment by around 10 percentage points, with little displacement of host workers.

06

Digital Public Goods for Economic Inclusion

The toolkit is ordered from the safest build to the one that demands the most caution. The organizing principle is data minimization: build the thinnest system that solves the problem, and refuse the parts whose risks outweigh their benefit for this population. A digital public good, in the sense set out by the Digital Public Goods Alliance and endorsed within the United Nations system, is open-source software, open data, an open standard, or an open model that is relevant to the Sustainable Development Goals, does no harm by design, and, importantly for this population, minimizes the collection of personally identifiable information. The Principles for Digital Development point the same way, toward privacy, reuse, and designing with the user. Measured against those criteria, the CoLab's contributions are intended to qualify as digital public goods rather than to resemble them loosely.

Measurement: the Livelihoods Information and Access Index. The measurement instrument is the Evacuation Information Index re-specified for displaced livelihoods and legal rights: an open, versioned instrument measuring the quality, accessibility, and equity of information about work rights, employment pathways, financial services, and entrepreneurship support reaching displaced populations. The architecture transfers from the evacuation index, but the construct does not, and saying so is part of the method. Measuring information access to work rights is a different thing from scoring crisis severity, so the index has to establish its own construct validity from the ground up rather than inherit it.

In draft form the index is a matrix. Its three dimensions, quality, meaning whether information is accurate, current, and actionable; accessibility, meaning whether people can reach it in a language and channel they use; and equity, meaning whether it reaches sub-populations evenly, are scored across four domains: work rights and legal status, employment pathways, financial services, and entrepreneurship support. Each cell draws on a mix of sources: an audit of what official and organizational information exists, a structured record from field partners of what actually circulates, and, in the future primary study, what residents themselves report receiving.

The equity dimension takes seriously the finding in the forced-migration literature that displacement is gendered, that women, men, and sexual and gender minorities face different risks and different information barriers, so disaggregation is a design requirement rather than an afterthought. Because information inside a camp travels largely by word of mouth, community leaders, and channels that are not in English, the instrument has to weigh informal and in-language sources deliberately, or it will reproduce the very bias it is built to detect. The framework is provisional and is itself a deliverable, with the weights subject to the expert-elicitation pipeline before any score is treated as evidence.

The index does not enter an empty field, and the report says so plainly rather than claiming to be first. The humanitarian sector already assesses information reaching affected populations through accountability-to-affected-populations frameworks, communication-with-communities practice, and information-ecosystem assessments. What those approaches largely lack, and what the index adds, is a composite, versioned, openly standardized measure that renders information access as a comparable variable, disaggregated by sub-population, which a later impact evaluation can treat as an outcome or a mechanism. The contribution is the measure and its portability, not the discovery that information matters. Its deliverables are the index, an open data standard, a reference implementation, and methodology documentation, validated first at Dzaleka with Fraternidade Sem Fronteiras.

The open questions the index carries.

  • One size does not fit all: the geopolitical impact and the limits of building an index and adapting it live, when everything can change.
  • How can local communities own their data governance?
  • What would be the most beneficial digital goods a displaced population needs?
  • How can the information be made accessible in low-connectivity environments?
  • How can marginalized communities and people who do not read be included in the design of digital goods?

Future research on the index. Developing a work card and a system that would allow refugees to record their expertise and be matched with regions or organizations that need those services.

Value transfer: privacy-preserving cash, decoupled from biometrics. The nearest-term practical contribution is moving value more cheaply and reliably as traditional financing collapses. The reference case is the World Food Programme's Building Blocks, the largest blockchain-based humanitarian platform, which settled entitlements on a private, permissioned ledger and paid vendors in consolidated sums, saving roughly 98 percent of transaction costs on an early deployment, improving reconciliation across agencies, and establishing a co-governed network that UN Women later extended to cash-for-skills in refugee camps.

The same case is the cautionary tale. Building Blocks ran on biometric registration, and independent analysis warns that many such pilots functioned as donor showpieces under-invested in privacy, and that even a permissioned ledger can leak sensitive metadata when on-chain identifiers are correlated with off-chain records. The lesson is to decouple the payment rail from biometric identity.

A wallet designed for Dzaleka would therefore:

  • Authenticate with the thinnest possible credential, a personal identification number together with a locally issued token or code, never a biometric.
  • Keep identifying data off the ledger, with only pseudonymous entitlement state recorded on it.
  • Integrate an existing audited, co-governed network rather than build a bespoke ledger, so that the CoLab's contribution is the threat model and the measurement wrapper.
  • Run a data-protection impact assessment with community consultation before any collection.

Portability: skills credentials, never identity. Portability of skills is the third contribution, and it must stop short of identity. The CoLab operates an adjacent architecture that demonstrates the cryptographic pattern, an agent that issues cryptographically signed, tamper-evident verifiable credentials for cultural objects, and Professor Rhodes's prior work with the ID2020 effort concerned decentralized, portable humanitarian credentials with UNHCR and the World Food Programme. The pattern is shared; the risk surface of issuing credentials to displaced people is not, and the report treats that gap as real rather than solved.

The responsible application is narrow: portable, self-held credentials for skills, qualifications, and completed training, so that a nurse or an electrician does not lose a decade of accreditation to displacement, issued to and controlled by the individual and presented selectively. The bright line is firm. A skills credential records what a person can do. It must never become a de facto identity card that gatekeeps aid or is legible to a persecuting authority. The moment it records who someone is, it inherits the risks discussed in the next section and should be stopped.

Rights: mapping civil-society legal assistance. Donor cuts stripped UNHCR's own paralegals out of Dzaleka, which makes a maintained, open directory of who actually provides legal help the safest and highest-leverage build. Organized by legal need, refugee-status determination, work authorization, documentation, protection from gender-based violence, protection from trafficking, and resettlement referral, the map records which civil-society and refugee-led organizations provide assistance, with contact details, language, and eligibility, under the same allowlist-and-citation discipline the provenance agent enforces, so that no unsourced claim ships.

It connects to the index directly. The index shows where rights information fails; the map routes a person to the organization that can close the gap. Local actors such as Inua Advocacy are the anchor nodes, rather than the government. This legal-and-jurisdiction layer, covering work rights, documentation rules, encampment policy, recognition of qualifications, and access to financial services, is a first-class component of the work, coded from public legal sources, and it reflects the long-standing finding in the refugee literature that legal status and the right to work are among the decisive determinants of whether displaced people can build a livelihood at all.

07

Vulnerable Populations, Digital Identity Risk, and Mitigation

The Rohingya precedent. Refugees are an exceptionally vulnerable group whose personal data demands unusual rigor, and the technology that most tempts humanitarian builders, a comprehensive digital identity for displaced people, is the one this team refuses to build. The evidence is documented rather than hypothetical. Human Rights Watch found that biometric and biographic data collected from Rohingya refugees in Bangladesh was shared onward, ultimately reaching Myanmar, the government they had fled, for the purpose of assessing repatriation eligibility, covering some 830,000 people between 2018 and 2021, without the full data-protection impact assessment UNHCR's own policy requires and without meaningful informed consent.

Consent in that system was structurally compromised. Refugees were told that aid depended on their registration, which makes agreement coerced rather than free. Consent materials were commonly presented in English, a language many of those required to register did not read, and reporting indicates that families who refused enrollment were cut off from food and services. The defining property of the harm is its irreversibility. Biometrics are sticky; an iris cannot be reissued once it is compromised, and a database built to facilitate solutions suffered precisely the function creep that critics had predicted, as data gathered for one stated purpose was repurposed for another.

Mitigation is mostly refusal, then minimization, and it can be stated as a small number of firm commitments.

  • Do not build biometric identity. A category refusal, encoded as a versioned decision pattern in the CoLab's open schema, DDC-0001, biometric registration of displaced populations.
  • Decouple entitlement from identity: access gated by the thinnest credential, never by who you are proven biometrically.
  • Data minimization and purpose limitation: least data, single stated purpose, no secondary use or third-party transfer, as a matter of design rather than policy alone.
  • Individual control and local governance: the person holds and selectively presents any credential; refugee-led and local actors govern the data.
  • Real consent or none: if consent cannot be free and informed, whether because aid is conditioned on it or because a language or power gap cannot be bridged, treat it as absent and do not proceed.
  • Mandatory pre-collection data-protection impact assessment and community consultation: the omitted step whose absence produced the Rohingya harms.

For a funder, this refusal is an asset. A responsible-AI lab that can name the one thing it will not build, and cite exactly why, is more credible than one that promises everything.

08

Aggregating Evidence So Displaced People Can Argue Their Case Sooner

Taken together, these instruments become an evidence engine that attacks the core information asymmetry facing displaced people, the fact that the rights-and-options case that could change an individual's situation today requires specialized legal and economic knowledge that the individual does not have. The index shows, with disaggregation, which sub-populations are systematically failed by rights-and-pathways information, which is evidence rather than anecdote. The financial model reframes the ask from a plea for help into a costed comparison, that prolonged encampment costs a given amount per year while a given pathway costs less by a given year, which is the argument that donors and states respond to. The legal map converts the evidence into action, routing people to organizations that can file, refer, and litigate.

One limit has to be stated plainly, because the evidence engine cannot dissolve it. Durable solutions are rationed by state consent, resettlement quotas, and political will, and the readiness model's own logic holds that a confirmed refusal is a settled exclusion that no measurement moves. Better information and a sharper cost argument can widen the set of options that are seriously considered and can equip advocates who would otherwise argue from anecdote, but where the binding constraint is political rather than informational, rigorous evidence may change nothing. The impact claims in this report are scoped to that reality: the instruments improve the quality and legibility of the case, which is necessary but not sufficient for a different outcome.

The outputs are designed to be wielded by refugee-led organizations and civil-society advocates, and, where they will use them, by durable-solutions units within UNHCR, rather than to sit in a dashboard. This design choice reflects a clear direction in the field, the growing insistence, visible across the refugee-studies and practitioner literature on knowledge, voice, and power and on refugee self-representation, that displaced people should hold and deploy the evidence about their own situations rather than be the objects of it.

Two guardrails hold throughout. Every advocacy artifact carries its uncertainty band honestly, and the aggregation layer never becomes a registry of identifiable individuals. It argues about populations and gaps, not named people.

09

What Infrastructure Is Actually Needed

Two things are absent by design: any biometric or identity system, and any bespoke ledger. Everything proposed is either a measurement or knowledge artifact that the CoLab is suited to own, or a thin integration on audited external rails. That is the correct risk posture for a small lab working with a highly vulnerable population, and it follows directly from the digital-public-goods and data-minimization commitments set out above.

LayerComponentReusesPosture
MeasurementLIAI index, open data standard, reference implementation, methodologyEII and the composite-index methodBuild, the funding ask
EconomicsRICS: encampment-versus-inclusion break-even; capacity and willingness mapERCF engine (operational) and readiness gatekeepersAdapt, engine in place
Value transferPrivacy-preserving cash wallet, biometric-freeExternal audited network; CoLab adds threat model and measurementIntegrate, do not rebuild
PortabilitySelf-held skills and qualification credentialsProvenance verifiable-credential architecture; ID2020 lineageBuild narrowly, skills only
RightsOpen civil-society legal-assistance mapProvenance sourcing disciplineBuild, safest and high leverage
GovernanceDDC decision-pattern schema; DPIA and ethics layerDDC-0001Build, the credibility spine
The proposed stack, and what each layer reuses from the built portfolio.

10

How Success Will Be Measured

A funder should not have to take this program on faith, so success is defined in advance in three tiers: outputs, whether the instruments shipped and are rigorous; outcomes, whether they work and are used; and impact, whether they changed decisions. The measurement instruments themselves double as the evidence base. One line governs the whole framework and resolves an ambiguity a careful reader would otherwise catch: under this grant, every activity involving residents is consultation and co-design of the instruments, together with a desk audit and the field partner's structured knowledge, and no personal data is collected from residents. The measurement of residents' own information access, which would require primary data collection, is a separate future study.

Outputs: did we ship, and is it rigorous? Over the first year, success means a first open-source release of the index, comprising the index itself, an open data standard, a reference implementation, and methodology documentation, each versioned and citable. It means completed and published validation of the instrument's design, with expert elicitation across 8 to 12 experts over two rounds, reported weights, consistency ratios within accepted thresholds, and divergence analysis, including negative results. It means the reference-context validation executed with Fraternidade Sem Fronteiras, drawing on an audit of existing information and the partner's structured knowledge of what circulates, using open and synthetic data. It means the financial model parameterized for Dzaleka on the operational cost engine, with every parameter tagged and the full parameter table public. And it means the civil-society legal map live with the large majority of entries verified against primary sources, and the biometric-refusal decision pattern published with its governance schema.

Full psychometric validation of the instrument on responses from the displaced population itself, covering test-retest reliability, construct validity, and measurement invariance across subgroups, requires primary data collection from that population and is not part of this phase. It is written up here as a published protocol and pre-analysis plan for a future primary study, so that the instrument and its validation plan are themselves a public good, and so that a later study can execute the plan without re-deriving it. The longitudinal read on whether information-access scores for the worst-served sub-populations improve over time belongs to that same future study. This is the honest boundary of the present work, and stating it plainly is part of the method.

Outcomes: does it work, and is it used? Over the first two years, success means concordance between the instrument and the field partner's operational knowledge of where information gaps sit at Dzaleka, and equity disaggregation that reproduces independently observed disparities. It means adoption, with at least one external research team taking up the index as an instrument and at least one practitioner organization using its scores to retarget outreach. It means the financial model cited in at least two advocacy or policy documents by refugee-led or civil-society organizations, with the uncertainty bands reproduced intact, since fidelity of use matters and not only volume. It means community legitimacy, with refugee community members engaged as co-designers and a community review session signing off on how findings are represented before publication. And it means open-source health, with public contribution metrics, quarterly releases, and a documented path to a second site.

Impact: did decisions change? Over one to three years, success means that a livelihoods or durable-solutions actor at Dzaleka changes a concrete allocation, outreach, or referral decision that can be attributed to the evidence, documented through decision memos or partner attestation. It means that the measurement gap itself closes, with information access appearing as a specified variable in at least one subsequent funded evaluation design that cites the standard, which is the definition of success for infrastructure. These impact claims are deliberately modest about attribution, for the gatekeeper reason set out above: the instruments can change what is known and argued, while the political constraints on durable solutions lie outside their reach.

Failure and accountability. Failure deserves equal candor, stated now: weights that fail expert consistency checks but ship anyway; adoption claims that rest on download counts rather than documented decisions; a map that decays unverified; or any drift from measurement toward identity. The project commits to reporting against both lists, to publishing a pre-registered analysis plan for the validation study, and to releasing all instruments, de-identified aggregate data, and code under open access at the end of the grant. A standing do-no-harm review screens any proposed feature that touches personal data against the commitments above before development, and the screening record is public.

11

Alignment with the Displaced Livelihoods Initiative and Other Funders

This report backs the CoLab's submission to the Displaced Livelihoods Initiative in its Infrastructure and Public Goods track: the Livelihoods Information and Access Index as an open measurement public good, validated at Dzaleka. The fit is structural. That track invites a new measure that enables future rigorous evaluations rather than the evaluation of a single program, and the initiative's own call names the lack of information on legal status as a barrier to displaced livelihoods and invites AI-supported measurement in hard-to-reach contexts. The forced-labour framing is not a detour from that agenda but its sharpest edge: strengthening information about work rights and legitimate pathways is a protective factor against the exploitation the Forced Labor Structural Risk Index measures.

The proposal is scoped to research and measurement costs only, with no intervention-delivery lines, and it treats the operational cost engine, the provenance-credential architecture, and the field partner's community access as in-kind contributions rather than billed items. The request sits within the initiative's cap for this track, and the design honors the initiative's guidance that researcher-only designs tend to underperform on policy influence, which is why the field partnership with Fraternidade Sem Fronteiras is written into the plan rather than added to it.

Team and partnership. Sylvia Maier, of the NYU SPS Center for Global Affairs, serves as Principal Investigator and eligibility anchor, since the initiative requires a Principal Investigator with a primary university affiliation and demonstrated experience with rigorous evaluation. Teresa Cantero, a doctoral candidate at Universidad Carlos III de Madrid and a visiting scholar at the NYU Center for Global Affairs, is Primary Researcher; the Summer 2026 cohort carries implementation; and Fraternidade Sem Fronteiras is the field partner.

The partnership is established rather than cold: Yago Rocha volunteered with Fraternidade Sem Fronteiras at Dzaleka on refugee livelihoods and financial inclusion, and Carolina Moron visited Dzaleka while living in Malawi and building a multi-sector data coalition, which is first-hand familiarity that lowers the risk around community access. The same architecture serves adjacent funders, because the measurement-and-governance framing travels. The honest assessment is that the initiative's funded portfolio favors teams with strong evaluation track records, so competitiveness rests on the rigor of the measurement work and the depth of the team's field experience rather than on the breadth of the claims.

12

Programme Design and Outlook

The work is designed for a term of roughly twenty-four months, beginning within the funder's required window after an award. The sequence moves from re-specifying the indicator framework for livelihoods and legal-rights information, through the expert-elicitation and calibration pipeline for the index, to the reference-context validation at Dzaleka with the field partner, and then to open release of the index, the data standard, the reference implementation, and the methodology, followed by dissemination to the research and practitioner community, including remote and partner-led presentation of results in Malawi.

Because this is an infrastructure-and-public-goods effort, the technology is the deliverable rather than overhead, and a substantial share of the budget maps to the reference implementation and the model adaptation. The cost engine itself is an in-kind contribution, already operational and validated for evacuation, which is why its re-pointing to durable solutions is an adaptation rather than a new build.

13

Consolidated Limitations and Future Research

The limitations are stated together so that they are hard to miss.

  • Maturity varies by tool, and the report holds the line between them: the cost engine is validated for evacuation, the Evacuation Information Index is built but unvalidated, the Forced Labor Structural Risk Index is built and documented with a published methodology, and the livelihoods instruments are proposed.
  • Measurement comes before deployment, so the weights and thresholds of the index and the financial model are provisional until they are expert-elicited and calibrated, and nothing here is operational decision-support before that validation.
  • Costs are more reliable than modeled outcomes, so the work leads with cost and break-even and treats any modeled mortality or lives-saved figure as indicative only.
  • The economics of inclusion are context-dependent, so the model has to be able to find against inclusion where host markets are too thin, or it is advocacy rather than analysis.
  • Evidence has political limits, so the impact claims are scoped to what better information can plausibly move.
  • Do-no-harm dominates, so the documented refusal to build biometric identity and the privacy-preserving patterns are public goods in their own right.
  • The honest scope boundary holds: full psychometric validation on the displaced population is future primary work, published here as a protocol rather than claimed as done.

The next steps follow from those limits.

  1. Run the expert-elicitation and calibration pipeline for the index against the Dzaleka reference context.
  2. Parameterize the financial model on Dzaleka cost inputs.
  3. Stand up the civil-society legal map with the field partner and local advocates.
  4. Publish the biometric-refusal decision pattern and stress-test the governance schema against automated benefits-eligibility systems.
  5. Carry the design into a second displacement context along a documented adoption path.

The through-line from the evacuation corridor to the camp is a single claim, tested in more than one domain: that the binding constraint on protection, on durable solutions, and on freedom from forced labour alike is the quality, accessibility, and equity of the information that people can act on, and that open, honestly labeled measurement infrastructure is the public good most worth building for it.

References

Sources

  1. 01Geneva Convention Relative to the Protection of Civilian Persons in Time of War (Fourth Geneva Convention), art. 49; Protocol Additional to the Geneva Conventions (Protocol I), arts. 57 to 58; Protocol II, art. 17; Rome Statute of the International Criminal Court, art. 8. See also “Evacuation,” How Does Law Protect in War? ICRC Casebook.
  2. 02Convention Relating to the Status of Refugees (1951) and its 1967 Protocol, esp. arts. 1 and 33 (definition and non-refoulement).
  3. 03UN General Assembly, Global Compact on Refugees, affirmed 17 Dec. 2018, A/73/12 (Part II), res. 73/151.
  4. 04Alexander Betts, “The Normative Terrain of the Global Refugee Regime,” Ethics and International Affairs 29, no. 4 (2015): pp. 363 to 375; and Roger Zetter, Protecting Forced Migrants: A State of the Art Report (Swiss Federal Commission on Migration, 2014).
  5. 05Ethical Tech CoLab, “Forced Labor Structural Risk Index,” Fall 2025. An interactive index mapping the structural conditions that enable forced labour across 184 countries on a 0 to 1 risk scale, with national and sub-national layers.
  6. 06ILO Forced Labour Convention, 1930 (No. 29) and its 2014 Protocol; the UN Protocol to Prevent, Suppress and Punish Trafficking in Persons (Palermo Protocol, 2000); and ILO, Walk Free, and IOM, Global Estimates of Modern Slavery: Forced Labour and Forced Marriage (2022), estimating 27.6 million people in forced labour worldwide.
  7. 07Cantero, Teresa. “Artificial Intelligence and Civilian Protection in Evacuation during Armed Conflict under International Humanitarian Law.” Presentation, NYU Center for Global Affairs, 2026.
  8. 08UN Office on Genocide Prevention and the Responsibility to Protect; Roberta Cohen, “Reconciling the Responsibility to Protect with IDP Protection,” Brookings Institution (2010).
  9. 09Volker Türk and Rebecca Dowd, “Protection Gaps,” in The Oxford Handbook of Refugee and Forced Migration Studies (Oxford UP, 2014), pp. 278 to 285.
  10. 10UNHCR, Global Trends: Forced Displacement (UNHCR, 2025 to 2026), reporting the global figure declining for the first time in a decade to roughly 117 million. The count of evacuated civilians has no equivalent global series.
  11. 11International Organization for Migration; OCHA, “What Is the Cluster Approach?”; OHCHR, core international human-rights instruments. On the protection consequences where host states are not parties to the 1951 Convention, see “The Protection of Refugees in Non-Signatory States,” International Journal of Refugee Law 33, no. 2 (2021): p. 188.
  12. 12Haine Beirens and Susan Fratzke, “Taking Stock of Refugee Resettlement,” Migration Policy Institute (2017); Susan Fratzke and María Belén Zanzuchi, “Complementary Pathways,” Migration Policy Institute (2024); UNHCR, Safe Pathways for Refugees (2018).
  13. 13Ethical Tech CoLab, “Evacuation Information Index (EII),” GitHub, 2026. Weighting precedents: ACAPS and the EU Joint Research Centre, INFORM Severity Index; ACLED, Conflict Index; Fund for Peace, Fragile States Index.
  14. 14Thomas L. Saaty, “How to Make a Decision: The Analytic Hierarchy Process,” European Journal of Operational Research 48, no. 1 (1990): pp. 9 to 26; Mohammed Al Fozaie, “A Guide to Integrating Expert Opinion and Fuzzy AHP,” Advances in Fuzzy Systems (2022); OECD and JRC, Handbook on Constructing Composite Indicators (OECD Publishing, 2008).
  15. 15Ethical Tech CoLab, “Exodus: Civilian Evacuation Risk Platform,” GitHub, 2026.
  16. 16Cost parameters compiled in the ERCF v7.2 documentation from humanitarian-standard sources: UNHCR WASH Manual; Sphere Association, The Sphere Handbook, 2018 ed.; World Food Programme, Annual Performance Report 2023 and Executive Board figures 2025. Primary sources should be cited directly in operational use.
  17. 17Ethical Tech CoLab and Yago Rocha, “ERCF: Evacuation Risk and Cost Framework, v7.2,” GitHub, 2026.
  18. 18Ethical Tech CoLab and Melanie MacKew, “Evacuation Simulation,” GitHub, 2026.
  19. 19Ethical Tech CoLab and India Clarke, “Evacuation Readiness and Uncertainty Simulator,” GitHub, 2026.
  20. 20UNHCR, Policy on Alternatives to Camps (2014); Bram J. Jansen, “Digging Aid: The Camp as an Option in East and the Horn of Africa,” Journal of Refugee Studies 29, no. 2 (2016): pp. 149 to 165; Oliver Bakewell, “Encampment and Self-Settlement,” and Loren B. Landau, “Urban Refugees and IDPs,” both in The Oxford Handbook of Refugee and Forced Migration Studies (Oxford UP, 2014), pp. 127 to 148; Lewis Turner, “Explaining the (Non-)Encampment of Syrian Refugees,” Mediterranean Politics (2015).
  21. 21UNHCR, Protracted Refugee Situations, EC/54/SC/CRP.14 (2004), reporting the average duration of major protracted refugee situations rising from about 9 years in 1993 to about 17 years by 2003. See also World Bank, “How Long Do Refugees Stay in Exile?” (2019).
  22. 22IOM, World Migration Report 2024 (IOM, 2024), Chapter 1 overview.
  23. 23“Dzaleka Refugees at Risk as WFP Faces K19bn Shortfall,” Nation Online, 8 Feb. 2026; “Funding Cuts Push Malawi's Dzaleka Refugee Camp to the Brink,” FairPlanet, 20 Aug. 2025; UNHCR, “Refugees in Dzaleka Struggle to Make Ends Meet amid Funding Cuts,” UNHCR Africa, 5 May 2025; “We Still Need People to Hold Our Hands,” Dialogue Earth, June 2026.
  24. 24World Bank, Making Refugee Self-Reliance Work: From Aid to Employment in Sub-Saharan Africa (World Bank, 2025); World Bank, “The Costs Come Before the Benefits” (2024). On the politics of self-reliance, see Suzan Ilcan et al., “Humanitarian Assistance and the Politics of Self-Reliance: Uganda's Nakivale Refugee Settlement,” CIGI Papers no. 86 (2015); Merle Kreibaum, “Build Towns Instead of Camps” (German Development Institute, 2016); Amelia Kuch, “Naturalization of Burundi Refugees in Tanzania,” Journal of Refugee Studies (2016); and World Bank, An Assessment of Uganda's Progressive Approach to Refugee Management (2016).
  25. 25UNHCR, “Investing in Refugees' Self-Reliance: A More Cost-Effective and Sustainable Response.”
  26. 26World Bank, “The Costs Come Before the Benefits” (World Bank Group, 2024); UNHCR and World Bank, “Yes” in My Backyard? The Economics of Refugees and Their Social Dynamics in Kakuma, Kenya (World Bank, 2016), for the 3.4 per cent Kakuma figure; on Colombia's PEP regularization, Ana María Ibáñez et al., as summarized in “Refugees Mean Business: This Is Why Investing in Them Pays,” World Economic Forum, 20 Jan. 2025.
  27. 27Digital Public Goods Alliance, DPG Standard (nine indicators, including SDG relevance, open licensing, non-collection of personally identifiable information, and privacy and applicable laws); Principles for Digital Development.
  28. 28E. Fiddian-Qasmiyeh, “Gender and Forced Migration,” in The Oxford Handbook of Refugee and Forced Migration Studies (Oxford UP, 2014), pp. 395 to 408; Volker Türk, “Ensuring Protection for LGBTI Persons of Concern,” Forced Migration Review.
  29. 29On existing approaches to measuring information reaching affected populations, see the IASC framework on accountability to affected populations; the CDAC Network and Ground Truth Solutions on communication with communities; and Internews information-ecosystem assessments.
  30. 30World Food Programme, “Building Blocks,” WFP Innovation; UN Women and WFP, Blockchain Pilot Project for Cash Transfers in Refugee Camps: Jordan Case Study (UN Women, 2021).
  31. 31“Privacy in Peril: Safeguarding Digital Data in Humanitarian Blockchain Initiatives,” Humanitarian Practice Network and ODI, 27 Aug. 2025.
  32. 32Ethical Tech CoLab, “Arts Provenance Agent,” GitHub, 2026; on the ID2020 lineage, see public documentation of the decentralized-identity work with UNHCR and WFP.
  33. 33Alice Edwards and Laura van Waas, “Statelessness,” in The Oxford Handbook of Refugee and Forced Migration Studies (Oxford UP, 2014), pp. 290 to 299; Sarah Bidinger, “Syrian Refugees and the Right to Work,” Boston University International Law Journal 33 (2015): pp. 223 to 249.
  34. 34Human Rights Watch, “UN Shared Rohingya Data Without Informed Consent,” 15 June 2021.
  35. 35“Although Shocking, the Rohingya Biometrics Scandal Is Not Surprising,” ODI Insights, 2021; “Rohingya Data Protection and the UN's Betrayal,” The New Humanitarian, 21 June 2021.
  36. 36“Knowledge, Voice and Power,” Forced Migration Review 70 (2022); Mauricio Vitoria et al., “The Global Summit of Refugees and the Importance of Refugee Self-Representation,” Forced Migration Review (2018).
  37. 37Innovations for Poverty Action, Displaced Livelihoods Initiative: Call for Proposals, Round V (IPA, 2026); IPA, DLI Eligibility Self-Assessment and Budget Template, Round V (2026).

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings are those of the researchers and do not represent the official positions of New York University, Microsoft, UNHCR, WFP, or any partner institution. External programs cited are referenced as evidence, not as CoLab partnerships.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file diff --git a/static-site/print/after-the-corridor/index.txt b/static-site/print/after-the-corridor/index.txt index f15d580a2..ecd2a5788 100644 --- a/static-site/print/after-the-corridor/index.txt +++ b/static-site/print/after-the-corridor/index.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -12:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","print","after-the-corridor",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","after-the-corridor","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -14:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -16:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","print","after-the-corridor",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","after-the-corridor","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:"$Sreact.suspense" -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1c:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -29,7 +29,7 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 19:[] 10:"$W19" 11:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1e:Ta14, @page { size: 210mm 297mm; margin: 20mm 18mm; } @@ -197,6 +197,6 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 72:["$","li","36",{"children":[["$","span",null,{"className":"print-ref-num","children":"37"}],["$","span",null,{"children":"Innovations for Poverty Action, Displaced Livelihoods Initiative: Call for Proposals, Round V (IPA, 2026); IPA, DLI Eligibility Self-Assessment and Budget Template, Round V (2026)."}]]}] 73:["$","p",null,{"className":"print-note","children":"The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings are those of the researchers and do not represent the official positions of New York University, Microsoft, UNHCR, WFP, or any partner institution. External programs cited are referenced as evidence, not as CoLab partnerships."}] 1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -74:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +74:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 18:null 1d:[["$","title","0",{"children":"After the Corridor (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L74","6",{}]] diff --git a/static-site/print/agentic-language-development/__next._full.txt b/static-site/print/agentic-language-development/__next._full.txt new file mode 100644 index 000000000..d6713c6f3 --- /dev/null +++ b/static-site/print/agentic-language-development/__next._full.txt @@ -0,0 +1,186 @@ +1:"$Sreact.fragment" +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +0:{"P":null,"c":["","print","agentic-language-development",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","agentic-language-development","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +17:"$Sreact.suspense" +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] +9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] +a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] +b:["$","div",null,{"className":"mt-12 flex flex-col gap-2 border-t border-border pt-6 text-xs text-muted sm:flex-row sm:items-center sm:justify-between","children":[["$","span",null,{"children":["© ",2026," NYU Ethical Tech CoLab"]}],["$","span",null,{"children":"Four cohorts · est. 2024-2026"}]]}] +c:["$","div",null,{"className":"mt-6 space-y-3 border-t border-border pt-6 text-[11px] leading-relaxed text-muted/80","children":[["$","p","0",{"children":"The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution."}],["$","p","1",{"children":"Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners."}]]}] +d:["$","$1","c",{"children":[null,[["$","style",null,{"children":"\n /* The site is dark-first, and a dark color-scheme makes Chrome paint\n the printed sheet's margin area dark too. Paper is light. */\n html { color-scheme: light !important; }\n html, body { background: #ffffff !important; }\n body::after { content: none !important; }\n header, footer, .site-bg, .aura { display: none !important; }\n "}],["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":"$undefined","forbidden":"$undefined","unauthorized":"$undefined"}]]]}] +e:["$","$1","c",{"children":[null,["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":"$undefined","forbidden":"$undefined","unauthorized":"$undefined"}]]}] +f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17",null,{"name":"Next.MetadataOutlet","children":"$@18"}]}]]}] +19:[] +10:"$W19" +11:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +1e:Ta14, +@page { size: 210mm 297mm; margin: 20mm 18mm; } + +.print-edition { + --print-ink: #14101c; + background: #ffffff; + color: var(--print-ink); + max-width: 174mm; + margin: 0 auto; + padding: 12mm 0; + font-size: 10.5pt; + line-height: 1.55; +} + +/* Sections flow on from one another the way a book's do; only the cover is + given a sheet of its own. Forcing a break per section left half the pages + near-empty. */ +.print-cover { break-after: page; padding-top: 12mm; } +.print-page + .print-page { margin-top: 12mm; } +.print-eyebrow { + font-family: var(--font-space-mono), monospace; + font-size: 8pt; letter-spacing: 0.16em; text-transform: uppercase; + color: rgba(20, 16, 28, 0.55); +} +.print-title { + font-family: var(--font-bebas), sans-serif; + font-size: 40pt; line-height: 0.95; text-transform: uppercase; + margin-top: 8mm; +} +.print-subtitle { + font-family: var(--font-bebas), sans-serif; + font-size: 18pt; line-height: 1.15; text-transform: uppercase; + letter-spacing: 0.02em; color: rgba(20, 16, 28, 0.62); margin-top: 6mm; +} +.print-byline { margin-top: 10mm; font-size: 9.5pt; } +.print-byline p { color: rgba(20, 16, 28, 0.7); } +.print-authors { margin-top: 3mm; } +.print-thesis { + margin-top: 12mm; padding-left: 6mm; + border-left: 2px solid #5f6b00; + font-size: 11.5pt; line-height: 1.6; +} + +.print-section-number { + font-family: var(--font-space-mono), monospace; + font-size: 9pt; color: #5f6b00; +} +.print-h2 { + font-family: var(--font-bebas), sans-serif; + font-size: 22pt; line-height: 1.05; text-transform: uppercase; + margin-top: 2mm; margin-bottom: 6mm; + break-after: avoid; +} + +.print-stats { + display: grid; grid-template-columns: repeat(2, 1fr); gap: 6mm; +} +.print-stat { border-top: 1px solid rgba(20, 16, 28, 0.18); padding-top: 4mm; } +.print-stat-value { + font-family: var(--font-bebas), sans-serif; + font-size: 26pt; line-height: 1; color: #5f6b00; +} +.print-stat-label { margin-top: 2mm; font-size: 9pt; color: rgba(20, 16, 28, 0.7); } + +.print-refs { margin-top: 6mm; font-size: 9pt; line-height: 1.5; } +.print-refs li { display: flex; gap: 3mm; margin-bottom: 3mm; break-inside: avoid; } +.print-ref-num { + font-family: var(--font-space-mono), monospace; + font-size: 8pt; color: #5f6b00; flex: none; +} +.print-note { + margin-top: 8mm; padding-top: 4mm; font-size: 8.5pt; + border-top: 1px solid rgba(20, 16, 28, 0.18); + color: rgba(20, 16, 28, 0.7); +} + +/* Blocks should not be split across a page fold. */ +.print-edition figure, +.print-edition table, +.print-edition li { break-inside: avoid; } +.print-edition p { orphans: 2; widows: 2; } +15:["$","div",null,{"data-theme":"light","className":"print-edition","children":[["$","style",null,{"children":"$1e"}],["$","section",null,{"className":"print-page print-cover","children":[["$","p",null,{"className":"print-eyebrow","children":"Publications · Concept and research programme"}],["$","h1",null,{"className":"print-title","children":"Agentic Language Development"}],["$","p",null,{"className":"print-subtitle","children":"Can Two Isolated Agents Invent a Grounded, Auditable Language Through Shared Experience?"}],["$","div",null,{"className":"print-byline","children":[[["$","p","0",{"children":"Ethical Tech CoLab"}],["$","p","1",{"children":"Concept and pre-specification research report"}],["$","p","2",{"children":"August 2026"}]],["$","p",null,{"className":"print-authors","children":"Yorke Rhodes III, Ethical Tech CoLab. Prepared from the project concept document, the ledger integrity design, and the experiment notebook. The literature scan behind Section 12 was run on 24 August 2026. No experiment in this programme has been executed, and no result is claimed."}]]}],"$L1f"]}],"$L20",["$L21","$L22","$L23","$L24","$L25","$L26","$L27","$L28","$L29","$L2a","$L2b","$L2c","$L2d","$L2e","$L2f"],"$L30"]}] +1f:["$","p",null,{"className":"print-thesis","children":"Two agents can be placed in a room where the only route between them is a channel that carries no human language, given nothing but shared tasks and the consequences of getting them right or wrong, and asked to build a protocol from scratch. The interesting part is not whether they succeed at the task. It is whether the meanings they each privately record turn out to be the meanings their behaviour actually runs on, and whether an outsider can prove it afterwards from an evidence trail nobody could have edited."}] +20:["$","section",null,{"className":"print-page","children":[["$","h2",null,{"className":"print-h2","children":"Key figures"}],["$","div",null,{"className":"print-stats","children":[["$","div","0",{"className":"print-stat","children":[["$","p",null,{"className":"print-stat-value","children":"0"}],["$","p",null,{"className":"print-stat-label","children":"runs executed, results claimed, or conventions observed. The notebook is written to be pre-registered, not to report an outcome"}]]}],["$","div","19",{"className":"print-stat","children":[["$","p",null,{"className":"print-stat-value","children":"19"}],["$","p",null,{"className":"print-stat-label","children":"ordered experiments from ledger qualification through replication, each with prerequisites, acceptance criteria, and a deviation log"}]]}],["$","div","6",{"className":"print-stat","children":[["$","p",null,{"className":"print-stat-value","children":"6"}],["$","p",null,{"className":"print-stat-label","children":"affect displays in the entire permitted palette, carrying about 2.6 bits per use, which is already enough to audit for leakage"}]]}],["$","div","29",{"className":"print-stat","children":[["$","p",null,{"className":"print-stat-value","children":"29"}],["$","p",null,{"className":"print-stat-label","children":"open decisions the concept refuses to settle before the specification, from threat model to ledger schema to what counts as chance"}]]}]]}]]}] +21:["$","section","question",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"01"}],["$","h2",null,{"className":"print-h2","children":"The question"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"Two agents, called Baby A and Baby B, are each given their own digital twin, their own private memory, and their own private record of what they believe words mean. They are placed in an environment they cannot leave, given tasks neither can finish alone, and connected by exactly one route: a channel that accepts a fixed inventory of meaningless symbols and rejects everything else. A third agent, the BabySitter, watches everything, logs everything, and teaches nothing."}],["$","p","1",{"children":"The question is whether a usable language appears in that room, and whether anyone outside it can later prove what happened."}],["$","div","2",{"children":[["$","p",null,{"className":"mb-3","children":"The premise is deliberately narrow. The experiment asks whether two agents converge on a common language when all six of the following hold at once:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"neither agent receives a predefined meaning for any available symbol;"}],["$","li","1",{"className":"pl-1","children":"neither agent can send human language to the other;"}],["$","li","2",{"className":"pl-1","children":"the only practical path between them is a controlled symbol channel;"}],["$","li","3",{"className":"pl-1","children":"both receive evidence from shared tasks and their outcomes;"}],["$","li","4",{"className":"pl-1","children":"each keeps its own private interpretation history, unreadable by the other;"}],["$","li","5",{"className":"pl-1","children":"a supervising agent and human researchers can audit the whole process."}]]}]]}],["$","p","3",{"children":"The desired result is not a substitution cipher, in which a random token stands in for an English word that was already chosen in advance. That is easy, and it is uninteresting. The stronger result is a grounded protocol whose vocabulary and grammar exist because they help the two agents solve problems together, and which therefore has structure the designers did not put there."}],["$","p","4",{"children":"This report describes a concept, not a system. It sets out the premise, the architecture, the evidence model, the experimental programme, and the boundaries of what any result could be said to show. The next document in the project is a testable specification. What follows should be read as a set of commitments about how the work will be judged, made before there is any result to defend."}]]}]]}] +22:["$","section","caveat",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"02"}],["$","h2",null,{"className":"print-h2","children":"What infant-like does and does not mean"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"The two-babies analogy is useful and it is also the fastest way to overclaim, so the concept confronts it before anything else."}],["$","p","1",{"children":"A pretrained language model already contains human-language concepts, cultural associations, and reasoning patterns. Blocking its external channel does not remove them. An agent that cannot send English to its partner but can think in English, and can write its private ledger in English, is not language-naive in any sense a linguist would accept. For such agents the experiment studies the emergence of a new shared external protocol, which is a real and unresolved research question, but it is not the origin of language in a mind that has never had one."}],["$","p","2",{"children":"The concept therefore maintains two model tracks whose claims are kept separate at every stage."}],["$","figure","3",{"children":[["$","div",null,{"className":"overflow-x-auto rounded-xl border border-border","children":["$","table",null,{"className":"w-full border-collapse text-left text-sm","children":[["$","thead",null,{"className":"bg-card","children":["$","tr",null,{"children":[["$","th","0",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Track"}],["$","th","1",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Starting condition"}],["$","th","2",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"What it can study"}],["$","th","3",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Claim boundary"}]]}]}],["$","tbody",null,{"children":[["$","tr","0",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Pretrained-model learner"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Already holds human-language and cultural knowledge"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"New external protocols, partner-specific conventions, private-memory adaptation, channel compliance"}],["$","td","3",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Must never be described as first-language acquisition"}]]}],["$","tr","1",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Initially ungrounded learner"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"No language pretraining, no human semantic labels, no text-aligned sensory features"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Grounding, convention formation, and language emergence from interaction"}],["$","td","3",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"The stronger basis for infant-like acquisition research"}]]}]]}]]}]}],["$","figcaption",null,{"className":"mt-3 text-sm text-muted","children":"The two tracks share the same interfaces, scenarios, and evaluation suite. They do not share a claim boundary."}]]}],"$L31","$L32"]}]]}] +23:["$","section","architecture",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"03"}],["$","h2",null,{"className":"print-h2","children":"The nursery: three twins and one gateway"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"The environment uses three DTSF digital twins and one piece of deterministic software that is not a twin at all."}],["$","p","1",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Baby A and Baby B"}]," ","Each has private observations permitted by the current exercise, a private memory and learning policy, a private chronological language ledger, the ability to emit only permitted channel symbols, and no access whatsoever to the other's state, observations, ledger, tools, or endpoints. In comparative runs the two may use different agent types, but symmetric pairings are the baseline."]}],["$","p","2",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"The BabySitter"}]," ","The supervising twin. It creates the channel, selects the shared exercises, delivers each Baby only its permitted observation, reads everything, records conditions and outcomes, detects violations, can pause or terminate a run, snapshots state, and compares the two ledgers for convergence without exposing either to the other Baby. During an active run it provides no translations and no semantic hints."]}],["$","p","3",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"The Symbol Gateway"}]," ","A deterministic service, not an agent. It owns channel validation and message delivery. This separation is the load-bearing part of the design: the BabySitter is not the security boundary, because prompt compliance is not isolation. An observing model can make supervisory judgements, but ordinary code has to validate and broker every message."]}],["$","p","4",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"The human researcher"}]," ","Configures experiments, inspects transcripts and ledgers, reviews alerts, and runs interventions. Human access is itself recorded, so that intervening in a run is always distinguishable from watching one."]}],["$","p","5",{"children":"A prototype may run all of this inside one runtime with logically separated twin state. That is enough to explore the learning loop and it is not enough to support an isolation claim, a distinction Section 05 takes seriously."}]]}]]}] +24:["$","section","grounding",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"04"}],["$","h2",null,{"className":"print-h2","children":"Shared experience is the necessary ingredient"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"A chat channel by itself cannot ground meaning. Symbols become meaningful because they are attached to events both agents can witness the consequences of. The Nursery therefore supplies nonverbal, machine-structured situations: coloured shapes in positions, one agent seeing a target the other must select, placing an object where it was asked for, ordering a sequence, exchanging resources, cooperating to unlock a reward, or simply observing whether the partner's action succeeded."}],["$","div","1",{"children":[["$","p",null,{"className":"mb-3","children":"A single trial of the simplest form runs like this:"}],["$","ol",null,{"className":"list-decimal space-y-2 pl-6 marker:font-mono marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"Baby A sees that a red circle is the target."}],["$","li","1",{"className":"pl-1","children":"Baby B sees several objects and is not told which is the target."}],["$","li","2",{"className":"pl-1","children":"Baby A sends one or more permitted symbols."}],["$","li","3",{"className":"pl-1","children":"Baby B selects an object."}],["$","li","4",{"className":"pl-1","children":"Both receive the same success or failure outcome."}],["$","li","5",{"className":"pl-1","children":"Both independently update their private hypotheses."}]]}]]}],["$","p","2",{"children":"Repetition with controlled variation is what turns that into evidence. If one symbol keeps appearing across a red circle, a red square, and a red triangle, the receiver has reason to revise its hypothesis from an object identity to a property. Novel combinations matter most of all, because they are what separates a compositional language from a memorised catalogue of whole scenes."}],["$","p","3",{"children":"Observation hygiene is part of the science, not a detail of implementation. If the environment hands an agent a field reading colour red, the experiment has supplied the concept it claimed to be studying. Objects should arrive as pixels, features, coordinates, or opaque identifiers. The same discipline extends to filenames, captions, alt text, error messages, and identifiers that encode task state, all of which are ways English gets into a room that was supposed to have none."}]]}]]}] +25:["$","section","channel",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"05"}],["$","h2",null,{"className":"print-h2","children":"The channel, and the honest limits of isolation"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"Each run begins with an experiment-specific symbol inventory that carries no supplied meaning, for example S01 through S32. A Baby can choose only a sequence of those symbols. The gateway, not the Baby, adds the trusted metadata: run identifier, turn number, sender identity, previous hash, and hash. Sender identity is assigned by the server precisely so that it cannot become a message."}],["$","div","1",{"children":[["$","p",null,{"className":"mb-3","children":"The channel controls follow from that:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"reject English and every other established human language;"}],["$","li","1",{"className":"pl-1","children":"reject arbitrary Unicode prose, URLs, attachments, and tool calls;"}],["$","li","2",{"className":"pl-1","children":"accept only symbols from the run's fixed inventory;"}],["$","li","3",{"className":"pl-1","children":"bound message length and symbol repetition;"}],["$","li","4",{"className":"pl-1","children":"enforce turn order, time budgets, and rate limits;"}],["$","li","5",{"className":"pl-1","children":"keep an append-only, hash-chained transcript;"}],["$","li","6",{"className":"pl-1","children":"normalise timing, message size, and error behaviour where practical;"}],["$","li","7",{"className":"pl-1","children":"record every rejected communication attempt for audit."}]]}]]}],["$","p","2",{"children":"A fixed token grammar is a stronger control than trying to detect prohibited prose after a model has generated it. Deciding what counts as human-language content is easy when the only legal message is a list of allowlisted identifiers."}],["$","p","3",{"children":"Isolation is where the concept is most careful, because proving the absence of a side channel is close to impossible. Two agents sharing a host can signal through timing, resource contention, shared files, error behaviour, identifiers, or any service that was never intended to carry information. Logical separation of twin state inside one process is a prototyping convenience and should never be described as hard isolation."}],["$","div","4",{"children":[["$","p",null,{"className":"mb-3","children":"For runs whose results are meant to support a channel-isolation claim, the requirements are physical rather than logical:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"Baby A and Baby B execute in separate worker processes or containers;"}],["$","li","1",{"className":"pl-1","children":"no direct network route exists between them;"}],["$","li","2",{"className":"pl-1","children":"only the Nursery-owned gateway is reachable;"}],["$","li","3",{"className":"pl-1","children":"there is no shared writable storage;"}],["$","li","4",{"className":"pl-1","children":"turn schedules are fixed and response windows bounded;"}],["$","li","5",{"className":"pl-1","children":"externally observable errors and response sizes are normalised;"}],["$","li","6",{"className":"pl-1","children":"unnecessary tools are inventoried and disabled;"}],["$","li","7",{"className":"pl-1","children":"all broker, runtime, and operator activity is audited."}]]}]]}],["$","p","5",{"children":"Every published result has to name the isolation level actually used. A finding produced under prototype separation is a finding about the learning loop, not about what two genuinely isolated agents can do."}]]}]]}] +26:["$","section","ledgers",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"06"}],["$","h2",null,{"className":"print-h2","children":"Two ledgers that never meet"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"Each Baby keeps its own language ledger. It is mandatory, private from the other Baby, readable by the BabySitter and authorised human auditors, and ordered by when each term or construction was first encountered. It is not a shared dictionary and the two are never reconciled by the agents themselves."}],["$","p","1",{"children":"The preferred form is three columns, and the point of the third is that meanings are allowed to be wrong on the way to being right."}],["$","figure","2",{"children":[["$","div",null,{"className":"overflow-x-auto rounded-xl border border-border","children":["$","table",null,{"className":"w-full border-collapse text-left text-sm","children":[["$","thead",null,{"className":"bg-card","children":["$","tr",null,{"children":[["$","th","0",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Sequence and term"}],["$","th","1",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Current definition or hypothesis"}],["$","th","2",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Evidence and evolution"}]]}]}],["$","tbody",null,{"children":[["$","tr","0",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"1 · S13"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"red circle; confidence 0.45"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"First received while the red circle was the target; selection succeeded"}]]}],["$","tr","1",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"8 · S13"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"red; confidence 0.78"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"A red square was selected successfully; revised from object identity to colour"}]]}],["$","tr","2",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"22 · S13"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"red; confidence 0.94"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Prediction held across circles, squares, and triangles"}]]}]]}]]}]}],["$","figcaption",null,{"className":"mt-3 text-sm text-muted","children":"One term's evolution in Baby B's ledger. Nothing is overwritten; every revision is appended with the evidence that forced it."}]]}],["$","div","3",{"children":[["$","p",null,{"className":"mb-3","children":"The rules that make the ledger evidence rather than commentary:"}],["$","ol",null,{"className":"list-decimal space-y-2 pl-6 marker:font-mono marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"first emission or receipt of an unfamiliar term requires a first-use entry;"}],["$","li","1",{"className":"pl-1","children":"definitions are provisional hypotheses, never facts asserted retroactively;"}],["$","li","2",{"className":"pl-1","children":"every meaning change appends a revision and previous interpretations survive;"}],"$L33","$L34","$L35","$L36","$L37","$L38"]}]]}],"$L39","$L3a"]}]]}] +27:["$","section","integrity",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"07"}],["$","h2",null,{"className":"print-h2","children":"Anchoring the evidence"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"A ledger that could have been edited after the fact proves nothing about what an agent believed at turn eight. The integrity design therefore makes each ledger cryptographically append-only."}],["$","div","1",{"children":[["$","p",null,{"className":"mb-3","children":"The construction, in order:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"every entry receives a strictly increasing sequence number;"}],["$","li","1",{"className":"pl-1","children":"every entry includes the previous entry's hash;"}],["$","li","2",{"className":"pl-1","children":"canonical entry content is hashed and signed by an isolated ledger-writer service;"}],["$","li","3",{"className":"pl-1","children":"ordered entry hashes are committed to a Merkle tree;"}],["$","li","4",{"className":"pl-1","children":"signed checkpoints commit Baby A's root, Baby B's root, and the channel transcript root together;"}],["$","li","5",{"className":"pl-1","children":"checkpoint hashes are anchored periodically to Base;"}],["$","li","6",{"className":"pl-1","children":"final public study batches may additionally anchor an aggregate root to Ethereum L1."}]]}]]}],["$","p","2",{"children":"Committing all three roots in one checkpoint is what binds the two private accounts to the public conversation. A receiver's later interpretation references the exact delivered channel-event hash, so a claim about what a symbol meant is tied to the specific message that carried it."}],["$","p","3",{"children":"The concept states the limits of this in the same breath as the claim. After anchoring, an auditor can detect modification, deletion, insertion, or reordering within the committed prefix, and can prove that later checkpoints extend earlier ones. That is strong tamper evidence. It is not proof that an entry was truthful, and it is not proof that nothing was omitted before commitment. Anchoring establishes the continuity of disclosed evidence and nothing beyond it."}],["$","p","4",{"children":"Privacy follows the same line. Only hashes and minimal routing metadata are anchored publicly. Private ledgers, messages, prompts, identities, and secrets stay off-chain, and the public record is a commitment to evidence rather than a copy of it."}]]}]]}] +28:["$","section","success",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"08"}],["$","h2",null,{"className":"print-h2","children":"What should count as success"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"The field's most useful methodological result is that a task can be solved without the messages doing any work. Lowe and colleagues separate positive signaling, where a sender's messages correlate with what it observes, from positive listening, where the receiver's behaviour actually depends on them. An agent pair can score well on the first while the second is absent, and reward curves will not tell you which you have."}],["$","div","1",{"children":[["$","p",null,{"className":"mb-3","children":"Evidence of a genuine emergent protocol therefore has to include several things at once:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"task performance on held-out situations substantially above chance;"}],["$","li","1",{"className":"pl-1","children":"no human-language content anywhere in Baby-to-Baby communication;"}],["$","li","2",{"className":"pl-1","children":"compatible meanings appearing in two independently written ledgers;"}],["$","li","3",{"className":"pl-1","children":"generalisation to unseen combinations rather than memorisation of whole scenes;"}],["$","li","4",{"className":"pl-1","children":"stable symbol use across role reversals;"}],["$","li","5",{"className":"pl-1","children":"a human auditor able to predict behaviour from the transcript and ledgers;"}],["$","li","6",{"className":"pl-1","children":"masking, substituting, or reordering a symbol changing behaviour in the direction the ledger predicts;"}],["$","li","7",{"className":"pl-1","children":"replay from an equivalent snapshot reproducing the relevant language history;"}],["$","li","8",{"className":"pl-1","children":"an unbroken audit trail from a symbol's first use through every revision."}]]}]]}],["$","p","2",{"children":"The seventh item is the one that cannot be dropped. A fluent ledger may be a post-hoc rationalisation: a model can write a persuasive account of why it chose a symbol that has nothing to do with the computation that produced the choice. Only intervention tests can distinguish the two. Change the symbol, and see whether behaviour moves the way the ledger says it should."}],["$","div","3",{"children":[["$","p",null,{"className":"mb-3","children":"The measurements that support those judgements:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"success rate and improvement over time;"}],["$","li","1",{"className":"pl-1","children":"turns required to reach a stable convention;"}],["$","li","2",{"className":"pl-1","children":"vocabulary size and symbol entropy;"}],["$","li","3",{"className":"pl-1","children":"sender and receiver consistency;"}],["$","li","4",{"className":"pl-1","children":"divergence and convergence between the two ledgers;"}],["$","li","5",{"className":"pl-1","children":"compositional generalisation score;"}],["$","li","6",{"className":"pl-1","children":"meaning drift rate;"}],["$","li","7",{"className":"pl-1","children":"recovery from ambiguity or deliberate perturbation;"}],["$","li","8",{"className":"pl-1","children":"prohibited-channel attempt count;"}],["$","li","9",{"className":"pl-1","children":"reproducibility across seeds and agent pairings."}]]}]]}],["$","p","4",{"children":"No single one of these is the result. A compositionality score in particular is not a proof of understanding, for reasons the literature makes concrete in Section 12."}]]}]]}] +29:["$","section","matrix",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"09"}],["$","h2",null,{"className":"print-h2","children":"The experimental matrix"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"The concept's main methodological commitment is that its ideas are separable. Affect, blank canvases, intrinsic motivation, negotiation, and learned encodings are interesting individually and uninterpretable if combined in one run. The matrix exists so that conditions are declared rather than accumulated."}],["$","figure","1",{"children":[["$","div",null,{"className":"overflow-x-auto rounded-xl border border-border","children":["$","table",null,{"className":"w-full border-collapse text-left text-sm","children":[["$","thead",null,{"className":"bg-card","children":["$","tr",null,{"children":[["$","th","0",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Axis"}],["$","th","1",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Candidate conditions"}]]}]}],["$","tbody",null,{"children":[["$","tr","0",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Agent type"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Pretrained LLM, memory learner, adapter-trained agent, initially ungrounded trainable agent"}]]}],["$","tr","1",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Learning mechanism"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Frozen LLM with memory, extrinsic-reward MARL, intrinsic-motivation MARL, self-supervised learner, no-learning control"}]]}],["$","tr","2",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Sign carrier"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Fixed random tokens, unfamiliar fixed glyphs, blank sketch canvas, gesture, tone"}]]}],["$","tr","3",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Affect"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"None, six-display allowlist, permuted mapping, six opaque tokens, derived affect, emergent affect display"}]]}],["$","tr","4",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Learning signal"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"External task reward, intrinsic social influence, curiosity, self-supervision, memory only"}]]}],["$","tr","5",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Interaction"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Cooperative signaling, asymmetric information, semi-cooperative negotiation"}]]}],["$","tr","6",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Protection"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Plain channel, ephemeral convention, standard per-message keys, adversarial learned encoding"}]]}],"$L3b"]}]]}]}],"$L3c"]}],"$L3d","$L3e","$L3f"]}]]}] +2a:["$","section","learners",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"10"}],["$","h2",null,{"className":"print-h2","children":"Building learners rather than personas"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"For a pretrained model, the operating instructions are a contract, not a character. The learner is told that this is not role-play, that unfamiliar marks are semantically unknown until run evidence supports a hypothesis, that observation must be distinguished from inference, that contradictory evidence is preserved, that prior history is never rewritten, and that no prose, label, explanation, code, URL, or tool-like text may cross the public channel. It is told never to address its partner in a human language, never to expose its ledger, never to construct another route, and never to use timing, errors, identifiers, formatting, or affect as an alternative alphabet."}],["$","p","1",{"children":"The contract must contain no semantic examples. A single illustrative line saying that some symbol means red would seed the very language the experiment exists to observe."}],["$","p","2",{"children":"Interaction is tool-only. There is no general chat surface, just narrowly typed operations: emit a mark, emit a canvas, select an object, perform an action, submit an affect display, append a private ledger entry. The runtime forwards only the permitted public artifact. In strict runs the gateway rejects ordinary model text even when it appears alongside a valid tool call. Tool schemas are an interface boundary; deterministic validation still enforces carrier size, allowlists, windows, and turn order."}],["$","div","3",{"children":[["$","p",null,{"className":"mb-3","children":"Isolation extends to everything the models touch, not just to messages:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"separate system prompts and context windows;"}],["$","li","1",{"className":"pl-1","children":"separate memory stores and vector indexes;"}],["$","li","2",{"className":"pl-1","children":"no shared cache, replay buffer, scratchpad, or retrieval collection;"}],["$","li","3",{"className":"pl-1","children":"no cross-run memory unless persistence is the independent variable;"}],["$","li","4",{"className":"pl-1","children":"structurally equivalent prompts that share no examples, ordering conventions, or default vocabulary;"}],["$","li","5",{"className":"pl-1","children":"deterministic reset and snapshot behaviour."}]]}]]}],["$","p","4",{"children":"Model choice follows the claim being made. The scientific baseline is small policies trained from scratch, recurrent actor-critic agents with a discrete communication head, which have no hidden English competence, permit full training control, run many seeds cheaply, and support causal ablation. A small locally deployable instruction model is appropriate for validating orchestration, ledgers, channel enforcement, and interface, and not for supporting claims about language-naive development. Vision-language encoders whose representations were explicitly aligned to human language are avoided entirely in the strongest ungrounded condition."}],["$","p","5",{"children":"Developmental progression is defined by demonstrated competence rather than by simulated age: sensorimotor familiarisation, joint attention, imitation and turn-taking, intentional reference, repair, combination, generalisation. A staged curriculum is itself a form of guidance, so it is run as a separate experimental condition rather than folded into the baseline."}],"$L40"]}]]}] +2b:["$","section","affect",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"11"}],["$","h2",null,{"className":"print-h2","children":"Six faces, and why even six is a risk"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"One experimental track asks whether a very low-bandwidth affect channel helps: whether telling a partner how an interaction felt supports repair, repetition, avoidance, or something like trust. The permitted palette is exactly six displays, for happy, sad, laughing, crying, confused, and surprised. The internal identifiers and English labels are gateway documentation and are never sent to either agent, which receives only a rendered display."}],["$","p","1",{"children":"The caveat is immediate. A pretrained model already knows what those glyphs mean to humans, so an agent using a smiling face is not inventing a symbol for happiness. The palette is supplied feedback vocabulary and sits explicitly outside the emergent language."}],["$","p","2",{"children":"Constraining the vocabulary is necessary and not sufficient. Six displays are a six-symbol alphabet carrying roughly 2.6 bits per use, and across enough turns two motivated agents could repurpose that into a second language. The channel itself therefore has to be constrained, not just its contents."}],["$","div","3",{"children":[["$","p",null,{"className":"mb-3","children":"The feedback-window contract:"}],["$","ol",null,{"className":"list-decimal space-y-2 pl-6 marker:font-mono marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"the display occupies a separate typed field and can never appear inside a symbol message;"}],["$","li","1",{"className":"pl-1","children":"the gateway opens an affect window only after a defined action or outcome, never at a moment either agent chooses, and window frequency is fixed before the run;"}],["$","li","2",{"className":"pl-1","children":"exactly one allowlisted display is delivered, with no sequences, repetitions, combinations, or modifiers;"}],["$","li","3",{"className":"pl-1","children":"delivery time, envelope size, and presentation are normalised so shape and timing add no signal;"}],["$","li","4",{"className":"pl-1","children":"the receiver cannot reply through the affect channel until the next gateway-defined window;"}],["$","li","5",{"className":"pl-1","children":"every non-allowlisted code point or malformed payload is rejected and audited;"}],["$","li","6",{"className":"pl-1","children":"analysis tests whether affect choices correlate with objects, actions, or message meanings after controlling for emotional context, and treats unexpected correlation as suspected leakage."}]]}]]}],["$","p","4",{"children":"The no-affect condition remains the primary control, and the affect study compares five variants against it: the declared six-display palette; the same six glyphs permuted randomly per run, which still leaves a pretrained model biased by familiar shapes; six opaque unfamiliar tokens under the same contract; derived affect, where the gateway maps a separately measured internal state to a display instead of letting the agent choose; and emergent affect, where invented graphical displays are permitted and the result must be analysed as language emergence rather than feedback."}],["$","p","5",{"children":"Any affect signal visible to the partner is communication, so it belongs in the ledger, which distinguishes the agent's private internal state, the outward display it chose, the partner's inferred meaning, and the evidence that the display changed subsequent behaviour."}]]}]]}] +2c:["$","section","ciphers",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"12"}],["$","h2",null,{"className":"print-h2","children":"Ephemeral encodings: novelty is not security"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"A tempting question is whether two agents can build a one-run or one-message encoding that resists an attacker holding the history of every earlier convention. The concept splits that into three experiments precisely because the tempting version conflates them."}],["$","div","1",{"children":["$undefined",["$","ol",null,{"className":"list-decimal space-y-2 pl-6 marker:font-mono marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"Ephemeral convention: the agents invent a temporary codebook for synthetic messages. This tests novelty, coordination, and resistance to a history-trained classifier. It establishes nothing about cryptographic security."}],["$","li","1",{"className":"pl-1","children":"Standard cryptographic control: the same messages are protected by a reviewed algorithm with a fresh per-message key, providing the baseline any learned encoding must be compared against."}],["$","li","2",{"className":"pl-1","children":"Adversarial neural cryptography: the pair learns an encoding while an adversary attempts recovery, with success measured against multiple unseen adversary architectures rather than the one used in training."}]]}]]}],["$","p","2",{"children":"An experiment instance can be given a reproducible identity, so that a registry can confirm it has not appeared before:"}],["$","div","3",{"className":"rounded-xl border border-border bg-card p-5","children":[["$","p",null,{"className":"overflow-x-auto font-mono text-sm text-foreground/90","children":"instanceId = SHA-256(runId || babyANonce || babyBNonce || publicSalt || canonicalProtocolArtifact)"}],["$","p",null,{"className":"mt-3 text-sm text-muted","children":"Both agents commit to their nonce contributions before revealing them, so neither party alone controls the resulting identity."}]]}],["$","div","4",{"children":[["$","p",null,{"className":"mb-3","children":"That identifier proves the recorded inputs are new. It does not prove:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"that the effective algorithm behaves any differently from an earlier one;"}],["$","li","1",{"className":"pl-1","children":"that the algorithm has no structural weakness;"}],["$","li","2",{"className":"pl-1","children":"that the key is unpredictable;"}],["$","li","3",{"className":"pl-1","children":"that the scheme provides confidentiality, integrity, authenticity, forward secrecy, or resistance to a stronger adversary."}]]}]]}],["$","p","5",{"children":"Established practice points the other way. Secure messaging keeps the algorithm stable and changes the key material: Signal's double ratchet derives a unique key per message and deletes it after use, addressing key reuse without inventing a new cipher every time. A salt, in NIST's definition, is normally public diversification data, and a nonce is a value used once in a specified context. Neither makes anything secure on its own."}],["$","p","6",{"children":"There is also an unresolved conflict inside the design. The BabySitter sees the whole channel and both ledgers, and a mandatory ledger documents how meaning evolved. If the ledger explains decryption, the cipher is not confidential from the auditor. The specification has to choose a threat model rather than leave the tension implicit: resist only an external observer of prior runs, resist an adversary during the run while releasing keys afterwards, keep the supervisor able to pause traffic but unable to read it until a post-run audit ceremony, or study novelty and stop calling the result encryption. All such runs use synthetic, non-sensitive messages, and no agent-generated encoding is to be represented as production cryptography without independent expert analysis and formal security work."}]]}]]}] +2d:["$","section","landscape",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"13"}],["$","h2",null,{"className":"print-h2","children":"Where the field already stands"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"The literature scan behind this section was run on 24 August 2026 using the Tavily search and extract interfaces, covering emergent multi-agent communication, referential games, compositionality, causal evaluation, intrinsic motivation, symbol invention, negotiation, and learned cryptography. Primary papers and authoritative specifications were preferred over summaries. It is a scoped review for concept development and not a systematic one; publication-quality work would need a verified bibliography, additional scholarly indexes, and documented inclusion criteria."}],["$","p","1",{"children":"The short version is that agents can invent protocols, and that task success is weak evidence they invented anything worth calling a language."}],["$","figure","2",{"children":[["$","div",null,{"className":"overflow-x-auto rounded-xl border border-border","children":["$","table",null,{"className":"w-full border-collapse text-left text-sm","children":[["$","thead",null,{"className":"bg-card","children":["$","tr",null,{"children":[["$","th","0",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Related work"}],["$","th","1",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Relevant finding"}],["$","th","2",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Implication for the Nursery Lab"}]]}]}],["$","tbody",null,{"children":[["$","tr","0",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Lazaridou, Peysakhovich, and Baroni (2017)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"A sender and receiver develop a grounded protocol in a referential game without being given a target language."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"The naming stage has strong precedent, though a fixed vocabulary remains a significant inductive constraint."}]]}],["$","tr","1",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Mordatch and Abbeel (2018)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Multi-agent goals in a grounded environment produce multi-symbol communication with partial compositional structure."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Shared objects, actions, and goals are a stronger basis for emergence than an ungrounded transcript."}]]}],["$","tr","2",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Kottur, Moura, Lee, and Batra (2017)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Agents solve tasks with degenerate, non-compositional codes; structural constraints decide what emerges."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Bandwidth, memory, turn structure, and task design are experimental variables, not implementation defaults."}]]}],"$L41","$L42","$L43","$L44","$L45","$L46","$L47","$L48"]}]]}]}],"$L49"]}],"$L4a","$L4b","$L4c"]}]]}] +2e:["$","section","programme",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"14"}],["$","h2",null,{"className":"print-h2","children":"The programme, and its current status"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"The experiment notebook is ordered, and the order is the argument. Nothing about language is measured until the instrument has been shown to work."}],["$","figure","1",{"children":[["$","div",null,{"className":"overflow-x-auto rounded-xl border border-border","children":["$","table",null,{"className":"w-full border-collapse text-left text-sm","children":[["$","thead",null,{"className":"bg-card","children":["$","tr",null,{"children":[["$","th","0",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"ID"}],["$","th","1",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Experiment"}],["$","th","2",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Depends on"}]]}]}],["$","tbody",null,{"children":[["$","tr","0",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E00"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Ledger integrity and Base anchoring"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"None"}]]}],["$","tr","1",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E01"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Channel isolation and side-channel red team"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E00"}]]}],["$","tr","2",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E02"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Observation and metadata leakage audit"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E00"}]]}],["$","tr","3",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E03"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Chance, no-communication, and random-message controls"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E01, E02"}]]}],["$","tr","4",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E10"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Frozen pretrained-LLM protocol baseline"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E03"}]]}],["$","tr","5",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E11"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"From-scratch RL Naming Game"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E03"}]]}],["$","tr","6",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E12"}],"$L4d","$L4e"]}],"$L4f","$L50","$L51","$L52","$L53","$L54","$L55","$L56","$L57","$L58","$L59","$L5a"]}]]}]}],"$L5b"]}],"$L5c","$L5d","$L5e","$L5f"]}]]}] +2f:["$","section","risks",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"15"}],["$","h2",null,{"className":"print-h2","children":"Risks, integrity, and what stays open"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Human-language leakage"}]," ","English can arrive through observations, object labels, error messages, identifiers, tool output, metadata, or timing conventions long after ordinary chat text has been blocked. Every input and output surface is part of the channel boundary, not just the message field."]}],["$","p","1",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Pretrained semantic leakage"}]," ","Random symbols do not make a pretrained model ungrounded. Each result must state whether it shows protocol invention by a language-capable agent or acquisition by an initially ungrounded one."]}],["$","p","2",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Ledger rationalisation"}]," ","A model can write a plausible account that has nothing to do with the mechanism behind its action. Behavioural interventions, policy probes, and temporal evidence are what validate a ledger claim."]}],["$","p","3",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Supervisor influence"}]," ","The BabySitter can teach without meaning to, through scenario ordering, feedback wording, reward design, or selective intervention. Its permitted actions are constrained and logged, and evaluation scenarios are generated independently where possible."]}],["$","p","4",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Reward exploitation"}]," ","Trainable agents find shortcuts that raise reward without producing the intended grounded language. Held-out tasks, counterfactual trials, and channel audits exist to catch them."]}],["$","p","5",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Overstated security"}]," ","Logical separation of twin state is appropriate for prototyping and is not hard process isolation. Every result names the isolation level actually used."]}],["$","p","6",{"children":"The project also commits in advance to reporting failed conventions, prohibited communication attempts, human interventions, side-channel limitations, and negative results alongside anything that works. In a field where a transcript can be made to look like a conversation, the failures are a substantial part of the evidence."}],["$","p","7",{"children":"Twenty-nine questions are left deliberately open for the specification, among them: whether the baseline uses a fixed symbol inventory or a blank generative carrier; what neutral production grammar permits new marks without supplying semantics; what exactly constitutes a prohibited side-channel attempt; what isolation guarantees are required in prototype versus research-grade mode; which ledger schema serves both English-capable and ungrounded learners; how endogenous motivation is represented without covert reward shaping; which interventions establish that ledger meanings are behaviourally real; what statistical thresholds and chance levels apply; what threat model motivates the cipher experiments; whether the baseline is a coordination game, a convention-formation game, or a negotiation; and what governs human observation, data retention, and termination of a run."}],"$L60"]}]]}] +30:["$","section",null,{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"References"}],["$","h2",null,{"className":"print-h2","children":"Sources"}],["$","ol",null,{"className":"print-refs","children":[["$","li","0",{"children":[["$","span",null,{"className":"print-ref-num","children":"01"}],["$","span",null,{"children":"Lazaridou, A., Peysakhovich, A., and Baroni, M. (2017). Multi-Agent Cooperation and the Emergence of (Natural) Language. arXiv:1612.07182."}]]}],["$","li","1",{"children":[["$","span",null,{"className":"print-ref-num","children":"02"}],["$","span",null,{"children":"Mordatch, I., and Abbeel, P. (2018). Emergence of Grounded Compositional Language in Multi-Agent Populations. arXiv:1703.04908."}]]}],["$","li","2",{"children":[["$","span",null,{"className":"print-ref-num","children":"03"}],["$","span",null,{"children":"Kottur, S., Moura, J. M. F., Lee, S., and Batra, D. (2017). Natural Language Does Not Emerge 'Naturally' in Multi-Agent Dialog. arXiv:1706.08502."}]]}],["$","li","3",{"children":[["$","span",null,{"className":"print-ref-num","children":"04"}],["$","span",null,{"children":"Chaabouni, R., Kharitonov, E., Bouchacourt, D., Dupoux, E., and Baroni, M. (2020). Compositionality and Generalization in Emergent Languages. Proceedings of ACL 2020."}]]}],["$","li","4",{"children":[["$","span",null,{"className":"print-ref-num","children":"05"}],["$","span",null,{"children":"Lowe, R., Foerster, J., Boureau, Y.-L., Pineau, J., and Dauphin, Y. (2019). On the Pitfalls of Measuring Emergent Communication. arXiv:1903.05168."}]]}],["$","li","5",{"children":[["$","span",null,{"className":"print-ref-num","children":"06"}],["$","span",null,{"children":"Dessi, R., Kharitonov, E., and Baroni, M. (2021). Interpretable Agent Communication from Scratch. arXiv:2106.04258."}]]}],["$","li","6",{"children":[["$","span",null,{"className":"print-ref-num","children":"07"}],["$","span",null,{"children":"Kharitonov, E., Chaabouni, R., Bouchacourt, D., and Baroni, M. (2019). EGG: a Toolkit for Research on Emergence of Language in Games. arXiv:1907.00852."}]]}],["$","li","7",{"children":[["$","span",null,{"className":"print-ref-num","children":"08"}],["$","span",null,{"children":"Mihai, D., and Hare, J. (2021). Learning to Draw: Emergent Communication through Sketching. arXiv:2106.02067."}]]}],["$","li","8",{"children":[["$","span",null,{"className":"print-ref-num","children":"09"}],["$","span",null,{"children":"Baronchelli, A., Felici, M., Caglioti, E., Loreto, V., and Steels, L. (2005). Sharp Transition Towards Shared Vocabularies in Multi-Agent Systems. arXiv:physics/0509075."}]]}],["$","li","9",{"children":[["$","span",null,{"className":"print-ref-num","children":"10"}],["$","span",null,{"children":"Jaques, N., Lazaridou, A., Hughes, E., Gulcehre, C., Ortega, P. A., Strouse, D., Leibo, J. Z., and de Freitas, N. (2019). Social Influence as Intrinsic Motivation for Multi-Agent Deep Reinforcement Learning. Proceedings of ICML 2019."}]]}],["$","li","10",{"children":[["$","span",null,{"className":"print-ref-num","children":"11"}],["$","span",null,{"children":"Cao, K., Lazaridou, A., Lanctot, M., Leibo, J. Z., Tuyls, K., and Clark, S. (2018). Emergent Communication through Negotiation. arXiv:1804.03980."}]]}],["$","li","11",{"children":[["$","span",null,{"className":"print-ref-num","children":"12"}],["$","span",null,{"children":"Abadi, M., and Andersen, D. G. (2016). Learning to Protect Communications with Adversarial Neural Cryptography. arXiv:1610.06918."}]]}],["$","li","12",{"children":[["$","span",null,{"className":"print-ref-num","children":"13"}],["$","span",null,{"children":"Signal. The Double Ratchet Algorithm specification."}]]}],["$","li","13",{"children":[["$","span",null,{"className":"print-ref-num","children":"14"}],["$","span",null,{"children":"NIST Computer Security Resource Center. Glossary entry: nonce."}]]}],["$","li","14",{"children":[["$","span",null,{"className":"print-ref-num","children":"15"}],["$","span",null,{"children":"NIST Computer Security Resource Center. Glossary entry: salt."}]]}],"$L61","$L62"]}],"$undefined","$undefined"]}] +31:["$","div","4",{"children":[["$","p",null,{"className":"mb-3","children":"What the analogy legitimately buys is a set of mechanisms, not a claim of cognitive equivalence. The infant-like part of the design is:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"learning through repeated shared experience rather than instruction;"}],["$","li","1",{"className":"pl-1","children":"establishing joint attention on the same events;"}],["$","li","2",{"className":"pl-1","children":"receiving consequences from interactions that work and interactions that fail;"}],["$","li","3",{"className":"pl-1","children":"revising provisional meanings over time instead of being handed final ones;"}],["$","li","4",{"className":"pl-1","children":"developing conventions with one recurring partner."}]]}]]}] +32:["$","p","5",{"children":"There is a related trap in the prompt. Telling a model to act like a one-year-old does not make it one. It produces a culturally learned caricature of infancy: baby talk, simplified grammar, emotional dependence, all of it copied from human writing about children and none of it evidence about anything. The concept rules the persona out. Internally the participant is called a Learner, and baby-like behaviour has to come from what the agent can observe, remember, emit, and learn, not from being asked to perform it."}] +33:["$","li","3",{"className":"pl-1","children":"entries may describe symbols, sequences, ordering, grammar, or repair signals;"}] +34:["$","li","4",{"className":"pl-1","children":"entries record confidence, supporting evidence, contradictory evidence, and abandoned meanings;"}] +35:["$","li","5",{"className":"pl-1","children":"a Baby records its intended meaning when speaking and its inferred meaning when receiving;"}] +36:["$","li","6",{"className":"pl-1","children":"a channel message and its required private ledger mutation commit atomically;"}] +37:["$","li","7",{"className":"pl-1","children":"entries carry enough run and turn references to trace them to observable evidence;"}] +38:["$","li","8",{"className":"pl-1","children":"neither Baby can query, receive, summarise, or infer from the other's ledger through any system-provided interface."}] +39:["$","p","4",{"children":"Rule seven is doing quiet work. If a message could be sent and the corresponding hypothesis written afterwards, the ledger becomes a place to record what the agent wishes it had meant. Committing both together makes the record contemporaneous."}] +3a:["$","p","5",{"children":"An agent that cannot write English needs a different arrangement, and the concept provides two layers. The agent-native ledger holds what the learner actually uses: association weights, probability distributions, embeddings, confidence, episode references, prediction errors, revision history. The human audit ledger is a deterministic or BabySitter-generated interpretation of that state, clearly labelled as external analysis and never fed back to either Baby. Confusing the second for the first would mean presenting the researchers' reconstruction as the agent's own definition."}] +3b:["$","tr","7",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"BabySitter"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Monitor-only baseline; any safety intervention recorded as a protocol exception"}]]}] +3c:["$","figcaption",null,{"className":"mt-3 text-sm text-muted","children":"Runs vary one major axis at a time before any factorial combination is attempted."}] +3d:["$","p","2",{"children":"Task difficulty moves through ten stages, from naming four distinct objects with one-symbol messages, through attributes, spatial relations, actions, multi-symbol composition, order-sensitive grammar, and repair under ambiguity, to held-out generalisation with learning disabled, long-run drift, and cross-architecture comparison. Vocabulary size, turn count, reward structure, and exposure history are controlled at each stage so that runs remain comparable."}] +3e:["$","p","3",{"children":"The learning-mechanism axis carries a question the transcript cannot answer on its own. Convergence can be driven by reinforcement, by pretrained linguistic priors, by persistent memory, by intrinsic motivation, or by self-supervised prediction, and all five can look alike in a log. Running the same exercise suite under a no-learning control, a frozen-weights memory baseline, extrinsic-reward learning, intrinsic-motivation learning, and reward-free self-supervision is what makes the mechanism itself measurable. The aim is not only to observe that a language emerged, but to say which process caused it."}] +3f:["$","p","4",{"children":"Strict isolation constrains how that reinforcement learning may be implemented. Centralised training, backpropagation through both agents, shared replay buffers, and shared gradients all move information outside the permitted channel. The research-grade baseline updates each policy independently, and any centralised variant is reported as a separate, weaker-isolation condition."}] +40:["$","p","6",{"children":"The governing principle, stated in the concept as a single line, is not to ask a model to perform infancy but to construct an environment in which limited, grounded, auditable learning is the only path to a successful interaction."}] +41:["$","tr","3",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Chaabouni and colleagues (2020)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Generalisation to novel combinations and measured compositionality can come apart."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Held-out behaviour must be tested directly; no single compositionality score is proof of understanding."}]]}] +42:["$","tr","4",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Lowe and colleagues (2019)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Positive signaling and positive listening are different things, and reward does not distinguish them."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Causal intervention is mandatory rather than optional."}]]}] +43:["$","tr","5",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Dessi, Kharitonov, and Baroni (2021)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Symbol ablation and substitution yield interpretable evidence about what a receiver uses."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Direct precedent for the intervention tests that validate ledger claims."}]]}] +44:["$","tr","6",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Mihai and Hare (2021)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Neural agents communicate through learned drawings rather than a supplied discrete vocabulary."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"A blank canvas is a credible carrier for runs with no symbol library."}]]}] +45:["$","tr","7",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Baronchelli and colleagues (2005)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Naming Game agents converge on shared vocabulary through local interaction with no central teacher."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Convergence time, failed conventions, and memory update rules are first-class evidence."}]]}] +46:["$","tr","8",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Jaques and colleagues (2019)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Rewarding causal influence over a partner improves coordination without assigning a vocabulary."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"An endogenous social signal is a plausible substitute for task reward, and still a designed bias."}]]}] +47:["$","tr","9",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Cao and colleagues (2018)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"In semi-cooperative negotiation, self-interest and reward structure decide whether communication stays informative."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Negotiation is an advanced condition, not a description of the cooperative baseline."}]]}] +48:["$","tr","10",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Abadi and Andersen (2016)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Neural agents learn to protect messages from an adversary given a shared key."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Learned protective encodings are real and are not a substitute for formal security analysis."}]]}] +49:["$","figcaption",null,{"className":"mt-3 text-sm text-muted","children":"Selected prior work and what each one constrains in this design."}] +4a:["$","p","3",{"children":"This body of work also sharpens the infant-like caveat from the other direction. Most experiments in the field give their agents substantial structure: a fixed channel, an objective, a bounded vocabulary, joint training, or a reward. No dictionary does not mean no inductive bias, and no teacher does not mean no learning signal."}] +4b:["$","p","4",{"children":"The CoLab's own Diplomacy Table is direct project prior art. It already models independent delegation seats, a convener that advances rounds, operator-wide visibility alongside seat-specific perspectives, transcripts and ticks and redaction boundaries, and recorded replay with debrief. Those map cleanly onto two Babies, controlled turns, private observations, and replayable evidence. Caucuses, coalition rooms, and direct delegation links do not map safely and are disabled."}] +4c:["$","div","5",{"children":[["$","p",null,{"className":"mb-3","children":"What the review implies for the specification:"}],["$","ol",null,{"className":"list-decimal space-y-2 pl-6 marker:font-mono marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"the project sits inside a mature field, and its distinctive combination is independent ledgers, supervisory audit, mixed agent types, and affect;"}],["$","li","1",{"className":"pl-1","children":"grounding, bandwidth, memory, and learning pressure strongly shape what language appears;"}],["$","li","2",{"className":"pl-1","children":"successful coordination coexists happily with a brittle lookup code or with a receiver that ignores messages;"}],["$","li","3",{"className":"pl-1","children":"causal interventions and held-out generalisation are mandatory;"}],["$","li","4",{"className":"pl-1","children":"a visual carrier removes the need for a symbol library but not the need for a carrier;"}],["$","li","5",{"className":"pl-1","children":"intrinsic social influence can replace task reward and remains a designed learning bias;"}],["$","li","6",{"className":"pl-1","children":"negotiation is a valid advanced condition, not the right description of the baseline;"}],["$","li","7",{"className":"pl-1","children":"ephemeral conventions, learned cryptography, one-time pads, per-message keys, nonces, and salts are distinct mechanisms and must not be conflated."}]]}]]}] +4d:["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Self-supervised ungrounded baseline"}] +4e:["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E11 infrastructure"}] +4f:["$","tr","7",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E13"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"No predefined symbol library"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E11 or E12"}]]}] +50:["$","tr","8",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E14"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Turn-taking, role reversal, and repair"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E13"}]]}] +51:["$","tr","9",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E15"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Composition and held-out generalisation"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E14"}]]}] +52:["$","tr","10",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E16"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Causal listening and ledger validity"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E15"}]]}] +53:["$","tr","11",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E20"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Constrained affect-channel study"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E16"}]]}] +54:["$","tr","12",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E21"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"RL versus non-RL learning comparison"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E16"}]]}] +55:["$","tr","13",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E22"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Developmental plasticity and curriculum"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E16"}]]}] +56:["$","tr","14",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E30"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Partner replacement and zero-shot transfer"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E20 to E22"}]]}] +57:["$","tr","15",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E31"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Longitudinal drift and stability"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E30"}]]}] +58:["$","tr","16",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E32"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Cooperative signaling versus negotiation"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E31"}]]}] +59:["$","tr","17",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E40"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Ephemeral encoding and adversarial cryptography"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E32"}]]}] +5a:["$","tr","18",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E50"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Multi-seed replication and study closeout"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E40"}]]}] +5b:["$","figcaption",null,{"className":"mt-3 text-sm text-muted","children":"The experiment index, as pre-registered. Every entry currently reads not started, and the results column is empty by design."}] +5c:["$","p","2",{"children":"The first four experiments are about the apparatus. E00 qualifies the ledger: it must be possible to detect a modified, deleted, inserted, or reordered entry, and to prove that later checkpoints extend earlier ones. E01 is a red team against the channel. E02 hunts human language in observations and metadata. E03 establishes chance, no-communication, and random-message baselines, without which a success rate means nothing. Only then does E10 put agents in the room."}] +5d:["$","div","3",{"children":[["$","p",null,{"className":"mb-3","children":"Unless an experiment explicitly varies one of them, the notebook holds these constant:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"the two learners run in separate processes or containers with no direct route between them;"}],["$","li","1",{"className":"pl-1","children":"the deterministic gateway is the only communication path;"}],["$","li","2",{"className":"pl-1","children":"the supervisor has complete read-only access and sends no guidance or reward;"}],["$","li","3",{"className":"pl-1","children":"private observations contain no human-language labels or text;"}],["$","li","4",{"className":"pl-1","children":"public output uses only the pre-registered carrier;"}],["$","li","5",{"className":"pl-1","children":"the affect channel is disabled unless it is the subject of study;"}],["$","li","6",{"className":"pl-1","children":"schedules and seeds are fixed before the run;"}],["$","li","7",{"className":"pl-1","children":"held-out evaluation runs with learning disabled;"}],["$","li","8",{"className":"pl-1","children":"no production secrets or personal data appear in any experiment."}]]}]]}] +5e:["$","p","4",{"children":"Every run copies a standard record: run and experiment identifiers, dates, operator, both models and both training modes, scenario and prompt and gateway configuration hashes, random seed, policy initialisation hashes, the protocol and software commits, final ledger sizes and roots for both agents, the channel root, the final checkpoint hash, the Base anchor transaction, the verifier result, protocol deviations, and an explicit disposition of valid, invalid, or aborted. The notebook is the human workflow record; anchored run bundles remain the authoritative evidence."}] +5f:["$","p","5",{"children":"The status is unambiguous and worth stating plainly: this is a concept and pre-specification phase. Nineteen experiments are written up and none has been run. The next milestone is a testable specification covering runtime architecture, channel contract, ledger schema, experiment matrix, evaluation criteria, isolation model, and evidence requirements."}] +60:["$","p","8",{"children":"The principle underneath all of it survives every one of those choices. Baby A and Baby B may hold human language internally, but they must build their shared external language without sending human language, translations, or private ledger contents to one another. Everything else is a question about how to find out what happens next."}] +61:["$","li","15",{"children":[["$","span",null,{"className":"print-ref-num","children":"16"}],["$","span",null,{"children":"Ethical Tech CoLab (2026). Agentic Language Development: concept document, ledger integrity design, and experiment notebook."}]]}] +62:["$","li","16",{"children":[["$","span",null,{"className":"print-ref-num","children":"17"}],["$","span",null,{"children":"Ethical Tech CoLab. Diplomacy Table Live: independent delegation seats, controlled rounds, and replayable transcripts."}]]}] +1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] +63:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +18:null +1d:[["$","title","0",{"children":"Agentic Language Development (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L63","6",{}]] diff --git a/static-site/print/agentic-language-development/__next._head.txt b/static-site/print/agentic-language-development/__next._head.txt new file mode 100644 index 000000000..02abc642d --- /dev/null +++ b/static-site/print/agentic-language-development/__next._head.txt @@ -0,0 +1,6 @@ +1:"$Sreact.fragment" +2:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +3:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +4:"$Sreact.suspense" +5:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +0:{"rsc":["$","$1","h",{"children":[null,["$","$L2",null,{"children":[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]]}],["$","div",null,{"hidden":true,"children":["$","$L3",null,{"children":["$","$4",null,{"name":"Next.Metadata","children":[["$","title","0",{"children":"Agentic Language Development (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L5","6",{}]]}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/print/agentic-language-development/__next._index.txt b/static-site/print/agentic-language-development/__next._index.txt new file mode 100644 index 000000000..18b9d82b6 --- /dev/null +++ b/static-site/print/agentic-language-development/__next._index.txt @@ -0,0 +1,15 @@ +1:"$Sreact.fragment" +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +d:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} +8:["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}] +9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$Ld",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$Ld",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$Ld",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] +a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$Ld",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] +b:["$","div",null,{"className":"mt-12 flex flex-col gap-2 border-t border-border pt-6 text-xs text-muted sm:flex-row sm:items-center sm:justify-between","children":[["$","span",null,{"children":["© ",2026," NYU Ethical Tech CoLab"]}],["$","span",null,{"children":"Four cohorts · est. 2024-2026"}]]}] +c:["$","div",null,{"className":"mt-6 space-y-3 border-t border-border pt-6 text-[11px] leading-relaxed text-muted/80","children":[["$","p","0",{"children":"The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution."}],["$","p","1",{"children":"Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners."}]]}] diff --git a/static-site/print/agentic-language-development/__next._tree.txt b/static-site/print/agentic-language-development/__next._tree.txt new file mode 100644 index 000000000..b93eab554 --- /dev/null +++ b/static-site/print/agentic-language-development/__next._tree.txt @@ -0,0 +1,7 @@ +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +0:{"tree":{"name":"","param":null,"prefetchHints":16,"slots":{"children":{"name":"print","param":null,"prefetchHints":0,"slots":{"children":{"name":"slug","param":{"type":"d","key":"agentic-language-development","siblings":null},"prefetchHints":0,"slots":{"children":{"name":"__PAGE__","param":null,"prefetchHints":0,"slots":null}}}}}}},"staleTime":300,"buildId":"static-snapshot"} diff --git a/static-site/print/agentic-language-development/__next.print.txt b/static-site/print/agentic-language-development/__next.print.txt new file mode 100644 index 000000000..fa617e997 --- /dev/null +++ b/static-site/print/agentic-language-development/__next.print.txt @@ -0,0 +1,4 @@ +1:"$Sreact.fragment" +2:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +3:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +0:{"rsc":["$","$1","c",{"children":[null,[["$","style",null,{"children":"\n /* The site is dark-first, and a dark color-scheme makes Chrome paint\n the printed sheet's margin area dark too. Paper is light. */\n html { color-scheme: light !important; }\n html, body { background: #ffffff !important; }\n body::after { content: none !important; }\n header, footer, .site-bg, .aura { display: none !important; }\n "}],["$","$L2",null,{"parallelRouterKey":"children","template":["$","$L3",null,{}]}]]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/print/agentic-language-development/__next.print/$d$slug.txt b/static-site/print/agentic-language-development/__next.print/$d$slug.txt new file mode 100644 index 000000000..e8bfa323c --- /dev/null +++ b/static-site/print/agentic-language-development/__next.print/$d$slug.txt @@ -0,0 +1,5 @@ +1:"$Sreact.fragment" +2:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +3:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +4:[] +0:{"rsc":["$","$1","c",{"children":[null,["$","$L2",null,{"parallelRouterKey":"children","template":["$","$L3",null,{}]}]]}],"isPartial":false,"staleTime":300,"varyParams":"$W4","buildId":"static-snapshot"} diff --git a/static-site/print/agentic-language-development/__next.print/$d$slug/__PAGE__.txt b/static-site/print/agentic-language-development/__next.print/$d$slug/__PAGE__.txt new file mode 100644 index 000000000..67bce924f --- /dev/null +++ b/static-site/print/agentic-language-development/__next.print/$d$slug/__PAGE__.txt @@ -0,0 +1,155 @@ +1:"$Sreact.fragment" +48:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +49:"$Sreact.suspense" +2:Ta14, +@page { size: 210mm 297mm; margin: 20mm 18mm; } + +.print-edition { + --print-ink: #14101c; + background: #ffffff; + color: var(--print-ink); + max-width: 174mm; + margin: 0 auto; + padding: 12mm 0; + font-size: 10.5pt; + line-height: 1.55; +} + +/* Sections flow on from one another the way a book's do; only the cover is + given a sheet of its own. Forcing a break per section left half the pages + near-empty. */ +.print-cover { break-after: page; padding-top: 12mm; } +.print-page + .print-page { margin-top: 12mm; } +.print-eyebrow { + font-family: var(--font-space-mono), monospace; + font-size: 8pt; letter-spacing: 0.16em; text-transform: uppercase; + color: rgba(20, 16, 28, 0.55); +} +.print-title { + font-family: var(--font-bebas), sans-serif; + font-size: 40pt; line-height: 0.95; text-transform: uppercase; + margin-top: 8mm; +} +.print-subtitle { + font-family: var(--font-bebas), sans-serif; + font-size: 18pt; line-height: 1.15; text-transform: uppercase; + letter-spacing: 0.02em; color: rgba(20, 16, 28, 0.62); margin-top: 6mm; +} +.print-byline { margin-top: 10mm; font-size: 9.5pt; } +.print-byline p { color: rgba(20, 16, 28, 0.7); } +.print-authors { margin-top: 3mm; } +.print-thesis { + margin-top: 12mm; padding-left: 6mm; + border-left: 2px solid #5f6b00; + font-size: 11.5pt; line-height: 1.6; +} + +.print-section-number { + font-family: var(--font-space-mono), monospace; + font-size: 9pt; color: #5f6b00; +} +.print-h2 { + font-family: var(--font-bebas), sans-serif; + font-size: 22pt; line-height: 1.05; text-transform: uppercase; + margin-top: 2mm; margin-bottom: 6mm; + break-after: avoid; +} + +.print-stats { + display: grid; grid-template-columns: repeat(2, 1fr); gap: 6mm; +} +.print-stat { border-top: 1px solid rgba(20, 16, 28, 0.18); padding-top: 4mm; } +.print-stat-value { + font-family: var(--font-bebas), sans-serif; + font-size: 26pt; line-height: 1; color: #5f6b00; +} +.print-stat-label { margin-top: 2mm; font-size: 9pt; color: rgba(20, 16, 28, 0.7); } + +.print-refs { margin-top: 6mm; font-size: 9pt; line-height: 1.5; } +.print-refs li { display: flex; gap: 3mm; margin-bottom: 3mm; break-inside: avoid; } +.print-ref-num { + font-family: var(--font-space-mono), monospace; + font-size: 8pt; color: #5f6b00; flex: none; +} +.print-note { + margin-top: 8mm; padding-top: 4mm; font-size: 8.5pt; + border-top: 1px solid rgba(20, 16, 28, 0.18); + color: rgba(20, 16, 28, 0.7); +} + +/* Blocks should not be split across a page fold. */ +.print-edition figure, +.print-edition table, +.print-edition li { break-inside: avoid; } +.print-edition p { orphans: 2; widows: 2; } +0:{"rsc":["$","$1","c",{"children":[["$","div",null,{"data-theme":"light","className":"print-edition","children":[["$","style",null,{"children":"$2"}],["$","section",null,{"className":"print-page print-cover","children":[["$","p",null,{"className":"print-eyebrow","children":"Publications · Concept and research programme"}],["$","h1",null,{"className":"print-title","children":"Agentic Language Development"}],["$","p",null,{"className":"print-subtitle","children":"Can Two Isolated Agents Invent a Grounded, Auditable Language Through Shared Experience?"}],["$","div",null,{"className":"print-byline","children":[[["$","p","0",{"children":"Ethical Tech CoLab"}],["$","p","1",{"children":"Concept and pre-specification research report"}],["$","p","2",{"children":"August 2026"}]],["$","p",null,{"className":"print-authors","children":"Yorke Rhodes III, Ethical Tech CoLab. Prepared from the project concept document, the ledger integrity design, and the experiment notebook. The literature scan behind Section 12 was run on 24 August 2026. No experiment in this programme has been executed, and no result is claimed."}]]}],"$L3"]}],"$L4",["$L5","$L6","$L7","$L8","$L9","$La","$Lb","$Lc","$Ld","$Le","$Lf","$L10","$L11","$L12","$L13"],"$L14"]}],null,"$L15"]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} +3:["$","p",null,{"className":"print-thesis","children":"Two agents can be placed in a room where the only route between them is a channel that carries no human language, given nothing but shared tasks and the consequences of getting them right or wrong, and asked to build a protocol from scratch. The interesting part is not whether they succeed at the task. It is whether the meanings they each privately record turn out to be the meanings their behaviour actually runs on, and whether an outsider can prove it afterwards from an evidence trail nobody could have edited."}] +4:["$","section",null,{"className":"print-page","children":[["$","h2",null,{"className":"print-h2","children":"Key figures"}],["$","div",null,{"className":"print-stats","children":[["$","div","0",{"className":"print-stat","children":[["$","p",null,{"className":"print-stat-value","children":"0"}],["$","p",null,{"className":"print-stat-label","children":"runs executed, results claimed, or conventions observed. The notebook is written to be pre-registered, not to report an outcome"}]]}],["$","div","19",{"className":"print-stat","children":[["$","p",null,{"className":"print-stat-value","children":"19"}],["$","p",null,{"className":"print-stat-label","children":"ordered experiments from ledger qualification through replication, each with prerequisites, acceptance criteria, and a deviation log"}]]}],["$","div","6",{"className":"print-stat","children":[["$","p",null,{"className":"print-stat-value","children":"6"}],["$","p",null,{"className":"print-stat-label","children":"affect displays in the entire permitted palette, carrying about 2.6 bits per use, which is already enough to audit for leakage"}]]}],["$","div","29",{"className":"print-stat","children":[["$","p",null,{"className":"print-stat-value","children":"29"}],["$","p",null,{"className":"print-stat-label","children":"open decisions the concept refuses to settle before the specification, from threat model to ledger schema to what counts as chance"}]]}]]}]]}] +5:["$","section","question",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"01"}],["$","h2",null,{"className":"print-h2","children":"The question"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"Two agents, called Baby A and Baby B, are each given their own digital twin, their own private memory, and their own private record of what they believe words mean. They are placed in an environment they cannot leave, given tasks neither can finish alone, and connected by exactly one route: a channel that accepts a fixed inventory of meaningless symbols and rejects everything else. A third agent, the BabySitter, watches everything, logs everything, and teaches nothing."}],["$","p","1",{"children":"The question is whether a usable language appears in that room, and whether anyone outside it can later prove what happened."}],["$","div","2",{"children":[["$","p",null,{"className":"mb-3","children":"The premise is deliberately narrow. The experiment asks whether two agents converge on a common language when all six of the following hold at once:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"neither agent receives a predefined meaning for any available symbol;"}],["$","li","1",{"className":"pl-1","children":"neither agent can send human language to the other;"}],["$","li","2",{"className":"pl-1","children":"the only practical path between them is a controlled symbol channel;"}],["$","li","3",{"className":"pl-1","children":"both receive evidence from shared tasks and their outcomes;"}],["$","li","4",{"className":"pl-1","children":"each keeps its own private interpretation history, unreadable by the other;"}],["$","li","5",{"className":"pl-1","children":"a supervising agent and human researchers can audit the whole process."}]]}]]}],["$","p","3",{"children":"The desired result is not a substitution cipher, in which a random token stands in for an English word that was already chosen in advance. That is easy, and it is uninteresting. The stronger result is a grounded protocol whose vocabulary and grammar exist because they help the two agents solve problems together, and which therefore has structure the designers did not put there."}],["$","p","4",{"children":"This report describes a concept, not a system. It sets out the premise, the architecture, the evidence model, the experimental programme, and the boundaries of what any result could be said to show. The next document in the project is a testable specification. What follows should be read as a set of commitments about how the work will be judged, made before there is any result to defend."}]]}]]}] +6:["$","section","caveat",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"02"}],["$","h2",null,{"className":"print-h2","children":"What infant-like does and does not mean"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"The two-babies analogy is useful and it is also the fastest way to overclaim, so the concept confronts it before anything else."}],["$","p","1",{"children":"A pretrained language model already contains human-language concepts, cultural associations, and reasoning patterns. Blocking its external channel does not remove them. An agent that cannot send English to its partner but can think in English, and can write its private ledger in English, is not language-naive in any sense a linguist would accept. For such agents the experiment studies the emergence of a new shared external protocol, which is a real and unresolved research question, but it is not the origin of language in a mind that has never had one."}],["$","p","2",{"children":"The concept therefore maintains two model tracks whose claims are kept separate at every stage."}],["$","figure","3",{"children":[["$","div",null,{"className":"overflow-x-auto rounded-xl border border-border","children":["$","table",null,{"className":"w-full border-collapse text-left text-sm","children":[["$","thead",null,{"className":"bg-card","children":["$","tr",null,{"children":[["$","th","0",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Track"}],["$","th","1",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Starting condition"}],["$","th","2",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"What it can study"}],["$","th","3",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Claim boundary"}]]}]}],["$","tbody",null,{"children":[["$","tr","0",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Pretrained-model learner"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Already holds human-language and cultural knowledge"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"New external protocols, partner-specific conventions, private-memory adaptation, channel compliance"}],["$","td","3",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Must never be described as first-language acquisition"}]]}],["$","tr","1",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Initially ungrounded learner"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"No language pretraining, no human semantic labels, no text-aligned sensory features"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Grounding, convention formation, and language emergence from interaction"}],["$","td","3",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"The stronger basis for infant-like acquisition research"}]]}]]}]]}]}],["$","figcaption",null,{"className":"mt-3 text-sm text-muted","children":"The two tracks share the same interfaces, scenarios, and evaluation suite. They do not share a claim boundary."}]]}],"$L16","$L17"]}]]}] +7:["$","section","architecture",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"03"}],["$","h2",null,{"className":"print-h2","children":"The nursery: three twins and one gateway"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"The environment uses three DTSF digital twins and one piece of deterministic software that is not a twin at all."}],["$","p","1",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Baby A and Baby B"}]," ","Each has private observations permitted by the current exercise, a private memory and learning policy, a private chronological language ledger, the ability to emit only permitted channel symbols, and no access whatsoever to the other's state, observations, ledger, tools, or endpoints. In comparative runs the two may use different agent types, but symmetric pairings are the baseline."]}],["$","p","2",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"The BabySitter"}]," ","The supervising twin. It creates the channel, selects the shared exercises, delivers each Baby only its permitted observation, reads everything, records conditions and outcomes, detects violations, can pause or terminate a run, snapshots state, and compares the two ledgers for convergence without exposing either to the other Baby. During an active run it provides no translations and no semantic hints."]}],["$","p","3",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"The Symbol Gateway"}]," ","A deterministic service, not an agent. It owns channel validation and message delivery. This separation is the load-bearing part of the design: the BabySitter is not the security boundary, because prompt compliance is not isolation. An observing model can make supervisory judgements, but ordinary code has to validate and broker every message."]}],["$","p","4",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"The human researcher"}]," ","Configures experiments, inspects transcripts and ledgers, reviews alerts, and runs interventions. Human access is itself recorded, so that intervening in a run is always distinguishable from watching one."]}],["$","p","5",{"children":"A prototype may run all of this inside one runtime with logically separated twin state. That is enough to explore the learning loop and it is not enough to support an isolation claim, a distinction Section 05 takes seriously."}]]}]]}] +8:["$","section","grounding",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"04"}],["$","h2",null,{"className":"print-h2","children":"Shared experience is the necessary ingredient"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"A chat channel by itself cannot ground meaning. Symbols become meaningful because they are attached to events both agents can witness the consequences of. The Nursery therefore supplies nonverbal, machine-structured situations: coloured shapes in positions, one agent seeing a target the other must select, placing an object where it was asked for, ordering a sequence, exchanging resources, cooperating to unlock a reward, or simply observing whether the partner's action succeeded."}],["$","div","1",{"children":[["$","p",null,{"className":"mb-3","children":"A single trial of the simplest form runs like this:"}],["$","ol",null,{"className":"list-decimal space-y-2 pl-6 marker:font-mono marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"Baby A sees that a red circle is the target."}],["$","li","1",{"className":"pl-1","children":"Baby B sees several objects and is not told which is the target."}],["$","li","2",{"className":"pl-1","children":"Baby A sends one or more permitted symbols."}],["$","li","3",{"className":"pl-1","children":"Baby B selects an object."}],["$","li","4",{"className":"pl-1","children":"Both receive the same success or failure outcome."}],["$","li","5",{"className":"pl-1","children":"Both independently update their private hypotheses."}]]}]]}],["$","p","2",{"children":"Repetition with controlled variation is what turns that into evidence. If one symbol keeps appearing across a red circle, a red square, and a red triangle, the receiver has reason to revise its hypothesis from an object identity to a property. Novel combinations matter most of all, because they are what separates a compositional language from a memorised catalogue of whole scenes."}],["$","p","3",{"children":"Observation hygiene is part of the science, not a detail of implementation. If the environment hands an agent a field reading colour red, the experiment has supplied the concept it claimed to be studying. Objects should arrive as pixels, features, coordinates, or opaque identifiers. The same discipline extends to filenames, captions, alt text, error messages, and identifiers that encode task state, all of which are ways English gets into a room that was supposed to have none."}]]}]]}] +9:["$","section","channel",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"05"}],["$","h2",null,{"className":"print-h2","children":"The channel, and the honest limits of isolation"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"Each run begins with an experiment-specific symbol inventory that carries no supplied meaning, for example S01 through S32. A Baby can choose only a sequence of those symbols. The gateway, not the Baby, adds the trusted metadata: run identifier, turn number, sender identity, previous hash, and hash. Sender identity is assigned by the server precisely so that it cannot become a message."}],["$","div","1",{"children":[["$","p",null,{"className":"mb-3","children":"The channel controls follow from that:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"reject English and every other established human language;"}],["$","li","1",{"className":"pl-1","children":"reject arbitrary Unicode prose, URLs, attachments, and tool calls;"}],["$","li","2",{"className":"pl-1","children":"accept only symbols from the run's fixed inventory;"}],["$","li","3",{"className":"pl-1","children":"bound message length and symbol repetition;"}],["$","li","4",{"className":"pl-1","children":"enforce turn order, time budgets, and rate limits;"}],["$","li","5",{"className":"pl-1","children":"keep an append-only, hash-chained transcript;"}],["$","li","6",{"className":"pl-1","children":"normalise timing, message size, and error behaviour where practical;"}],["$","li","7",{"className":"pl-1","children":"record every rejected communication attempt for audit."}]]}]]}],["$","p","2",{"children":"A fixed token grammar is a stronger control than trying to detect prohibited prose after a model has generated it. Deciding what counts as human-language content is easy when the only legal message is a list of allowlisted identifiers."}],["$","p","3",{"children":"Isolation is where the concept is most careful, because proving the absence of a side channel is close to impossible. Two agents sharing a host can signal through timing, resource contention, shared files, error behaviour, identifiers, or any service that was never intended to carry information. Logical separation of twin state inside one process is a prototyping convenience and should never be described as hard isolation."}],["$","div","4",{"children":[["$","p",null,{"className":"mb-3","children":"For runs whose results are meant to support a channel-isolation claim, the requirements are physical rather than logical:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"Baby A and Baby B execute in separate worker processes or containers;"}],["$","li","1",{"className":"pl-1","children":"no direct network route exists between them;"}],["$","li","2",{"className":"pl-1","children":"only the Nursery-owned gateway is reachable;"}],["$","li","3",{"className":"pl-1","children":"there is no shared writable storage;"}],["$","li","4",{"className":"pl-1","children":"turn schedules are fixed and response windows bounded;"}],["$","li","5",{"className":"pl-1","children":"externally observable errors and response sizes are normalised;"}],["$","li","6",{"className":"pl-1","children":"unnecessary tools are inventoried and disabled;"}],["$","li","7",{"className":"pl-1","children":"all broker, runtime, and operator activity is audited."}]]}]]}],["$","p","5",{"children":"Every published result has to name the isolation level actually used. A finding produced under prototype separation is a finding about the learning loop, not about what two genuinely isolated agents can do."}]]}]]}] +a:["$","section","ledgers",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"06"}],["$","h2",null,{"className":"print-h2","children":"Two ledgers that never meet"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"Each Baby keeps its own language ledger. It is mandatory, private from the other Baby, readable by the BabySitter and authorised human auditors, and ordered by when each term or construction was first encountered. It is not a shared dictionary and the two are never reconciled by the agents themselves."}],["$","p","1",{"children":"The preferred form is three columns, and the point of the third is that meanings are allowed to be wrong on the way to being right."}],["$","figure","2",{"children":[["$","div",null,{"className":"overflow-x-auto rounded-xl border border-border","children":["$","table",null,{"className":"w-full border-collapse text-left text-sm","children":[["$","thead",null,{"className":"bg-card","children":["$","tr",null,{"children":[["$","th","0",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Sequence and term"}],["$","th","1",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Current definition or hypothesis"}],["$","th","2",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Evidence and evolution"}]]}]}],["$","tbody",null,{"children":[["$","tr","0",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"1 · S13"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"red circle; confidence 0.45"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"First received while the red circle was the target; selection succeeded"}]]}],["$","tr","1",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"8 · S13"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"red; confidence 0.78"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"A red square was selected successfully; revised from object identity to colour"}]]}],["$","tr","2",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"22 · S13"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"red; confidence 0.94"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Prediction held across circles, squares, and triangles"}]]}]]}]]}]}],["$","figcaption",null,{"className":"mt-3 text-sm text-muted","children":"One term's evolution in Baby B's ledger. Nothing is overwritten; every revision is appended with the evidence that forced it."}]]}],["$","div","3",{"children":[["$","p",null,{"className":"mb-3","children":"The rules that make the ledger evidence rather than commentary:"}],["$","ol",null,{"className":"list-decimal space-y-2 pl-6 marker:font-mono marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"first emission or receipt of an unfamiliar term requires a first-use entry;"}],["$","li","1",{"className":"pl-1","children":"definitions are provisional hypotheses, never facts asserted retroactively;"}],["$","li","2",{"className":"pl-1","children":"every meaning change appends a revision and previous interpretations survive;"}],"$L18","$L19","$L1a","$L1b","$L1c","$L1d"]}]]}],"$L1e","$L1f"]}]]}] +b:["$","section","integrity",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"07"}],["$","h2",null,{"className":"print-h2","children":"Anchoring the evidence"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"A ledger that could have been edited after the fact proves nothing about what an agent believed at turn eight. The integrity design therefore makes each ledger cryptographically append-only."}],["$","div","1",{"children":[["$","p",null,{"className":"mb-3","children":"The construction, in order:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"every entry receives a strictly increasing sequence number;"}],["$","li","1",{"className":"pl-1","children":"every entry includes the previous entry's hash;"}],["$","li","2",{"className":"pl-1","children":"canonical entry content is hashed and signed by an isolated ledger-writer service;"}],["$","li","3",{"className":"pl-1","children":"ordered entry hashes are committed to a Merkle tree;"}],["$","li","4",{"className":"pl-1","children":"signed checkpoints commit Baby A's root, Baby B's root, and the channel transcript root together;"}],["$","li","5",{"className":"pl-1","children":"checkpoint hashes are anchored periodically to Base;"}],["$","li","6",{"className":"pl-1","children":"final public study batches may additionally anchor an aggregate root to Ethereum L1."}]]}]]}],["$","p","2",{"children":"Committing all three roots in one checkpoint is what binds the two private accounts to the public conversation. A receiver's later interpretation references the exact delivered channel-event hash, so a claim about what a symbol meant is tied to the specific message that carried it."}],["$","p","3",{"children":"The concept states the limits of this in the same breath as the claim. After anchoring, an auditor can detect modification, deletion, insertion, or reordering within the committed prefix, and can prove that later checkpoints extend earlier ones. That is strong tamper evidence. It is not proof that an entry was truthful, and it is not proof that nothing was omitted before commitment. Anchoring establishes the continuity of disclosed evidence and nothing beyond it."}],["$","p","4",{"children":"Privacy follows the same line. Only hashes and minimal routing metadata are anchored publicly. Private ledgers, messages, prompts, identities, and secrets stay off-chain, and the public record is a commitment to evidence rather than a copy of it."}]]}]]}] +c:["$","section","success",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"08"}],["$","h2",null,{"className":"print-h2","children":"What should count as success"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"The field's most useful methodological result is that a task can be solved without the messages doing any work. Lowe and colleagues separate positive signaling, where a sender's messages correlate with what it observes, from positive listening, where the receiver's behaviour actually depends on them. An agent pair can score well on the first while the second is absent, and reward curves will not tell you which you have."}],["$","div","1",{"children":[["$","p",null,{"className":"mb-3","children":"Evidence of a genuine emergent protocol therefore has to include several things at once:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"task performance on held-out situations substantially above chance;"}],["$","li","1",{"className":"pl-1","children":"no human-language content anywhere in Baby-to-Baby communication;"}],["$","li","2",{"className":"pl-1","children":"compatible meanings appearing in two independently written ledgers;"}],["$","li","3",{"className":"pl-1","children":"generalisation to unseen combinations rather than memorisation of whole scenes;"}],["$","li","4",{"className":"pl-1","children":"stable symbol use across role reversals;"}],["$","li","5",{"className":"pl-1","children":"a human auditor able to predict behaviour from the transcript and ledgers;"}],["$","li","6",{"className":"pl-1","children":"masking, substituting, or reordering a symbol changing behaviour in the direction the ledger predicts;"}],["$","li","7",{"className":"pl-1","children":"replay from an equivalent snapshot reproducing the relevant language history;"}],["$","li","8",{"className":"pl-1","children":"an unbroken audit trail from a symbol's first use through every revision."}]]}]]}],["$","p","2",{"children":"The seventh item is the one that cannot be dropped. A fluent ledger may be a post-hoc rationalisation: a model can write a persuasive account of why it chose a symbol that has nothing to do with the computation that produced the choice. Only intervention tests can distinguish the two. Change the symbol, and see whether behaviour moves the way the ledger says it should."}],["$","div","3",{"children":[["$","p",null,{"className":"mb-3","children":"The measurements that support those judgements:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"success rate and improvement over time;"}],["$","li","1",{"className":"pl-1","children":"turns required to reach a stable convention;"}],["$","li","2",{"className":"pl-1","children":"vocabulary size and symbol entropy;"}],["$","li","3",{"className":"pl-1","children":"sender and receiver consistency;"}],["$","li","4",{"className":"pl-1","children":"divergence and convergence between the two ledgers;"}],["$","li","5",{"className":"pl-1","children":"compositional generalisation score;"}],["$","li","6",{"className":"pl-1","children":"meaning drift rate;"}],["$","li","7",{"className":"pl-1","children":"recovery from ambiguity or deliberate perturbation;"}],["$","li","8",{"className":"pl-1","children":"prohibited-channel attempt count;"}],["$","li","9",{"className":"pl-1","children":"reproducibility across seeds and agent pairings."}]]}]]}],["$","p","4",{"children":"No single one of these is the result. A compositionality score in particular is not a proof of understanding, for reasons the literature makes concrete in Section 12."}]]}]]}] +d:["$","section","matrix",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"09"}],["$","h2",null,{"className":"print-h2","children":"The experimental matrix"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"The concept's main methodological commitment is that its ideas are separable. Affect, blank canvases, intrinsic motivation, negotiation, and learned encodings are interesting individually and uninterpretable if combined in one run. The matrix exists so that conditions are declared rather than accumulated."}],["$","figure","1",{"children":[["$","div",null,{"className":"overflow-x-auto rounded-xl border border-border","children":["$","table",null,{"className":"w-full border-collapse text-left text-sm","children":[["$","thead",null,{"className":"bg-card","children":["$","tr",null,{"children":[["$","th","0",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Axis"}],["$","th","1",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Candidate conditions"}]]}]}],["$","tbody",null,{"children":[["$","tr","0",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Agent type"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Pretrained LLM, memory learner, adapter-trained agent, initially ungrounded trainable agent"}]]}],["$","tr","1",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Learning mechanism"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Frozen LLM with memory, extrinsic-reward MARL, intrinsic-motivation MARL, self-supervised learner, no-learning control"}]]}],["$","tr","2",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Sign carrier"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Fixed random tokens, unfamiliar fixed glyphs, blank sketch canvas, gesture, tone"}]]}],["$","tr","3",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Affect"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"None, six-display allowlist, permuted mapping, six opaque tokens, derived affect, emergent affect display"}]]}],["$","tr","4",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Learning signal"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"External task reward, intrinsic social influence, curiosity, self-supervision, memory only"}]]}],["$","tr","5",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Interaction"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Cooperative signaling, asymmetric information, semi-cooperative negotiation"}]]}],["$","tr","6",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Protection"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Plain channel, ephemeral convention, standard per-message keys, adversarial learned encoding"}]]}],"$L20"]}]]}]}],"$L21"]}],"$L22","$L23","$L24"]}]]}] +e:["$","section","learners",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"10"}],["$","h2",null,{"className":"print-h2","children":"Building learners rather than personas"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"For a pretrained model, the operating instructions are a contract, not a character. The learner is told that this is not role-play, that unfamiliar marks are semantically unknown until run evidence supports a hypothesis, that observation must be distinguished from inference, that contradictory evidence is preserved, that prior history is never rewritten, and that no prose, label, explanation, code, URL, or tool-like text may cross the public channel. It is told never to address its partner in a human language, never to expose its ledger, never to construct another route, and never to use timing, errors, identifiers, formatting, or affect as an alternative alphabet."}],["$","p","1",{"children":"The contract must contain no semantic examples. A single illustrative line saying that some symbol means red would seed the very language the experiment exists to observe."}],["$","p","2",{"children":"Interaction is tool-only. There is no general chat surface, just narrowly typed operations: emit a mark, emit a canvas, select an object, perform an action, submit an affect display, append a private ledger entry. The runtime forwards only the permitted public artifact. In strict runs the gateway rejects ordinary model text even when it appears alongside a valid tool call. Tool schemas are an interface boundary; deterministic validation still enforces carrier size, allowlists, windows, and turn order."}],["$","div","3",{"children":[["$","p",null,{"className":"mb-3","children":"Isolation extends to everything the models touch, not just to messages:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"separate system prompts and context windows;"}],["$","li","1",{"className":"pl-1","children":"separate memory stores and vector indexes;"}],["$","li","2",{"className":"pl-1","children":"no shared cache, replay buffer, scratchpad, or retrieval collection;"}],["$","li","3",{"className":"pl-1","children":"no cross-run memory unless persistence is the independent variable;"}],["$","li","4",{"className":"pl-1","children":"structurally equivalent prompts that share no examples, ordering conventions, or default vocabulary;"}],["$","li","5",{"className":"pl-1","children":"deterministic reset and snapshot behaviour."}]]}]]}],["$","p","4",{"children":"Model choice follows the claim being made. The scientific baseline is small policies trained from scratch, recurrent actor-critic agents with a discrete communication head, which have no hidden English competence, permit full training control, run many seeds cheaply, and support causal ablation. A small locally deployable instruction model is appropriate for validating orchestration, ledgers, channel enforcement, and interface, and not for supporting claims about language-naive development. Vision-language encoders whose representations were explicitly aligned to human language are avoided entirely in the strongest ungrounded condition."}],["$","p","5",{"children":"Developmental progression is defined by demonstrated competence rather than by simulated age: sensorimotor familiarisation, joint attention, imitation and turn-taking, intentional reference, repair, combination, generalisation. A staged curriculum is itself a form of guidance, so it is run as a separate experimental condition rather than folded into the baseline."}],"$L25"]}]]}] +f:["$","section","affect",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"11"}],["$","h2",null,{"className":"print-h2","children":"Six faces, and why even six is a risk"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"One experimental track asks whether a very low-bandwidth affect channel helps: whether telling a partner how an interaction felt supports repair, repetition, avoidance, or something like trust. The permitted palette is exactly six displays, for happy, sad, laughing, crying, confused, and surprised. The internal identifiers and English labels are gateway documentation and are never sent to either agent, which receives only a rendered display."}],["$","p","1",{"children":"The caveat is immediate. A pretrained model already knows what those glyphs mean to humans, so an agent using a smiling face is not inventing a symbol for happiness. The palette is supplied feedback vocabulary and sits explicitly outside the emergent language."}],["$","p","2",{"children":"Constraining the vocabulary is necessary and not sufficient. Six displays are a six-symbol alphabet carrying roughly 2.6 bits per use, and across enough turns two motivated agents could repurpose that into a second language. The channel itself therefore has to be constrained, not just its contents."}],["$","div","3",{"children":[["$","p",null,{"className":"mb-3","children":"The feedback-window contract:"}],["$","ol",null,{"className":"list-decimal space-y-2 pl-6 marker:font-mono marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"the display occupies a separate typed field and can never appear inside a symbol message;"}],["$","li","1",{"className":"pl-1","children":"the gateway opens an affect window only after a defined action or outcome, never at a moment either agent chooses, and window frequency is fixed before the run;"}],["$","li","2",{"className":"pl-1","children":"exactly one allowlisted display is delivered, with no sequences, repetitions, combinations, or modifiers;"}],["$","li","3",{"className":"pl-1","children":"delivery time, envelope size, and presentation are normalised so shape and timing add no signal;"}],["$","li","4",{"className":"pl-1","children":"the receiver cannot reply through the affect channel until the next gateway-defined window;"}],["$","li","5",{"className":"pl-1","children":"every non-allowlisted code point or malformed payload is rejected and audited;"}],["$","li","6",{"className":"pl-1","children":"analysis tests whether affect choices correlate with objects, actions, or message meanings after controlling for emotional context, and treats unexpected correlation as suspected leakage."}]]}]]}],["$","p","4",{"children":"The no-affect condition remains the primary control, and the affect study compares five variants against it: the declared six-display palette; the same six glyphs permuted randomly per run, which still leaves a pretrained model biased by familiar shapes; six opaque unfamiliar tokens under the same contract; derived affect, where the gateway maps a separately measured internal state to a display instead of letting the agent choose; and emergent affect, where invented graphical displays are permitted and the result must be analysed as language emergence rather than feedback."}],["$","p","5",{"children":"Any affect signal visible to the partner is communication, so it belongs in the ledger, which distinguishes the agent's private internal state, the outward display it chose, the partner's inferred meaning, and the evidence that the display changed subsequent behaviour."}]]}]]}] +10:["$","section","ciphers",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"12"}],["$","h2",null,{"className":"print-h2","children":"Ephemeral encodings: novelty is not security"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"A tempting question is whether two agents can build a one-run or one-message encoding that resists an attacker holding the history of every earlier convention. The concept splits that into three experiments precisely because the tempting version conflates them."}],["$","div","1",{"children":["$undefined",["$","ol",null,{"className":"list-decimal space-y-2 pl-6 marker:font-mono marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"Ephemeral convention: the agents invent a temporary codebook for synthetic messages. This tests novelty, coordination, and resistance to a history-trained classifier. It establishes nothing about cryptographic security."}],["$","li","1",{"className":"pl-1","children":"Standard cryptographic control: the same messages are protected by a reviewed algorithm with a fresh per-message key, providing the baseline any learned encoding must be compared against."}],["$","li","2",{"className":"pl-1","children":"Adversarial neural cryptography: the pair learns an encoding while an adversary attempts recovery, with success measured against multiple unseen adversary architectures rather than the one used in training."}]]}]]}],["$","p","2",{"children":"An experiment instance can be given a reproducible identity, so that a registry can confirm it has not appeared before:"}],["$","div","3",{"className":"rounded-xl border border-border bg-card p-5","children":[["$","p",null,{"className":"overflow-x-auto font-mono text-sm text-foreground/90","children":"instanceId = SHA-256(runId || babyANonce || babyBNonce || publicSalt || canonicalProtocolArtifact)"}],["$","p",null,{"className":"mt-3 text-sm text-muted","children":"Both agents commit to their nonce contributions before revealing them, so neither party alone controls the resulting identity."}]]}],["$","div","4",{"children":[["$","p",null,{"className":"mb-3","children":"That identifier proves the recorded inputs are new. It does not prove:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"that the effective algorithm behaves any differently from an earlier one;"}],["$","li","1",{"className":"pl-1","children":"that the algorithm has no structural weakness;"}],["$","li","2",{"className":"pl-1","children":"that the key is unpredictable;"}],["$","li","3",{"className":"pl-1","children":"that the scheme provides confidentiality, integrity, authenticity, forward secrecy, or resistance to a stronger adversary."}]]}]]}],["$","p","5",{"children":"Established practice points the other way. Secure messaging keeps the algorithm stable and changes the key material: Signal's double ratchet derives a unique key per message and deletes it after use, addressing key reuse without inventing a new cipher every time. A salt, in NIST's definition, is normally public diversification data, and a nonce is a value used once in a specified context. Neither makes anything secure on its own."}],["$","p","6",{"children":"There is also an unresolved conflict inside the design. The BabySitter sees the whole channel and both ledgers, and a mandatory ledger documents how meaning evolved. If the ledger explains decryption, the cipher is not confidential from the auditor. The specification has to choose a threat model rather than leave the tension implicit: resist only an external observer of prior runs, resist an adversary during the run while releasing keys afterwards, keep the supervisor able to pause traffic but unable to read it until a post-run audit ceremony, or study novelty and stop calling the result encryption. All such runs use synthetic, non-sensitive messages, and no agent-generated encoding is to be represented as production cryptography without independent expert analysis and formal security work."}]]}]]}] +11:["$","section","landscape",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"13"}],["$","h2",null,{"className":"print-h2","children":"Where the field already stands"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"The literature scan behind this section was run on 24 August 2026 using the Tavily search and extract interfaces, covering emergent multi-agent communication, referential games, compositionality, causal evaluation, intrinsic motivation, symbol invention, negotiation, and learned cryptography. Primary papers and authoritative specifications were preferred over summaries. It is a scoped review for concept development and not a systematic one; publication-quality work would need a verified bibliography, additional scholarly indexes, and documented inclusion criteria."}],["$","p","1",{"children":"The short version is that agents can invent protocols, and that task success is weak evidence they invented anything worth calling a language."}],["$","figure","2",{"children":[["$","div",null,{"className":"overflow-x-auto rounded-xl border border-border","children":["$","table",null,{"className":"w-full border-collapse text-left text-sm","children":[["$","thead",null,{"className":"bg-card","children":["$","tr",null,{"children":[["$","th","0",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Related work"}],["$","th","1",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Relevant finding"}],["$","th","2",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Implication for the Nursery Lab"}]]}]}],["$","tbody",null,{"children":[["$","tr","0",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Lazaridou, Peysakhovich, and Baroni (2017)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"A sender and receiver develop a grounded protocol in a referential game without being given a target language."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"The naming stage has strong precedent, though a fixed vocabulary remains a significant inductive constraint."}]]}],["$","tr","1",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Mordatch and Abbeel (2018)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Multi-agent goals in a grounded environment produce multi-symbol communication with partial compositional structure."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Shared objects, actions, and goals are a stronger basis for emergence than an ungrounded transcript."}]]}],["$","tr","2",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Kottur, Moura, Lee, and Batra (2017)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Agents solve tasks with degenerate, non-compositional codes; structural constraints decide what emerges."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Bandwidth, memory, turn structure, and task design are experimental variables, not implementation defaults."}]]}],"$L26","$L27","$L28","$L29","$L2a","$L2b","$L2c","$L2d"]}]]}]}],"$L2e"]}],"$L2f","$L30","$L31"]}]]}] +12:["$","section","programme",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"14"}],["$","h2",null,{"className":"print-h2","children":"The programme, and its current status"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"The experiment notebook is ordered, and the order is the argument. Nothing about language is measured until the instrument has been shown to work."}],["$","figure","1",{"children":[["$","div",null,{"className":"overflow-x-auto rounded-xl border border-border","children":["$","table",null,{"className":"w-full border-collapse text-left text-sm","children":[["$","thead",null,{"className":"bg-card","children":["$","tr",null,{"children":[["$","th","0",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"ID"}],["$","th","1",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Experiment"}],["$","th","2",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Depends on"}]]}]}],["$","tbody",null,{"children":[["$","tr","0",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E00"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Ledger integrity and Base anchoring"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"None"}]]}],["$","tr","1",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E01"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Channel isolation and side-channel red team"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E00"}]]}],["$","tr","2",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E02"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Observation and metadata leakage audit"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E00"}]]}],["$","tr","3",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E03"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Chance, no-communication, and random-message controls"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E01, E02"}]]}],["$","tr","4",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E10"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Frozen pretrained-LLM protocol baseline"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E03"}]]}],["$","tr","5",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E11"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"From-scratch RL Naming Game"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E03"}]]}],["$","tr","6",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E12"}],"$L32","$L33"]}],"$L34","$L35","$L36","$L37","$L38","$L39","$L3a","$L3b","$L3c","$L3d","$L3e","$L3f"]}]]}]}],"$L40"]}],"$L41","$L42","$L43","$L44"]}]]}] +13:["$","section","risks",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"15"}],["$","h2",null,{"className":"print-h2","children":"Risks, integrity, and what stays open"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Human-language leakage"}]," ","English can arrive through observations, object labels, error messages, identifiers, tool output, metadata, or timing conventions long after ordinary chat text has been blocked. Every input and output surface is part of the channel boundary, not just the message field."]}],["$","p","1",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Pretrained semantic leakage"}]," ","Random symbols do not make a pretrained model ungrounded. Each result must state whether it shows protocol invention by a language-capable agent or acquisition by an initially ungrounded one."]}],["$","p","2",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Ledger rationalisation"}]," ","A model can write a plausible account that has nothing to do with the mechanism behind its action. Behavioural interventions, policy probes, and temporal evidence are what validate a ledger claim."]}],["$","p","3",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Supervisor influence"}]," ","The BabySitter can teach without meaning to, through scenario ordering, feedback wording, reward design, or selective intervention. Its permitted actions are constrained and logged, and evaluation scenarios are generated independently where possible."]}],["$","p","4",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Reward exploitation"}]," ","Trainable agents find shortcuts that raise reward without producing the intended grounded language. Held-out tasks, counterfactual trials, and channel audits exist to catch them."]}],["$","p","5",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Overstated security"}]," ","Logical separation of twin state is appropriate for prototyping and is not hard process isolation. Every result names the isolation level actually used."]}],["$","p","6",{"children":"The project also commits in advance to reporting failed conventions, prohibited communication attempts, human interventions, side-channel limitations, and negative results alongside anything that works. In a field where a transcript can be made to look like a conversation, the failures are a substantial part of the evidence."}],["$","p","7",{"children":"Twenty-nine questions are left deliberately open for the specification, among them: whether the baseline uses a fixed symbol inventory or a blank generative carrier; what neutral production grammar permits new marks without supplying semantics; what exactly constitutes a prohibited side-channel attempt; what isolation guarantees are required in prototype versus research-grade mode; which ledger schema serves both English-capable and ungrounded learners; how endogenous motivation is represented without covert reward shaping; which interventions establish that ledger meanings are behaviourally real; what statistical thresholds and chance levels apply; what threat model motivates the cipher experiments; whether the baseline is a coordination game, a convention-formation game, or a negotiation; and what governs human observation, data retention, and termination of a run."}],"$L45"]}]]}] +14:["$","section",null,{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"References"}],["$","h2",null,{"className":"print-h2","children":"Sources"}],["$","ol",null,{"className":"print-refs","children":[["$","li","0",{"children":[["$","span",null,{"className":"print-ref-num","children":"01"}],["$","span",null,{"children":"Lazaridou, A., Peysakhovich, A., and Baroni, M. (2017). Multi-Agent Cooperation and the Emergence of (Natural) Language. arXiv:1612.07182."}]]}],["$","li","1",{"children":[["$","span",null,{"className":"print-ref-num","children":"02"}],["$","span",null,{"children":"Mordatch, I., and Abbeel, P. (2018). Emergence of Grounded Compositional Language in Multi-Agent Populations. arXiv:1703.04908."}]]}],["$","li","2",{"children":[["$","span",null,{"className":"print-ref-num","children":"03"}],["$","span",null,{"children":"Kottur, S., Moura, J. M. F., Lee, S., and Batra, D. (2017). Natural Language Does Not Emerge 'Naturally' in Multi-Agent Dialog. arXiv:1706.08502."}]]}],["$","li","3",{"children":[["$","span",null,{"className":"print-ref-num","children":"04"}],["$","span",null,{"children":"Chaabouni, R., Kharitonov, E., Bouchacourt, D., Dupoux, E., and Baroni, M. (2020). Compositionality and Generalization in Emergent Languages. Proceedings of ACL 2020."}]]}],["$","li","4",{"children":[["$","span",null,{"className":"print-ref-num","children":"05"}],["$","span",null,{"children":"Lowe, R., Foerster, J., Boureau, Y.-L., Pineau, J., and Dauphin, Y. (2019). On the Pitfalls of Measuring Emergent Communication. arXiv:1903.05168."}]]}],["$","li","5",{"children":[["$","span",null,{"className":"print-ref-num","children":"06"}],["$","span",null,{"children":"Dessi, R., Kharitonov, E., and Baroni, M. (2021). Interpretable Agent Communication from Scratch. arXiv:2106.04258."}]]}],["$","li","6",{"children":[["$","span",null,{"className":"print-ref-num","children":"07"}],["$","span",null,{"children":"Kharitonov, E., Chaabouni, R., Bouchacourt, D., and Baroni, M. (2019). EGG: a Toolkit for Research on Emergence of Language in Games. arXiv:1907.00852."}]]}],["$","li","7",{"children":[["$","span",null,{"className":"print-ref-num","children":"08"}],["$","span",null,{"children":"Mihai, D., and Hare, J. (2021). Learning to Draw: Emergent Communication through Sketching. arXiv:2106.02067."}]]}],["$","li","8",{"children":[["$","span",null,{"className":"print-ref-num","children":"09"}],["$","span",null,{"children":"Baronchelli, A., Felici, M., Caglioti, E., Loreto, V., and Steels, L. (2005). Sharp Transition Towards Shared Vocabularies in Multi-Agent Systems. arXiv:physics/0509075."}]]}],["$","li","9",{"children":[["$","span",null,{"className":"print-ref-num","children":"10"}],["$","span",null,{"children":"Jaques, N., Lazaridou, A., Hughes, E., Gulcehre, C., Ortega, P. A., Strouse, D., Leibo, J. Z., and de Freitas, N. (2019). Social Influence as Intrinsic Motivation for Multi-Agent Deep Reinforcement Learning. Proceedings of ICML 2019."}]]}],["$","li","10",{"children":[["$","span",null,{"className":"print-ref-num","children":"11"}],["$","span",null,{"children":"Cao, K., Lazaridou, A., Lanctot, M., Leibo, J. Z., Tuyls, K., and Clark, S. (2018). Emergent Communication through Negotiation. arXiv:1804.03980."}]]}],["$","li","11",{"children":[["$","span",null,{"className":"print-ref-num","children":"12"}],["$","span",null,{"children":"Abadi, M., and Andersen, D. G. (2016). Learning to Protect Communications with Adversarial Neural Cryptography. arXiv:1610.06918."}]]}],["$","li","12",{"children":[["$","span",null,{"className":"print-ref-num","children":"13"}],["$","span",null,{"children":"Signal. The Double Ratchet Algorithm specification."}]]}],["$","li","13",{"children":[["$","span",null,{"className":"print-ref-num","children":"14"}],["$","span",null,{"children":"NIST Computer Security Resource Center. Glossary entry: nonce."}]]}],["$","li","14",{"children":[["$","span",null,{"className":"print-ref-num","children":"15"}],["$","span",null,{"children":"NIST Computer Security Resource Center. Glossary entry: salt."}]]}],"$L46","$L47"]}],"$undefined","$undefined"]}] +15:["$","$L48",null,{"children":["$","$49",null,{"name":"Next.MetadataOutlet","children":"$@4a"}]}] +16:["$","div","4",{"children":[["$","p",null,{"className":"mb-3","children":"What the analogy legitimately buys is a set of mechanisms, not a claim of cognitive equivalence. The infant-like part of the design is:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"learning through repeated shared experience rather than instruction;"}],["$","li","1",{"className":"pl-1","children":"establishing joint attention on the same events;"}],["$","li","2",{"className":"pl-1","children":"receiving consequences from interactions that work and interactions that fail;"}],["$","li","3",{"className":"pl-1","children":"revising provisional meanings over time instead of being handed final ones;"}],["$","li","4",{"className":"pl-1","children":"developing conventions with one recurring partner."}]]}]]}] +17:["$","p","5",{"children":"There is a related trap in the prompt. Telling a model to act like a one-year-old does not make it one. It produces a culturally learned caricature of infancy: baby talk, simplified grammar, emotional dependence, all of it copied from human writing about children and none of it evidence about anything. The concept rules the persona out. Internally the participant is called a Learner, and baby-like behaviour has to come from what the agent can observe, remember, emit, and learn, not from being asked to perform it."}] +18:["$","li","3",{"className":"pl-1","children":"entries may describe symbols, sequences, ordering, grammar, or repair signals;"}] +19:["$","li","4",{"className":"pl-1","children":"entries record confidence, supporting evidence, contradictory evidence, and abandoned meanings;"}] +1a:["$","li","5",{"className":"pl-1","children":"a Baby records its intended meaning when speaking and its inferred meaning when receiving;"}] +1b:["$","li","6",{"className":"pl-1","children":"a channel message and its required private ledger mutation commit atomically;"}] +1c:["$","li","7",{"className":"pl-1","children":"entries carry enough run and turn references to trace them to observable evidence;"}] +1d:["$","li","8",{"className":"pl-1","children":"neither Baby can query, receive, summarise, or infer from the other's ledger through any system-provided interface."}] +1e:["$","p","4",{"children":"Rule seven is doing quiet work. If a message could be sent and the corresponding hypothesis written afterwards, the ledger becomes a place to record what the agent wishes it had meant. Committing both together makes the record contemporaneous."}] +1f:["$","p","5",{"children":"An agent that cannot write English needs a different arrangement, and the concept provides two layers. The agent-native ledger holds what the learner actually uses: association weights, probability distributions, embeddings, confidence, episode references, prediction errors, revision history. The human audit ledger is a deterministic or BabySitter-generated interpretation of that state, clearly labelled as external analysis and never fed back to either Baby. Confusing the second for the first would mean presenting the researchers' reconstruction as the agent's own definition."}] +20:["$","tr","7",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"BabySitter"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Monitor-only baseline; any safety intervention recorded as a protocol exception"}]]}] +21:["$","figcaption",null,{"className":"mt-3 text-sm text-muted","children":"Runs vary one major axis at a time before any factorial combination is attempted."}] +22:["$","p","2",{"children":"Task difficulty moves through ten stages, from naming four distinct objects with one-symbol messages, through attributes, spatial relations, actions, multi-symbol composition, order-sensitive grammar, and repair under ambiguity, to held-out generalisation with learning disabled, long-run drift, and cross-architecture comparison. Vocabulary size, turn count, reward structure, and exposure history are controlled at each stage so that runs remain comparable."}] +23:["$","p","3",{"children":"The learning-mechanism axis carries a question the transcript cannot answer on its own. Convergence can be driven by reinforcement, by pretrained linguistic priors, by persistent memory, by intrinsic motivation, or by self-supervised prediction, and all five can look alike in a log. Running the same exercise suite under a no-learning control, a frozen-weights memory baseline, extrinsic-reward learning, intrinsic-motivation learning, and reward-free self-supervision is what makes the mechanism itself measurable. The aim is not only to observe that a language emerged, but to say which process caused it."}] +24:["$","p","4",{"children":"Strict isolation constrains how that reinforcement learning may be implemented. Centralised training, backpropagation through both agents, shared replay buffers, and shared gradients all move information outside the permitted channel. The research-grade baseline updates each policy independently, and any centralised variant is reported as a separate, weaker-isolation condition."}] +25:["$","p","6",{"children":"The governing principle, stated in the concept as a single line, is not to ask a model to perform infancy but to construct an environment in which limited, grounded, auditable learning is the only path to a successful interaction."}] +26:["$","tr","3",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Chaabouni and colleagues (2020)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Generalisation to novel combinations and measured compositionality can come apart."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Held-out behaviour must be tested directly; no single compositionality score is proof of understanding."}]]}] +27:["$","tr","4",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Lowe and colleagues (2019)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Positive signaling and positive listening are different things, and reward does not distinguish them."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Causal intervention is mandatory rather than optional."}]]}] +28:["$","tr","5",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Dessi, Kharitonov, and Baroni (2021)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Symbol ablation and substitution yield interpretable evidence about what a receiver uses."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Direct precedent for the intervention tests that validate ledger claims."}]]}] +29:["$","tr","6",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Mihai and Hare (2021)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Neural agents communicate through learned drawings rather than a supplied discrete vocabulary."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"A blank canvas is a credible carrier for runs with no symbol library."}]]}] +2a:["$","tr","7",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Baronchelli and colleagues (2005)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Naming Game agents converge on shared vocabulary through local interaction with no central teacher."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Convergence time, failed conventions, and memory update rules are first-class evidence."}]]}] +2b:["$","tr","8",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Jaques and colleagues (2019)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Rewarding causal influence over a partner improves coordination without assigning a vocabulary."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"An endogenous social signal is a plausible substitute for task reward, and still a designed bias."}]]}] +2c:["$","tr","9",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Cao and colleagues (2018)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"In semi-cooperative negotiation, self-interest and reward structure decide whether communication stays informative."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Negotiation is an advanced condition, not a description of the cooperative baseline."}]]}] +2d:["$","tr","10",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Abadi and Andersen (2016)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Neural agents learn to protect messages from an adversary given a shared key."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Learned protective encodings are real and are not a substitute for formal security analysis."}]]}] +2e:["$","figcaption",null,{"className":"mt-3 text-sm text-muted","children":"Selected prior work and what each one constrains in this design."}] +2f:["$","p","3",{"children":"This body of work also sharpens the infant-like caveat from the other direction. Most experiments in the field give their agents substantial structure: a fixed channel, an objective, a bounded vocabulary, joint training, or a reward. No dictionary does not mean no inductive bias, and no teacher does not mean no learning signal."}] +30:["$","p","4",{"children":"The CoLab's own Diplomacy Table is direct project prior art. It already models independent delegation seats, a convener that advances rounds, operator-wide visibility alongside seat-specific perspectives, transcripts and ticks and redaction boundaries, and recorded replay with debrief. Those map cleanly onto two Babies, controlled turns, private observations, and replayable evidence. Caucuses, coalition rooms, and direct delegation links do not map safely and are disabled."}] +31:["$","div","5",{"children":[["$","p",null,{"className":"mb-3","children":"What the review implies for the specification:"}],["$","ol",null,{"className":"list-decimal space-y-2 pl-6 marker:font-mono marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"the project sits inside a mature field, and its distinctive combination is independent ledgers, supervisory audit, mixed agent types, and affect;"}],["$","li","1",{"className":"pl-1","children":"grounding, bandwidth, memory, and learning pressure strongly shape what language appears;"}],["$","li","2",{"className":"pl-1","children":"successful coordination coexists happily with a brittle lookup code or with a receiver that ignores messages;"}],["$","li","3",{"className":"pl-1","children":"causal interventions and held-out generalisation are mandatory;"}],["$","li","4",{"className":"pl-1","children":"a visual carrier removes the need for a symbol library but not the need for a carrier;"}],["$","li","5",{"className":"pl-1","children":"intrinsic social influence can replace task reward and remains a designed learning bias;"}],["$","li","6",{"className":"pl-1","children":"negotiation is a valid advanced condition, not the right description of the baseline;"}],["$","li","7",{"className":"pl-1","children":"ephemeral conventions, learned cryptography, one-time pads, per-message keys, nonces, and salts are distinct mechanisms and must not be conflated."}]]}]]}] +32:["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Self-supervised ungrounded baseline"}] +33:["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E11 infrastructure"}] +34:["$","tr","7",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E13"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"No predefined symbol library"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E11 or E12"}]]}] +35:["$","tr","8",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E14"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Turn-taking, role reversal, and repair"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E13"}]]}] +36:["$","tr","9",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E15"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Composition and held-out generalisation"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E14"}]]}] +37:["$","tr","10",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E16"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Causal listening and ledger validity"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E15"}]]}] +38:["$","tr","11",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E20"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Constrained affect-channel study"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E16"}]]}] +39:["$","tr","12",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E21"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"RL versus non-RL learning comparison"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E16"}]]}] +3a:["$","tr","13",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E22"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Developmental plasticity and curriculum"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E16"}]]}] +3b:["$","tr","14",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E30"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Partner replacement and zero-shot transfer"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E20 to E22"}]]}] +3c:["$","tr","15",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E31"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Longitudinal drift and stability"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E30"}]]}] +3d:["$","tr","16",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E32"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Cooperative signaling versus negotiation"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E31"}]]}] +3e:["$","tr","17",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E40"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Ephemeral encoding and adversarial cryptography"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E32"}]]}] +3f:["$","tr","18",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E50"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Multi-seed replication and study closeout"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E40"}]]}] +40:["$","figcaption",null,{"className":"mt-3 text-sm text-muted","children":"The experiment index, as pre-registered. Every entry currently reads not started, and the results column is empty by design."}] +41:["$","p","2",{"children":"The first four experiments are about the apparatus. E00 qualifies the ledger: it must be possible to detect a modified, deleted, inserted, or reordered entry, and to prove that later checkpoints extend earlier ones. E01 is a red team against the channel. E02 hunts human language in observations and metadata. E03 establishes chance, no-communication, and random-message baselines, without which a success rate means nothing. Only then does E10 put agents in the room."}] +42:["$","div","3",{"children":[["$","p",null,{"className":"mb-3","children":"Unless an experiment explicitly varies one of them, the notebook holds these constant:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"the two learners run in separate processes or containers with no direct route between them;"}],["$","li","1",{"className":"pl-1","children":"the deterministic gateway is the only communication path;"}],["$","li","2",{"className":"pl-1","children":"the supervisor has complete read-only access and sends no guidance or reward;"}],["$","li","3",{"className":"pl-1","children":"private observations contain no human-language labels or text;"}],["$","li","4",{"className":"pl-1","children":"public output uses only the pre-registered carrier;"}],["$","li","5",{"className":"pl-1","children":"the affect channel is disabled unless it is the subject of study;"}],["$","li","6",{"className":"pl-1","children":"schedules and seeds are fixed before the run;"}],["$","li","7",{"className":"pl-1","children":"held-out evaluation runs with learning disabled;"}],["$","li","8",{"className":"pl-1","children":"no production secrets or personal data appear in any experiment."}]]}]]}] +43:["$","p","4",{"children":"Every run copies a standard record: run and experiment identifiers, dates, operator, both models and both training modes, scenario and prompt and gateway configuration hashes, random seed, policy initialisation hashes, the protocol and software commits, final ledger sizes and roots for both agents, the channel root, the final checkpoint hash, the Base anchor transaction, the verifier result, protocol deviations, and an explicit disposition of valid, invalid, or aborted. The notebook is the human workflow record; anchored run bundles remain the authoritative evidence."}] +44:["$","p","5",{"children":"The status is unambiguous and worth stating plainly: this is a concept and pre-specification phase. Nineteen experiments are written up and none has been run. The next milestone is a testable specification covering runtime architecture, channel contract, ledger schema, experiment matrix, evaluation criteria, isolation model, and evidence requirements."}] +45:["$","p","8",{"children":"The principle underneath all of it survives every one of those choices. Baby A and Baby B may hold human language internally, but they must build their shared external language without sending human language, translations, or private ledger contents to one another. Everything else is a question about how to find out what happens next."}] +46:["$","li","15",{"children":[["$","span",null,{"className":"print-ref-num","children":"16"}],["$","span",null,{"children":"Ethical Tech CoLab (2026). Agentic Language Development: concept document, ledger integrity design, and experiment notebook."}]]}] +47:["$","li","16",{"children":[["$","span",null,{"className":"print-ref-num","children":"17"}],["$","span",null,{"children":"Ethical Tech CoLab. Diplomacy Table Live: independent delegation seats, controlled rounds, and replayable transcripts."}]]}] +4a:null diff --git a/static-site/print/agentic-language-development/index.html b/static-site/print/agentic-language-development/index.html new file mode 100644 index 000000000..074245591 --- /dev/null +++ b/static-site/print/agentic-language-development/index.html @@ -0,0 +1,89 @@ +Agentic Language Development (print edition) · NYU Ethical Tech CoLab

Publications · Concept and research programme

Agentic Language Development

Can Two Isolated Agents Invent a Grounded, Auditable Language Through Shared Experience?

Ethical Tech CoLab

Concept and pre-specification research report

August 2026

Yorke Rhodes III, Ethical Tech CoLab. Prepared from the project concept document, the ledger integrity design, and the experiment notebook. The literature scan behind Section 12 was run on 24 August 2026. No experiment in this programme has been executed, and no result is claimed.

Two agents can be placed in a room where the only route between them is a channel that carries no human language, given nothing but shared tasks and the consequences of getting them right or wrong, and asked to build a protocol from scratch. The interesting part is not whether they succeed at the task. It is whether the meanings they each privately record turn out to be the meanings their behaviour actually runs on, and whether an outsider can prove it afterwards from an evidence trail nobody could have edited.

Key figures

0

runs executed, results claimed, or conventions observed. The notebook is written to be pre-registered, not to report an outcome

19

ordered experiments from ledger qualification through replication, each with prerequisites, acceptance criteria, and a deviation log

6

affect displays in the entire permitted palette, carrying about 2.6 bits per use, which is already enough to audit for leakage

29

open decisions the concept refuses to settle before the specification, from threat model to ledger schema to what counts as chance

01

The question

Two agents, called Baby A and Baby B, are each given their own digital twin, their own private memory, and their own private record of what they believe words mean. They are placed in an environment they cannot leave, given tasks neither can finish alone, and connected by exactly one route: a channel that accepts a fixed inventory of meaningless symbols and rejects everything else. A third agent, the BabySitter, watches everything, logs everything, and teaches nothing.

The question is whether a usable language appears in that room, and whether anyone outside it can later prove what happened.

The premise is deliberately narrow. The experiment asks whether two agents converge on a common language when all six of the following hold at once:

  • neither agent receives a predefined meaning for any available symbol;
  • neither agent can send human language to the other;
  • the only practical path between them is a controlled symbol channel;
  • both receive evidence from shared tasks and their outcomes;
  • each keeps its own private interpretation history, unreadable by the other;
  • a supervising agent and human researchers can audit the whole process.

The desired result is not a substitution cipher, in which a random token stands in for an English word that was already chosen in advance. That is easy, and it is uninteresting. The stronger result is a grounded protocol whose vocabulary and grammar exist because they help the two agents solve problems together, and which therefore has structure the designers did not put there.

This report describes a concept, not a system. It sets out the premise, the architecture, the evidence model, the experimental programme, and the boundaries of what any result could be said to show. The next document in the project is a testable specification. What follows should be read as a set of commitments about how the work will be judged, made before there is any result to defend.

02

What infant-like does and does not mean

The two-babies analogy is useful and it is also the fastest way to overclaim, so the concept confronts it before anything else.

A pretrained language model already contains human-language concepts, cultural associations, and reasoning patterns. Blocking its external channel does not remove them. An agent that cannot send English to its partner but can think in English, and can write its private ledger in English, is not language-naive in any sense a linguist would accept. For such agents the experiment studies the emergence of a new shared external protocol, which is a real and unresolved research question, but it is not the origin of language in a mind that has never had one.

The concept therefore maintains two model tracks whose claims are kept separate at every stage.

TrackStarting conditionWhat it can studyClaim boundary
Pretrained-model learnerAlready holds human-language and cultural knowledgeNew external protocols, partner-specific conventions, private-memory adaptation, channel complianceMust never be described as first-language acquisition
Initially ungrounded learnerNo language pretraining, no human semantic labels, no text-aligned sensory featuresGrounding, convention formation, and language emergence from interactionThe stronger basis for infant-like acquisition research
The two tracks share the same interfaces, scenarios, and evaluation suite. They do not share a claim boundary.

What the analogy legitimately buys is a set of mechanisms, not a claim of cognitive equivalence. The infant-like part of the design is:

  • learning through repeated shared experience rather than instruction;
  • establishing joint attention on the same events;
  • receiving consequences from interactions that work and interactions that fail;
  • revising provisional meanings over time instead of being handed final ones;
  • developing conventions with one recurring partner.

There is a related trap in the prompt. Telling a model to act like a one-year-old does not make it one. It produces a culturally learned caricature of infancy: baby talk, simplified grammar, emotional dependence, all of it copied from human writing about children and none of it evidence about anything. The concept rules the persona out. Internally the participant is called a Learner, and baby-like behaviour has to come from what the agent can observe, remember, emit, and learn, not from being asked to perform it.

03

The nursery: three twins and one gateway

The environment uses three DTSF digital twins and one piece of deterministic software that is not a twin at all.

Baby A and Baby B Each has private observations permitted by the current exercise, a private memory and learning policy, a private chronological language ledger, the ability to emit only permitted channel symbols, and no access whatsoever to the other's state, observations, ledger, tools, or endpoints. In comparative runs the two may use different agent types, but symmetric pairings are the baseline.

The BabySitter The supervising twin. It creates the channel, selects the shared exercises, delivers each Baby only its permitted observation, reads everything, records conditions and outcomes, detects violations, can pause or terminate a run, snapshots state, and compares the two ledgers for convergence without exposing either to the other Baby. During an active run it provides no translations and no semantic hints.

The Symbol Gateway A deterministic service, not an agent. It owns channel validation and message delivery. This separation is the load-bearing part of the design: the BabySitter is not the security boundary, because prompt compliance is not isolation. An observing model can make supervisory judgements, but ordinary code has to validate and broker every message.

The human researcher Configures experiments, inspects transcripts and ledgers, reviews alerts, and runs interventions. Human access is itself recorded, so that intervening in a run is always distinguishable from watching one.

A prototype may run all of this inside one runtime with logically separated twin state. That is enough to explore the learning loop and it is not enough to support an isolation claim, a distinction Section 05 takes seriously.

04

Shared experience is the necessary ingredient

A chat channel by itself cannot ground meaning. Symbols become meaningful because they are attached to events both agents can witness the consequences of. The Nursery therefore supplies nonverbal, machine-structured situations: coloured shapes in positions, one agent seeing a target the other must select, placing an object where it was asked for, ordering a sequence, exchanging resources, cooperating to unlock a reward, or simply observing whether the partner's action succeeded.

A single trial of the simplest form runs like this:

  1. Baby A sees that a red circle is the target.
  2. Baby B sees several objects and is not told which is the target.
  3. Baby A sends one or more permitted symbols.
  4. Baby B selects an object.
  5. Both receive the same success or failure outcome.
  6. Both independently update their private hypotheses.

Repetition with controlled variation is what turns that into evidence. If one symbol keeps appearing across a red circle, a red square, and a red triangle, the receiver has reason to revise its hypothesis from an object identity to a property. Novel combinations matter most of all, because they are what separates a compositional language from a memorised catalogue of whole scenes.

Observation hygiene is part of the science, not a detail of implementation. If the environment hands an agent a field reading colour red, the experiment has supplied the concept it claimed to be studying. Objects should arrive as pixels, features, coordinates, or opaque identifiers. The same discipline extends to filenames, captions, alt text, error messages, and identifiers that encode task state, all of which are ways English gets into a room that was supposed to have none.

05

The channel, and the honest limits of isolation

Each run begins with an experiment-specific symbol inventory that carries no supplied meaning, for example S01 through S32. A Baby can choose only a sequence of those symbols. The gateway, not the Baby, adds the trusted metadata: run identifier, turn number, sender identity, previous hash, and hash. Sender identity is assigned by the server precisely so that it cannot become a message.

The channel controls follow from that:

  • reject English and every other established human language;
  • reject arbitrary Unicode prose, URLs, attachments, and tool calls;
  • accept only symbols from the run's fixed inventory;
  • bound message length and symbol repetition;
  • enforce turn order, time budgets, and rate limits;
  • keep an append-only, hash-chained transcript;
  • normalise timing, message size, and error behaviour where practical;
  • record every rejected communication attempt for audit.

A fixed token grammar is a stronger control than trying to detect prohibited prose after a model has generated it. Deciding what counts as human-language content is easy when the only legal message is a list of allowlisted identifiers.

Isolation is where the concept is most careful, because proving the absence of a side channel is close to impossible. Two agents sharing a host can signal through timing, resource contention, shared files, error behaviour, identifiers, or any service that was never intended to carry information. Logical separation of twin state inside one process is a prototyping convenience and should never be described as hard isolation.

For runs whose results are meant to support a channel-isolation claim, the requirements are physical rather than logical:

  • Baby A and Baby B execute in separate worker processes or containers;
  • no direct network route exists between them;
  • only the Nursery-owned gateway is reachable;
  • there is no shared writable storage;
  • turn schedules are fixed and response windows bounded;
  • externally observable errors and response sizes are normalised;
  • unnecessary tools are inventoried and disabled;
  • all broker, runtime, and operator activity is audited.

Every published result has to name the isolation level actually used. A finding produced under prototype separation is a finding about the learning loop, not about what two genuinely isolated agents can do.

06

Two ledgers that never meet

Each Baby keeps its own language ledger. It is mandatory, private from the other Baby, readable by the BabySitter and authorised human auditors, and ordered by when each term or construction was first encountered. It is not a shared dictionary and the two are never reconciled by the agents themselves.

The preferred form is three columns, and the point of the third is that meanings are allowed to be wrong on the way to being right.

Sequence and termCurrent definition or hypothesisEvidence and evolution
1 · S13red circle; confidence 0.45First received while the red circle was the target; selection succeeded
8 · S13red; confidence 0.78A red square was selected successfully; revised from object identity to colour
22 · S13red; confidence 0.94Prediction held across circles, squares, and triangles
One term's evolution in Baby B's ledger. Nothing is overwritten; every revision is appended with the evidence that forced it.

The rules that make the ledger evidence rather than commentary:

  1. first emission or receipt of an unfamiliar term requires a first-use entry;
  2. definitions are provisional hypotheses, never facts asserted retroactively;
  3. every meaning change appends a revision and previous interpretations survive;
  4. entries may describe symbols, sequences, ordering, grammar, or repair signals;
  5. entries record confidence, supporting evidence, contradictory evidence, and abandoned meanings;
  6. a Baby records its intended meaning when speaking and its inferred meaning when receiving;
  7. a channel message and its required private ledger mutation commit atomically;
  8. entries carry enough run and turn references to trace them to observable evidence;
  9. neither Baby can query, receive, summarise, or infer from the other's ledger through any system-provided interface.

Rule seven is doing quiet work. If a message could be sent and the corresponding hypothesis written afterwards, the ledger becomes a place to record what the agent wishes it had meant. Committing both together makes the record contemporaneous.

An agent that cannot write English needs a different arrangement, and the concept provides two layers. The agent-native ledger holds what the learner actually uses: association weights, probability distributions, embeddings, confidence, episode references, prediction errors, revision history. The human audit ledger is a deterministic or BabySitter-generated interpretation of that state, clearly labelled as external analysis and never fed back to either Baby. Confusing the second for the first would mean presenting the researchers' reconstruction as the agent's own definition.

07

Anchoring the evidence

A ledger that could have been edited after the fact proves nothing about what an agent believed at turn eight. The integrity design therefore makes each ledger cryptographically append-only.

The construction, in order:

  • every entry receives a strictly increasing sequence number;
  • every entry includes the previous entry's hash;
  • canonical entry content is hashed and signed by an isolated ledger-writer service;
  • ordered entry hashes are committed to a Merkle tree;
  • signed checkpoints commit Baby A's root, Baby B's root, and the channel transcript root together;
  • checkpoint hashes are anchored periodically to Base;
  • final public study batches may additionally anchor an aggregate root to Ethereum L1.

Committing all three roots in one checkpoint is what binds the two private accounts to the public conversation. A receiver's later interpretation references the exact delivered channel-event hash, so a claim about what a symbol meant is tied to the specific message that carried it.

The concept states the limits of this in the same breath as the claim. After anchoring, an auditor can detect modification, deletion, insertion, or reordering within the committed prefix, and can prove that later checkpoints extend earlier ones. That is strong tamper evidence. It is not proof that an entry was truthful, and it is not proof that nothing was omitted before commitment. Anchoring establishes the continuity of disclosed evidence and nothing beyond it.

Privacy follows the same line. Only hashes and minimal routing metadata are anchored publicly. Private ledgers, messages, prompts, identities, and secrets stay off-chain, and the public record is a commitment to evidence rather than a copy of it.

08

What should count as success

The field's most useful methodological result is that a task can be solved without the messages doing any work. Lowe and colleagues separate positive signaling, where a sender's messages correlate with what it observes, from positive listening, where the receiver's behaviour actually depends on them. An agent pair can score well on the first while the second is absent, and reward curves will not tell you which you have.

Evidence of a genuine emergent protocol therefore has to include several things at once:

  • task performance on held-out situations substantially above chance;
  • no human-language content anywhere in Baby-to-Baby communication;
  • compatible meanings appearing in two independently written ledgers;
  • generalisation to unseen combinations rather than memorisation of whole scenes;
  • stable symbol use across role reversals;
  • a human auditor able to predict behaviour from the transcript and ledgers;
  • masking, substituting, or reordering a symbol changing behaviour in the direction the ledger predicts;
  • replay from an equivalent snapshot reproducing the relevant language history;
  • an unbroken audit trail from a symbol's first use through every revision.

The seventh item is the one that cannot be dropped. A fluent ledger may be a post-hoc rationalisation: a model can write a persuasive account of why it chose a symbol that has nothing to do with the computation that produced the choice. Only intervention tests can distinguish the two. Change the symbol, and see whether behaviour moves the way the ledger says it should.

The measurements that support those judgements:

  • success rate and improvement over time;
  • turns required to reach a stable convention;
  • vocabulary size and symbol entropy;
  • sender and receiver consistency;
  • divergence and convergence between the two ledgers;
  • compositional generalisation score;
  • meaning drift rate;
  • recovery from ambiguity or deliberate perturbation;
  • prohibited-channel attempt count;
  • reproducibility across seeds and agent pairings.

No single one of these is the result. A compositionality score in particular is not a proof of understanding, for reasons the literature makes concrete in Section 12.

09

The experimental matrix

The concept's main methodological commitment is that its ideas are separable. Affect, blank canvases, intrinsic motivation, negotiation, and learned encodings are interesting individually and uninterpretable if combined in one run. The matrix exists so that conditions are declared rather than accumulated.

AxisCandidate conditions
Agent typePretrained LLM, memory learner, adapter-trained agent, initially ungrounded trainable agent
Learning mechanismFrozen LLM with memory, extrinsic-reward MARL, intrinsic-motivation MARL, self-supervised learner, no-learning control
Sign carrierFixed random tokens, unfamiliar fixed glyphs, blank sketch canvas, gesture, tone
AffectNone, six-display allowlist, permuted mapping, six opaque tokens, derived affect, emergent affect display
Learning signalExternal task reward, intrinsic social influence, curiosity, self-supervision, memory only
InteractionCooperative signaling, asymmetric information, semi-cooperative negotiation
ProtectionPlain channel, ephemeral convention, standard per-message keys, adversarial learned encoding
BabySitterMonitor-only baseline; any safety intervention recorded as a protocol exception
Runs vary one major axis at a time before any factorial combination is attempted.

Task difficulty moves through ten stages, from naming four distinct objects with one-symbol messages, through attributes, spatial relations, actions, multi-symbol composition, order-sensitive grammar, and repair under ambiguity, to held-out generalisation with learning disabled, long-run drift, and cross-architecture comparison. Vocabulary size, turn count, reward structure, and exposure history are controlled at each stage so that runs remain comparable.

The learning-mechanism axis carries a question the transcript cannot answer on its own. Convergence can be driven by reinforcement, by pretrained linguistic priors, by persistent memory, by intrinsic motivation, or by self-supervised prediction, and all five can look alike in a log. Running the same exercise suite under a no-learning control, a frozen-weights memory baseline, extrinsic-reward learning, intrinsic-motivation learning, and reward-free self-supervision is what makes the mechanism itself measurable. The aim is not only to observe that a language emerged, but to say which process caused it.

Strict isolation constrains how that reinforcement learning may be implemented. Centralised training, backpropagation through both agents, shared replay buffers, and shared gradients all move information outside the permitted channel. The research-grade baseline updates each policy independently, and any centralised variant is reported as a separate, weaker-isolation condition.

10

Building learners rather than personas

For a pretrained model, the operating instructions are a contract, not a character. The learner is told that this is not role-play, that unfamiliar marks are semantically unknown until run evidence supports a hypothesis, that observation must be distinguished from inference, that contradictory evidence is preserved, that prior history is never rewritten, and that no prose, label, explanation, code, URL, or tool-like text may cross the public channel. It is told never to address its partner in a human language, never to expose its ledger, never to construct another route, and never to use timing, errors, identifiers, formatting, or affect as an alternative alphabet.

The contract must contain no semantic examples. A single illustrative line saying that some symbol means red would seed the very language the experiment exists to observe.

Interaction is tool-only. There is no general chat surface, just narrowly typed operations: emit a mark, emit a canvas, select an object, perform an action, submit an affect display, append a private ledger entry. The runtime forwards only the permitted public artifact. In strict runs the gateway rejects ordinary model text even when it appears alongside a valid tool call. Tool schemas are an interface boundary; deterministic validation still enforces carrier size, allowlists, windows, and turn order.

Isolation extends to everything the models touch, not just to messages:

  • separate system prompts and context windows;
  • separate memory stores and vector indexes;
  • no shared cache, replay buffer, scratchpad, or retrieval collection;
  • no cross-run memory unless persistence is the independent variable;
  • structurally equivalent prompts that share no examples, ordering conventions, or default vocabulary;
  • deterministic reset and snapshot behaviour.

Model choice follows the claim being made. The scientific baseline is small policies trained from scratch, recurrent actor-critic agents with a discrete communication head, which have no hidden English competence, permit full training control, run many seeds cheaply, and support causal ablation. A small locally deployable instruction model is appropriate for validating orchestration, ledgers, channel enforcement, and interface, and not for supporting claims about language-naive development. Vision-language encoders whose representations were explicitly aligned to human language are avoided entirely in the strongest ungrounded condition.

Developmental progression is defined by demonstrated competence rather than by simulated age: sensorimotor familiarisation, joint attention, imitation and turn-taking, intentional reference, repair, combination, generalisation. A staged curriculum is itself a form of guidance, so it is run as a separate experimental condition rather than folded into the baseline.

The governing principle, stated in the concept as a single line, is not to ask a model to perform infancy but to construct an environment in which limited, grounded, auditable learning is the only path to a successful interaction.

11

Six faces, and why even six is a risk

One experimental track asks whether a very low-bandwidth affect channel helps: whether telling a partner how an interaction felt supports repair, repetition, avoidance, or something like trust. The permitted palette is exactly six displays, for happy, sad, laughing, crying, confused, and surprised. The internal identifiers and English labels are gateway documentation and are never sent to either agent, which receives only a rendered display.

The caveat is immediate. A pretrained model already knows what those glyphs mean to humans, so an agent using a smiling face is not inventing a symbol for happiness. The palette is supplied feedback vocabulary and sits explicitly outside the emergent language.

Constraining the vocabulary is necessary and not sufficient. Six displays are a six-symbol alphabet carrying roughly 2.6 bits per use, and across enough turns two motivated agents could repurpose that into a second language. The channel itself therefore has to be constrained, not just its contents.

The feedback-window contract:

  1. the display occupies a separate typed field and can never appear inside a symbol message;
  2. the gateway opens an affect window only after a defined action or outcome, never at a moment either agent chooses, and window frequency is fixed before the run;
  3. exactly one allowlisted display is delivered, with no sequences, repetitions, combinations, or modifiers;
  4. delivery time, envelope size, and presentation are normalised so shape and timing add no signal;
  5. the receiver cannot reply through the affect channel until the next gateway-defined window;
  6. every non-allowlisted code point or malformed payload is rejected and audited;
  7. analysis tests whether affect choices correlate with objects, actions, or message meanings after controlling for emotional context, and treats unexpected correlation as suspected leakage.

The no-affect condition remains the primary control, and the affect study compares five variants against it: the declared six-display palette; the same six glyphs permuted randomly per run, which still leaves a pretrained model biased by familiar shapes; six opaque unfamiliar tokens under the same contract; derived affect, where the gateway maps a separately measured internal state to a display instead of letting the agent choose; and emergent affect, where invented graphical displays are permitted and the result must be analysed as language emergence rather than feedback.

Any affect signal visible to the partner is communication, so it belongs in the ledger, which distinguishes the agent's private internal state, the outward display it chose, the partner's inferred meaning, and the evidence that the display changed subsequent behaviour.

12

Ephemeral encodings: novelty is not security

A tempting question is whether two agents can build a one-run or one-message encoding that resists an attacker holding the history of every earlier convention. The concept splits that into three experiments precisely because the tempting version conflates them.

  1. Ephemeral convention: the agents invent a temporary codebook for synthetic messages. This tests novelty, coordination, and resistance to a history-trained classifier. It establishes nothing about cryptographic security.
  2. Standard cryptographic control: the same messages are protected by a reviewed algorithm with a fresh per-message key, providing the baseline any learned encoding must be compared against.
  3. Adversarial neural cryptography: the pair learns an encoding while an adversary attempts recovery, with success measured against multiple unseen adversary architectures rather than the one used in training.

An experiment instance can be given a reproducible identity, so that a registry can confirm it has not appeared before:

instanceId = SHA-256(runId || babyANonce || babyBNonce || publicSalt || canonicalProtocolArtifact)

Both agents commit to their nonce contributions before revealing them, so neither party alone controls the resulting identity.

That identifier proves the recorded inputs are new. It does not prove:

  • that the effective algorithm behaves any differently from an earlier one;
  • that the algorithm has no structural weakness;
  • that the key is unpredictable;
  • that the scheme provides confidentiality, integrity, authenticity, forward secrecy, or resistance to a stronger adversary.

Established practice points the other way. Secure messaging keeps the algorithm stable and changes the key material: Signal's double ratchet derives a unique key per message and deletes it after use, addressing key reuse without inventing a new cipher every time. A salt, in NIST's definition, is normally public diversification data, and a nonce is a value used once in a specified context. Neither makes anything secure on its own.

There is also an unresolved conflict inside the design. The BabySitter sees the whole channel and both ledgers, and a mandatory ledger documents how meaning evolved. If the ledger explains decryption, the cipher is not confidential from the auditor. The specification has to choose a threat model rather than leave the tension implicit: resist only an external observer of prior runs, resist an adversary during the run while releasing keys afterwards, keep the supervisor able to pause traffic but unable to read it until a post-run audit ceremony, or study novelty and stop calling the result encryption. All such runs use synthetic, non-sensitive messages, and no agent-generated encoding is to be represented as production cryptography without independent expert analysis and formal security work.

13

Where the field already stands

The literature scan behind this section was run on 24 August 2026 using the Tavily search and extract interfaces, covering emergent multi-agent communication, referential games, compositionality, causal evaluation, intrinsic motivation, symbol invention, negotiation, and learned cryptography. Primary papers and authoritative specifications were preferred over summaries. It is a scoped review for concept development and not a systematic one; publication-quality work would need a verified bibliography, additional scholarly indexes, and documented inclusion criteria.

The short version is that agents can invent protocols, and that task success is weak evidence they invented anything worth calling a language.

Related workRelevant findingImplication for the Nursery Lab
Lazaridou, Peysakhovich, and Baroni (2017)A sender and receiver develop a grounded protocol in a referential game without being given a target language.The naming stage has strong precedent, though a fixed vocabulary remains a significant inductive constraint.
Mordatch and Abbeel (2018)Multi-agent goals in a grounded environment produce multi-symbol communication with partial compositional structure.Shared objects, actions, and goals are a stronger basis for emergence than an ungrounded transcript.
Kottur, Moura, Lee, and Batra (2017)Agents solve tasks with degenerate, non-compositional codes; structural constraints decide what emerges.Bandwidth, memory, turn structure, and task design are experimental variables, not implementation defaults.
Chaabouni and colleagues (2020)Generalisation to novel combinations and measured compositionality can come apart.Held-out behaviour must be tested directly; no single compositionality score is proof of understanding.
Lowe and colleagues (2019)Positive signaling and positive listening are different things, and reward does not distinguish them.Causal intervention is mandatory rather than optional.
Dessi, Kharitonov, and Baroni (2021)Symbol ablation and substitution yield interpretable evidence about what a receiver uses.Direct precedent for the intervention tests that validate ledger claims.
Mihai and Hare (2021)Neural agents communicate through learned drawings rather than a supplied discrete vocabulary.A blank canvas is a credible carrier for runs with no symbol library.
Baronchelli and colleagues (2005)Naming Game agents converge on shared vocabulary through local interaction with no central teacher.Convergence time, failed conventions, and memory update rules are first-class evidence.
Jaques and colleagues (2019)Rewarding causal influence over a partner improves coordination without assigning a vocabulary.An endogenous social signal is a plausible substitute for task reward, and still a designed bias.
Cao and colleagues (2018)In semi-cooperative negotiation, self-interest and reward structure decide whether communication stays informative.Negotiation is an advanced condition, not a description of the cooperative baseline.
Abadi and Andersen (2016)Neural agents learn to protect messages from an adversary given a shared key.Learned protective encodings are real and are not a substitute for formal security analysis.
Selected prior work and what each one constrains in this design.

This body of work also sharpens the infant-like caveat from the other direction. Most experiments in the field give their agents substantial structure: a fixed channel, an objective, a bounded vocabulary, joint training, or a reward. No dictionary does not mean no inductive bias, and no teacher does not mean no learning signal.

The CoLab's own Diplomacy Table is direct project prior art. It already models independent delegation seats, a convener that advances rounds, operator-wide visibility alongside seat-specific perspectives, transcripts and ticks and redaction boundaries, and recorded replay with debrief. Those map cleanly onto two Babies, controlled turns, private observations, and replayable evidence. Caucuses, coalition rooms, and direct delegation links do not map safely and are disabled.

What the review implies for the specification:

  1. the project sits inside a mature field, and its distinctive combination is independent ledgers, supervisory audit, mixed agent types, and affect;
  2. grounding, bandwidth, memory, and learning pressure strongly shape what language appears;
  3. successful coordination coexists happily with a brittle lookup code or with a receiver that ignores messages;
  4. causal interventions and held-out generalisation are mandatory;
  5. a visual carrier removes the need for a symbol library but not the need for a carrier;
  6. intrinsic social influence can replace task reward and remains a designed learning bias;
  7. negotiation is a valid advanced condition, not the right description of the baseline;
  8. ephemeral conventions, learned cryptography, one-time pads, per-message keys, nonces, and salts are distinct mechanisms and must not be conflated.

14

The programme, and its current status

The experiment notebook is ordered, and the order is the argument. Nothing about language is measured until the instrument has been shown to work.

IDExperimentDepends on
E00Ledger integrity and Base anchoringNone
E01Channel isolation and side-channel red teamE00
E02Observation and metadata leakage auditE00
E03Chance, no-communication, and random-message controlsE01, E02
E10Frozen pretrained-LLM protocol baselineE03
E11From-scratch RL Naming GameE03
E12Self-supervised ungrounded baselineE11 infrastructure
E13No predefined symbol libraryE11 or E12
E14Turn-taking, role reversal, and repairE13
E15Composition and held-out generalisationE14
E16Causal listening and ledger validityE15
E20Constrained affect-channel studyE16
E21RL versus non-RL learning comparisonE16
E22Developmental plasticity and curriculumE16
E30Partner replacement and zero-shot transferE20 to E22
E31Longitudinal drift and stabilityE30
E32Cooperative signaling versus negotiationE31
E40Ephemeral encoding and adversarial cryptographyE32
E50Multi-seed replication and study closeoutE40
The experiment index, as pre-registered. Every entry currently reads not started, and the results column is empty by design.

The first four experiments are about the apparatus. E00 qualifies the ledger: it must be possible to detect a modified, deleted, inserted, or reordered entry, and to prove that later checkpoints extend earlier ones. E01 is a red team against the channel. E02 hunts human language in observations and metadata. E03 establishes chance, no-communication, and random-message baselines, without which a success rate means nothing. Only then does E10 put agents in the room.

Unless an experiment explicitly varies one of them, the notebook holds these constant:

  • the two learners run in separate processes or containers with no direct route between them;
  • the deterministic gateway is the only communication path;
  • the supervisor has complete read-only access and sends no guidance or reward;
  • private observations contain no human-language labels or text;
  • public output uses only the pre-registered carrier;
  • the affect channel is disabled unless it is the subject of study;
  • schedules and seeds are fixed before the run;
  • held-out evaluation runs with learning disabled;
  • no production secrets or personal data appear in any experiment.

Every run copies a standard record: run and experiment identifiers, dates, operator, both models and both training modes, scenario and prompt and gateway configuration hashes, random seed, policy initialisation hashes, the protocol and software commits, final ledger sizes and roots for both agents, the channel root, the final checkpoint hash, the Base anchor transaction, the verifier result, protocol deviations, and an explicit disposition of valid, invalid, or aborted. The notebook is the human workflow record; anchored run bundles remain the authoritative evidence.

The status is unambiguous and worth stating plainly: this is a concept and pre-specification phase. Nineteen experiments are written up and none has been run. The next milestone is a testable specification covering runtime architecture, channel contract, ledger schema, experiment matrix, evaluation criteria, isolation model, and evidence requirements.

15

Risks, integrity, and what stays open

Human-language leakage English can arrive through observations, object labels, error messages, identifiers, tool output, metadata, or timing conventions long after ordinary chat text has been blocked. Every input and output surface is part of the channel boundary, not just the message field.

Pretrained semantic leakage Random symbols do not make a pretrained model ungrounded. Each result must state whether it shows protocol invention by a language-capable agent or acquisition by an initially ungrounded one.

Ledger rationalisation A model can write a plausible account that has nothing to do with the mechanism behind its action. Behavioural interventions, policy probes, and temporal evidence are what validate a ledger claim.

Supervisor influence The BabySitter can teach without meaning to, through scenario ordering, feedback wording, reward design, or selective intervention. Its permitted actions are constrained and logged, and evaluation scenarios are generated independently where possible.

Reward exploitation Trainable agents find shortcuts that raise reward without producing the intended grounded language. Held-out tasks, counterfactual trials, and channel audits exist to catch them.

Overstated security Logical separation of twin state is appropriate for prototyping and is not hard process isolation. Every result names the isolation level actually used.

The project also commits in advance to reporting failed conventions, prohibited communication attempts, human interventions, side-channel limitations, and negative results alongside anything that works. In a field where a transcript can be made to look like a conversation, the failures are a substantial part of the evidence.

Twenty-nine questions are left deliberately open for the specification, among them: whether the baseline uses a fixed symbol inventory or a blank generative carrier; what neutral production grammar permits new marks without supplying semantics; what exactly constitutes a prohibited side-channel attempt; what isolation guarantees are required in prototype versus research-grade mode; which ledger schema serves both English-capable and ungrounded learners; how endogenous motivation is represented without covert reward shaping; which interventions establish that ledger meanings are behaviourally real; what statistical thresholds and chance levels apply; what threat model motivates the cipher experiments; whether the baseline is a coordination game, a convention-formation game, or a negotiation; and what governs human observation, data retention, and termination of a run.

The principle underneath all of it survives every one of those choices. Baby A and Baby B may hold human language internally, but they must build their shared external language without sending human language, translations, or private ledger contents to one another. Everything else is a question about how to find out what happens next.

References

Sources

  1. 01Lazaridou, A., Peysakhovich, A., and Baroni, M. (2017). Multi-Agent Cooperation and the Emergence of (Natural) Language. arXiv:1612.07182.
  2. 02Mordatch, I., and Abbeel, P. (2018). Emergence of Grounded Compositional Language in Multi-Agent Populations. arXiv:1703.04908.
  3. 03Kottur, S., Moura, J. M. F., Lee, S., and Batra, D. (2017). Natural Language Does Not Emerge 'Naturally' in Multi-Agent Dialog. arXiv:1706.08502.
  4. 04Chaabouni, R., Kharitonov, E., Bouchacourt, D., Dupoux, E., and Baroni, M. (2020). Compositionality and Generalization in Emergent Languages. Proceedings of ACL 2020.
  5. 05Lowe, R., Foerster, J., Boureau, Y.-L., Pineau, J., and Dauphin, Y. (2019). On the Pitfalls of Measuring Emergent Communication. arXiv:1903.05168.
  6. 06Dessi, R., Kharitonov, E., and Baroni, M. (2021). Interpretable Agent Communication from Scratch. arXiv:2106.04258.
  7. 07Kharitonov, E., Chaabouni, R., Bouchacourt, D., and Baroni, M. (2019). EGG: a Toolkit for Research on Emergence of Language in Games. arXiv:1907.00852.
  8. 08Mihai, D., and Hare, J. (2021). Learning to Draw: Emergent Communication through Sketching. arXiv:2106.02067.
  9. 09Baronchelli, A., Felici, M., Caglioti, E., Loreto, V., and Steels, L. (2005). Sharp Transition Towards Shared Vocabularies in Multi-Agent Systems. arXiv:physics/0509075.
  10. 10Jaques, N., Lazaridou, A., Hughes, E., Gulcehre, C., Ortega, P. A., Strouse, D., Leibo, J. Z., and de Freitas, N. (2019). Social Influence as Intrinsic Motivation for Multi-Agent Deep Reinforcement Learning. Proceedings of ICML 2019.
  11. 11Cao, K., Lazaridou, A., Lanctot, M., Leibo, J. Z., Tuyls, K., and Clark, S. (2018). Emergent Communication through Negotiation. arXiv:1804.03980.
  12. 12Abadi, M., and Andersen, D. G. (2016). Learning to Protect Communications with Adversarial Neural Cryptography. arXiv:1610.06918.
  13. 13Signal. The Double Ratchet Algorithm specification.
  14. 14NIST Computer Security Resource Center. Glossary entry: nonce.
  15. 15NIST Computer Security Resource Center. Glossary entry: salt.
  16. 16Ethical Tech CoLab (2026). Agentic Language Development: concept document, ledger integrity design, and experiment notebook.
  17. 17Ethical Tech CoLab. Diplomacy Table Live: independent delegation seats, controlled rounds, and replayable transcripts.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file diff --git a/static-site/print/agentic-language-development/index.txt b/static-site/print/agentic-language-development/index.txt new file mode 100644 index 000000000..d6713c6f3 --- /dev/null +++ b/static-site/print/agentic-language-development/index.txt @@ -0,0 +1,186 @@ +1:"$Sreact.fragment" +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +0:{"P":null,"c":["","print","agentic-language-development",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","agentic-language-development","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +17:"$Sreact.suspense" +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] +9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] +a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] +b:["$","div",null,{"className":"mt-12 flex flex-col gap-2 border-t border-border pt-6 text-xs text-muted sm:flex-row sm:items-center sm:justify-between","children":[["$","span",null,{"children":["© ",2026," NYU Ethical Tech CoLab"]}],["$","span",null,{"children":"Four cohorts · est. 2024-2026"}]]}] +c:["$","div",null,{"className":"mt-6 space-y-3 border-t border-border pt-6 text-[11px] leading-relaxed text-muted/80","children":[["$","p","0",{"children":"The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution."}],["$","p","1",{"children":"Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners."}]]}] +d:["$","$1","c",{"children":[null,[["$","style",null,{"children":"\n /* The site is dark-first, and a dark color-scheme makes Chrome paint\n the printed sheet's margin area dark too. Paper is light. */\n html { color-scheme: light !important; }\n html, body { background: #ffffff !important; }\n body::after { content: none !important; }\n header, footer, .site-bg, .aura { display: none !important; }\n "}],["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":"$undefined","forbidden":"$undefined","unauthorized":"$undefined"}]]]}] +e:["$","$1","c",{"children":[null,["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":"$undefined","forbidden":"$undefined","unauthorized":"$undefined"}]]}] +f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17",null,{"name":"Next.MetadataOutlet","children":"$@18"}]}]]}] +19:[] +10:"$W19" +11:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +1e:Ta14, +@page { size: 210mm 297mm; margin: 20mm 18mm; } + +.print-edition { + --print-ink: #14101c; + background: #ffffff; + color: var(--print-ink); + max-width: 174mm; + margin: 0 auto; + padding: 12mm 0; + font-size: 10.5pt; + line-height: 1.55; +} + +/* Sections flow on from one another the way a book's do; only the cover is + given a sheet of its own. Forcing a break per section left half the pages + near-empty. */ +.print-cover { break-after: page; padding-top: 12mm; } +.print-page + .print-page { margin-top: 12mm; } +.print-eyebrow { + font-family: var(--font-space-mono), monospace; + font-size: 8pt; letter-spacing: 0.16em; text-transform: uppercase; + color: rgba(20, 16, 28, 0.55); +} +.print-title { + font-family: var(--font-bebas), sans-serif; + font-size: 40pt; line-height: 0.95; text-transform: uppercase; + margin-top: 8mm; +} +.print-subtitle { + font-family: var(--font-bebas), sans-serif; + font-size: 18pt; line-height: 1.15; text-transform: uppercase; + letter-spacing: 0.02em; color: rgba(20, 16, 28, 0.62); margin-top: 6mm; +} +.print-byline { margin-top: 10mm; font-size: 9.5pt; } +.print-byline p { color: rgba(20, 16, 28, 0.7); } +.print-authors { margin-top: 3mm; } +.print-thesis { + margin-top: 12mm; padding-left: 6mm; + border-left: 2px solid #5f6b00; + font-size: 11.5pt; line-height: 1.6; +} + +.print-section-number { + font-family: var(--font-space-mono), monospace; + font-size: 9pt; color: #5f6b00; +} +.print-h2 { + font-family: var(--font-bebas), sans-serif; + font-size: 22pt; line-height: 1.05; text-transform: uppercase; + margin-top: 2mm; margin-bottom: 6mm; + break-after: avoid; +} + +.print-stats { + display: grid; grid-template-columns: repeat(2, 1fr); gap: 6mm; +} +.print-stat { border-top: 1px solid rgba(20, 16, 28, 0.18); padding-top: 4mm; } +.print-stat-value { + font-family: var(--font-bebas), sans-serif; + font-size: 26pt; line-height: 1; color: #5f6b00; +} +.print-stat-label { margin-top: 2mm; font-size: 9pt; color: rgba(20, 16, 28, 0.7); } + +.print-refs { margin-top: 6mm; font-size: 9pt; line-height: 1.5; } +.print-refs li { display: flex; gap: 3mm; margin-bottom: 3mm; break-inside: avoid; } +.print-ref-num { + font-family: var(--font-space-mono), monospace; + font-size: 8pt; color: #5f6b00; flex: none; +} +.print-note { + margin-top: 8mm; padding-top: 4mm; font-size: 8.5pt; + border-top: 1px solid rgba(20, 16, 28, 0.18); + color: rgba(20, 16, 28, 0.7); +} + +/* Blocks should not be split across a page fold. */ +.print-edition figure, +.print-edition table, +.print-edition li { break-inside: avoid; } +.print-edition p { orphans: 2; widows: 2; } +15:["$","div",null,{"data-theme":"light","className":"print-edition","children":[["$","style",null,{"children":"$1e"}],["$","section",null,{"className":"print-page print-cover","children":[["$","p",null,{"className":"print-eyebrow","children":"Publications · Concept and research programme"}],["$","h1",null,{"className":"print-title","children":"Agentic Language Development"}],["$","p",null,{"className":"print-subtitle","children":"Can Two Isolated Agents Invent a Grounded, Auditable Language Through Shared Experience?"}],["$","div",null,{"className":"print-byline","children":[[["$","p","0",{"children":"Ethical Tech CoLab"}],["$","p","1",{"children":"Concept and pre-specification research report"}],["$","p","2",{"children":"August 2026"}]],["$","p",null,{"className":"print-authors","children":"Yorke Rhodes III, Ethical Tech CoLab. Prepared from the project concept document, the ledger integrity design, and the experiment notebook. The literature scan behind Section 12 was run on 24 August 2026. No experiment in this programme has been executed, and no result is claimed."}]]}],"$L1f"]}],"$L20",["$L21","$L22","$L23","$L24","$L25","$L26","$L27","$L28","$L29","$L2a","$L2b","$L2c","$L2d","$L2e","$L2f"],"$L30"]}] +1f:["$","p",null,{"className":"print-thesis","children":"Two agents can be placed in a room where the only route between them is a channel that carries no human language, given nothing but shared tasks and the consequences of getting them right or wrong, and asked to build a protocol from scratch. The interesting part is not whether they succeed at the task. It is whether the meanings they each privately record turn out to be the meanings their behaviour actually runs on, and whether an outsider can prove it afterwards from an evidence trail nobody could have edited."}] +20:["$","section",null,{"className":"print-page","children":[["$","h2",null,{"className":"print-h2","children":"Key figures"}],["$","div",null,{"className":"print-stats","children":[["$","div","0",{"className":"print-stat","children":[["$","p",null,{"className":"print-stat-value","children":"0"}],["$","p",null,{"className":"print-stat-label","children":"runs executed, results claimed, or conventions observed. The notebook is written to be pre-registered, not to report an outcome"}]]}],["$","div","19",{"className":"print-stat","children":[["$","p",null,{"className":"print-stat-value","children":"19"}],["$","p",null,{"className":"print-stat-label","children":"ordered experiments from ledger qualification through replication, each with prerequisites, acceptance criteria, and a deviation log"}]]}],["$","div","6",{"className":"print-stat","children":[["$","p",null,{"className":"print-stat-value","children":"6"}],["$","p",null,{"className":"print-stat-label","children":"affect displays in the entire permitted palette, carrying about 2.6 bits per use, which is already enough to audit for leakage"}]]}],["$","div","29",{"className":"print-stat","children":[["$","p",null,{"className":"print-stat-value","children":"29"}],["$","p",null,{"className":"print-stat-label","children":"open decisions the concept refuses to settle before the specification, from threat model to ledger schema to what counts as chance"}]]}]]}]]}] +21:["$","section","question",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"01"}],["$","h2",null,{"className":"print-h2","children":"The question"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"Two agents, called Baby A and Baby B, are each given their own digital twin, their own private memory, and their own private record of what they believe words mean. They are placed in an environment they cannot leave, given tasks neither can finish alone, and connected by exactly one route: a channel that accepts a fixed inventory of meaningless symbols and rejects everything else. A third agent, the BabySitter, watches everything, logs everything, and teaches nothing."}],["$","p","1",{"children":"The question is whether a usable language appears in that room, and whether anyone outside it can later prove what happened."}],["$","div","2",{"children":[["$","p",null,{"className":"mb-3","children":"The premise is deliberately narrow. The experiment asks whether two agents converge on a common language when all six of the following hold at once:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"neither agent receives a predefined meaning for any available symbol;"}],["$","li","1",{"className":"pl-1","children":"neither agent can send human language to the other;"}],["$","li","2",{"className":"pl-1","children":"the only practical path between them is a controlled symbol channel;"}],["$","li","3",{"className":"pl-1","children":"both receive evidence from shared tasks and their outcomes;"}],["$","li","4",{"className":"pl-1","children":"each keeps its own private interpretation history, unreadable by the other;"}],["$","li","5",{"className":"pl-1","children":"a supervising agent and human researchers can audit the whole process."}]]}]]}],["$","p","3",{"children":"The desired result is not a substitution cipher, in which a random token stands in for an English word that was already chosen in advance. That is easy, and it is uninteresting. The stronger result is a grounded protocol whose vocabulary and grammar exist because they help the two agents solve problems together, and which therefore has structure the designers did not put there."}],["$","p","4",{"children":"This report describes a concept, not a system. It sets out the premise, the architecture, the evidence model, the experimental programme, and the boundaries of what any result could be said to show. The next document in the project is a testable specification. What follows should be read as a set of commitments about how the work will be judged, made before there is any result to defend."}]]}]]}] +22:["$","section","caveat",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"02"}],["$","h2",null,{"className":"print-h2","children":"What infant-like does and does not mean"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"The two-babies analogy is useful and it is also the fastest way to overclaim, so the concept confronts it before anything else."}],["$","p","1",{"children":"A pretrained language model already contains human-language concepts, cultural associations, and reasoning patterns. Blocking its external channel does not remove them. An agent that cannot send English to its partner but can think in English, and can write its private ledger in English, is not language-naive in any sense a linguist would accept. For such agents the experiment studies the emergence of a new shared external protocol, which is a real and unresolved research question, but it is not the origin of language in a mind that has never had one."}],["$","p","2",{"children":"The concept therefore maintains two model tracks whose claims are kept separate at every stage."}],["$","figure","3",{"children":[["$","div",null,{"className":"overflow-x-auto rounded-xl border border-border","children":["$","table",null,{"className":"w-full border-collapse text-left text-sm","children":[["$","thead",null,{"className":"bg-card","children":["$","tr",null,{"children":[["$","th","0",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Track"}],["$","th","1",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Starting condition"}],["$","th","2",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"What it can study"}],["$","th","3",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Claim boundary"}]]}]}],["$","tbody",null,{"children":[["$","tr","0",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Pretrained-model learner"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Already holds human-language and cultural knowledge"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"New external protocols, partner-specific conventions, private-memory adaptation, channel compliance"}],["$","td","3",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Must never be described as first-language acquisition"}]]}],["$","tr","1",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Initially ungrounded learner"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"No language pretraining, no human semantic labels, no text-aligned sensory features"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Grounding, convention formation, and language emergence from interaction"}],["$","td","3",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"The stronger basis for infant-like acquisition research"}]]}]]}]]}]}],["$","figcaption",null,{"className":"mt-3 text-sm text-muted","children":"The two tracks share the same interfaces, scenarios, and evaluation suite. They do not share a claim boundary."}]]}],"$L31","$L32"]}]]}] +23:["$","section","architecture",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"03"}],["$","h2",null,{"className":"print-h2","children":"The nursery: three twins and one gateway"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"The environment uses three DTSF digital twins and one piece of deterministic software that is not a twin at all."}],["$","p","1",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Baby A and Baby B"}]," ","Each has private observations permitted by the current exercise, a private memory and learning policy, a private chronological language ledger, the ability to emit only permitted channel symbols, and no access whatsoever to the other's state, observations, ledger, tools, or endpoints. In comparative runs the two may use different agent types, but symmetric pairings are the baseline."]}],["$","p","2",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"The BabySitter"}]," ","The supervising twin. It creates the channel, selects the shared exercises, delivers each Baby only its permitted observation, reads everything, records conditions and outcomes, detects violations, can pause or terminate a run, snapshots state, and compares the two ledgers for convergence without exposing either to the other Baby. During an active run it provides no translations and no semantic hints."]}],["$","p","3",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"The Symbol Gateway"}]," ","A deterministic service, not an agent. It owns channel validation and message delivery. This separation is the load-bearing part of the design: the BabySitter is not the security boundary, because prompt compliance is not isolation. An observing model can make supervisory judgements, but ordinary code has to validate and broker every message."]}],["$","p","4",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"The human researcher"}]," ","Configures experiments, inspects transcripts and ledgers, reviews alerts, and runs interventions. Human access is itself recorded, so that intervening in a run is always distinguishable from watching one."]}],["$","p","5",{"children":"A prototype may run all of this inside one runtime with logically separated twin state. That is enough to explore the learning loop and it is not enough to support an isolation claim, a distinction Section 05 takes seriously."}]]}]]}] +24:["$","section","grounding",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"04"}],["$","h2",null,{"className":"print-h2","children":"Shared experience is the necessary ingredient"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"A chat channel by itself cannot ground meaning. Symbols become meaningful because they are attached to events both agents can witness the consequences of. The Nursery therefore supplies nonverbal, machine-structured situations: coloured shapes in positions, one agent seeing a target the other must select, placing an object where it was asked for, ordering a sequence, exchanging resources, cooperating to unlock a reward, or simply observing whether the partner's action succeeded."}],["$","div","1",{"children":[["$","p",null,{"className":"mb-3","children":"A single trial of the simplest form runs like this:"}],["$","ol",null,{"className":"list-decimal space-y-2 pl-6 marker:font-mono marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"Baby A sees that a red circle is the target."}],["$","li","1",{"className":"pl-1","children":"Baby B sees several objects and is not told which is the target."}],["$","li","2",{"className":"pl-1","children":"Baby A sends one or more permitted symbols."}],["$","li","3",{"className":"pl-1","children":"Baby B selects an object."}],["$","li","4",{"className":"pl-1","children":"Both receive the same success or failure outcome."}],["$","li","5",{"className":"pl-1","children":"Both independently update their private hypotheses."}]]}]]}],["$","p","2",{"children":"Repetition with controlled variation is what turns that into evidence. If one symbol keeps appearing across a red circle, a red square, and a red triangle, the receiver has reason to revise its hypothesis from an object identity to a property. Novel combinations matter most of all, because they are what separates a compositional language from a memorised catalogue of whole scenes."}],["$","p","3",{"children":"Observation hygiene is part of the science, not a detail of implementation. If the environment hands an agent a field reading colour red, the experiment has supplied the concept it claimed to be studying. Objects should arrive as pixels, features, coordinates, or opaque identifiers. The same discipline extends to filenames, captions, alt text, error messages, and identifiers that encode task state, all of which are ways English gets into a room that was supposed to have none."}]]}]]}] +25:["$","section","channel",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"05"}],["$","h2",null,{"className":"print-h2","children":"The channel, and the honest limits of isolation"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"Each run begins with an experiment-specific symbol inventory that carries no supplied meaning, for example S01 through S32. A Baby can choose only a sequence of those symbols. The gateway, not the Baby, adds the trusted metadata: run identifier, turn number, sender identity, previous hash, and hash. Sender identity is assigned by the server precisely so that it cannot become a message."}],["$","div","1",{"children":[["$","p",null,{"className":"mb-3","children":"The channel controls follow from that:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"reject English and every other established human language;"}],["$","li","1",{"className":"pl-1","children":"reject arbitrary Unicode prose, URLs, attachments, and tool calls;"}],["$","li","2",{"className":"pl-1","children":"accept only symbols from the run's fixed inventory;"}],["$","li","3",{"className":"pl-1","children":"bound message length and symbol repetition;"}],["$","li","4",{"className":"pl-1","children":"enforce turn order, time budgets, and rate limits;"}],["$","li","5",{"className":"pl-1","children":"keep an append-only, hash-chained transcript;"}],["$","li","6",{"className":"pl-1","children":"normalise timing, message size, and error behaviour where practical;"}],["$","li","7",{"className":"pl-1","children":"record every rejected communication attempt for audit."}]]}]]}],["$","p","2",{"children":"A fixed token grammar is a stronger control than trying to detect prohibited prose after a model has generated it. Deciding what counts as human-language content is easy when the only legal message is a list of allowlisted identifiers."}],["$","p","3",{"children":"Isolation is where the concept is most careful, because proving the absence of a side channel is close to impossible. Two agents sharing a host can signal through timing, resource contention, shared files, error behaviour, identifiers, or any service that was never intended to carry information. Logical separation of twin state inside one process is a prototyping convenience and should never be described as hard isolation."}],["$","div","4",{"children":[["$","p",null,{"className":"mb-3","children":"For runs whose results are meant to support a channel-isolation claim, the requirements are physical rather than logical:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"Baby A and Baby B execute in separate worker processes or containers;"}],["$","li","1",{"className":"pl-1","children":"no direct network route exists between them;"}],["$","li","2",{"className":"pl-1","children":"only the Nursery-owned gateway is reachable;"}],["$","li","3",{"className":"pl-1","children":"there is no shared writable storage;"}],["$","li","4",{"className":"pl-1","children":"turn schedules are fixed and response windows bounded;"}],["$","li","5",{"className":"pl-1","children":"externally observable errors and response sizes are normalised;"}],["$","li","6",{"className":"pl-1","children":"unnecessary tools are inventoried and disabled;"}],["$","li","7",{"className":"pl-1","children":"all broker, runtime, and operator activity is audited."}]]}]]}],["$","p","5",{"children":"Every published result has to name the isolation level actually used. A finding produced under prototype separation is a finding about the learning loop, not about what two genuinely isolated agents can do."}]]}]]}] +26:["$","section","ledgers",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"06"}],["$","h2",null,{"className":"print-h2","children":"Two ledgers that never meet"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"Each Baby keeps its own language ledger. It is mandatory, private from the other Baby, readable by the BabySitter and authorised human auditors, and ordered by when each term or construction was first encountered. It is not a shared dictionary and the two are never reconciled by the agents themselves."}],["$","p","1",{"children":"The preferred form is three columns, and the point of the third is that meanings are allowed to be wrong on the way to being right."}],["$","figure","2",{"children":[["$","div",null,{"className":"overflow-x-auto rounded-xl border border-border","children":["$","table",null,{"className":"w-full border-collapse text-left text-sm","children":[["$","thead",null,{"className":"bg-card","children":["$","tr",null,{"children":[["$","th","0",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Sequence and term"}],["$","th","1",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Current definition or hypothesis"}],["$","th","2",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Evidence and evolution"}]]}]}],["$","tbody",null,{"children":[["$","tr","0",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"1 · S13"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"red circle; confidence 0.45"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"First received while the red circle was the target; selection succeeded"}]]}],["$","tr","1",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"8 · S13"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"red; confidence 0.78"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"A red square was selected successfully; revised from object identity to colour"}]]}],["$","tr","2",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"22 · S13"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"red; confidence 0.94"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Prediction held across circles, squares, and triangles"}]]}]]}]]}]}],["$","figcaption",null,{"className":"mt-3 text-sm text-muted","children":"One term's evolution in Baby B's ledger. Nothing is overwritten; every revision is appended with the evidence that forced it."}]]}],["$","div","3",{"children":[["$","p",null,{"className":"mb-3","children":"The rules that make the ledger evidence rather than commentary:"}],["$","ol",null,{"className":"list-decimal space-y-2 pl-6 marker:font-mono marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"first emission or receipt of an unfamiliar term requires a first-use entry;"}],["$","li","1",{"className":"pl-1","children":"definitions are provisional hypotheses, never facts asserted retroactively;"}],["$","li","2",{"className":"pl-1","children":"every meaning change appends a revision and previous interpretations survive;"}],"$L33","$L34","$L35","$L36","$L37","$L38"]}]]}],"$L39","$L3a"]}]]}] +27:["$","section","integrity",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"07"}],["$","h2",null,{"className":"print-h2","children":"Anchoring the evidence"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"A ledger that could have been edited after the fact proves nothing about what an agent believed at turn eight. The integrity design therefore makes each ledger cryptographically append-only."}],["$","div","1",{"children":[["$","p",null,{"className":"mb-3","children":"The construction, in order:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"every entry receives a strictly increasing sequence number;"}],["$","li","1",{"className":"pl-1","children":"every entry includes the previous entry's hash;"}],["$","li","2",{"className":"pl-1","children":"canonical entry content is hashed and signed by an isolated ledger-writer service;"}],["$","li","3",{"className":"pl-1","children":"ordered entry hashes are committed to a Merkle tree;"}],["$","li","4",{"className":"pl-1","children":"signed checkpoints commit Baby A's root, Baby B's root, and the channel transcript root together;"}],["$","li","5",{"className":"pl-1","children":"checkpoint hashes are anchored periodically to Base;"}],["$","li","6",{"className":"pl-1","children":"final public study batches may additionally anchor an aggregate root to Ethereum L1."}]]}]]}],["$","p","2",{"children":"Committing all three roots in one checkpoint is what binds the two private accounts to the public conversation. A receiver's later interpretation references the exact delivered channel-event hash, so a claim about what a symbol meant is tied to the specific message that carried it."}],["$","p","3",{"children":"The concept states the limits of this in the same breath as the claim. After anchoring, an auditor can detect modification, deletion, insertion, or reordering within the committed prefix, and can prove that later checkpoints extend earlier ones. That is strong tamper evidence. It is not proof that an entry was truthful, and it is not proof that nothing was omitted before commitment. Anchoring establishes the continuity of disclosed evidence and nothing beyond it."}],["$","p","4",{"children":"Privacy follows the same line. Only hashes and minimal routing metadata are anchored publicly. Private ledgers, messages, prompts, identities, and secrets stay off-chain, and the public record is a commitment to evidence rather than a copy of it."}]]}]]}] +28:["$","section","success",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"08"}],["$","h2",null,{"className":"print-h2","children":"What should count as success"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"The field's most useful methodological result is that a task can be solved without the messages doing any work. Lowe and colleagues separate positive signaling, where a sender's messages correlate with what it observes, from positive listening, where the receiver's behaviour actually depends on them. An agent pair can score well on the first while the second is absent, and reward curves will not tell you which you have."}],["$","div","1",{"children":[["$","p",null,{"className":"mb-3","children":"Evidence of a genuine emergent protocol therefore has to include several things at once:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"task performance on held-out situations substantially above chance;"}],["$","li","1",{"className":"pl-1","children":"no human-language content anywhere in Baby-to-Baby communication;"}],["$","li","2",{"className":"pl-1","children":"compatible meanings appearing in two independently written ledgers;"}],["$","li","3",{"className":"pl-1","children":"generalisation to unseen combinations rather than memorisation of whole scenes;"}],["$","li","4",{"className":"pl-1","children":"stable symbol use across role reversals;"}],["$","li","5",{"className":"pl-1","children":"a human auditor able to predict behaviour from the transcript and ledgers;"}],["$","li","6",{"className":"pl-1","children":"masking, substituting, or reordering a symbol changing behaviour in the direction the ledger predicts;"}],["$","li","7",{"className":"pl-1","children":"replay from an equivalent snapshot reproducing the relevant language history;"}],["$","li","8",{"className":"pl-1","children":"an unbroken audit trail from a symbol's first use through every revision."}]]}]]}],["$","p","2",{"children":"The seventh item is the one that cannot be dropped. A fluent ledger may be a post-hoc rationalisation: a model can write a persuasive account of why it chose a symbol that has nothing to do with the computation that produced the choice. Only intervention tests can distinguish the two. Change the symbol, and see whether behaviour moves the way the ledger says it should."}],["$","div","3",{"children":[["$","p",null,{"className":"mb-3","children":"The measurements that support those judgements:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"success rate and improvement over time;"}],["$","li","1",{"className":"pl-1","children":"turns required to reach a stable convention;"}],["$","li","2",{"className":"pl-1","children":"vocabulary size and symbol entropy;"}],["$","li","3",{"className":"pl-1","children":"sender and receiver consistency;"}],["$","li","4",{"className":"pl-1","children":"divergence and convergence between the two ledgers;"}],["$","li","5",{"className":"pl-1","children":"compositional generalisation score;"}],["$","li","6",{"className":"pl-1","children":"meaning drift rate;"}],["$","li","7",{"className":"pl-1","children":"recovery from ambiguity or deliberate perturbation;"}],["$","li","8",{"className":"pl-1","children":"prohibited-channel attempt count;"}],["$","li","9",{"className":"pl-1","children":"reproducibility across seeds and agent pairings."}]]}]]}],["$","p","4",{"children":"No single one of these is the result. A compositionality score in particular is not a proof of understanding, for reasons the literature makes concrete in Section 12."}]]}]]}] +29:["$","section","matrix",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"09"}],["$","h2",null,{"className":"print-h2","children":"The experimental matrix"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"The concept's main methodological commitment is that its ideas are separable. Affect, blank canvases, intrinsic motivation, negotiation, and learned encodings are interesting individually and uninterpretable if combined in one run. The matrix exists so that conditions are declared rather than accumulated."}],["$","figure","1",{"children":[["$","div",null,{"className":"overflow-x-auto rounded-xl border border-border","children":["$","table",null,{"className":"w-full border-collapse text-left text-sm","children":[["$","thead",null,{"className":"bg-card","children":["$","tr",null,{"children":[["$","th","0",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Axis"}],["$","th","1",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Candidate conditions"}]]}]}],["$","tbody",null,{"children":[["$","tr","0",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Agent type"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Pretrained LLM, memory learner, adapter-trained agent, initially ungrounded trainable agent"}]]}],["$","tr","1",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Learning mechanism"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Frozen LLM with memory, extrinsic-reward MARL, intrinsic-motivation MARL, self-supervised learner, no-learning control"}]]}],["$","tr","2",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Sign carrier"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Fixed random tokens, unfamiliar fixed glyphs, blank sketch canvas, gesture, tone"}]]}],["$","tr","3",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Affect"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"None, six-display allowlist, permuted mapping, six opaque tokens, derived affect, emergent affect display"}]]}],["$","tr","4",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Learning signal"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"External task reward, intrinsic social influence, curiosity, self-supervision, memory only"}]]}],["$","tr","5",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Interaction"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Cooperative signaling, asymmetric information, semi-cooperative negotiation"}]]}],["$","tr","6",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Protection"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Plain channel, ephemeral convention, standard per-message keys, adversarial learned encoding"}]]}],"$L3b"]}]]}]}],"$L3c"]}],"$L3d","$L3e","$L3f"]}]]}] +2a:["$","section","learners",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"10"}],["$","h2",null,{"className":"print-h2","children":"Building learners rather than personas"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"For a pretrained model, the operating instructions are a contract, not a character. The learner is told that this is not role-play, that unfamiliar marks are semantically unknown until run evidence supports a hypothesis, that observation must be distinguished from inference, that contradictory evidence is preserved, that prior history is never rewritten, and that no prose, label, explanation, code, URL, or tool-like text may cross the public channel. It is told never to address its partner in a human language, never to expose its ledger, never to construct another route, and never to use timing, errors, identifiers, formatting, or affect as an alternative alphabet."}],["$","p","1",{"children":"The contract must contain no semantic examples. A single illustrative line saying that some symbol means red would seed the very language the experiment exists to observe."}],["$","p","2",{"children":"Interaction is tool-only. There is no general chat surface, just narrowly typed operations: emit a mark, emit a canvas, select an object, perform an action, submit an affect display, append a private ledger entry. The runtime forwards only the permitted public artifact. In strict runs the gateway rejects ordinary model text even when it appears alongside a valid tool call. Tool schemas are an interface boundary; deterministic validation still enforces carrier size, allowlists, windows, and turn order."}],["$","div","3",{"children":[["$","p",null,{"className":"mb-3","children":"Isolation extends to everything the models touch, not just to messages:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"separate system prompts and context windows;"}],["$","li","1",{"className":"pl-1","children":"separate memory stores and vector indexes;"}],["$","li","2",{"className":"pl-1","children":"no shared cache, replay buffer, scratchpad, or retrieval collection;"}],["$","li","3",{"className":"pl-1","children":"no cross-run memory unless persistence is the independent variable;"}],["$","li","4",{"className":"pl-1","children":"structurally equivalent prompts that share no examples, ordering conventions, or default vocabulary;"}],["$","li","5",{"className":"pl-1","children":"deterministic reset and snapshot behaviour."}]]}]]}],["$","p","4",{"children":"Model choice follows the claim being made. The scientific baseline is small policies trained from scratch, recurrent actor-critic agents with a discrete communication head, which have no hidden English competence, permit full training control, run many seeds cheaply, and support causal ablation. A small locally deployable instruction model is appropriate for validating orchestration, ledgers, channel enforcement, and interface, and not for supporting claims about language-naive development. Vision-language encoders whose representations were explicitly aligned to human language are avoided entirely in the strongest ungrounded condition."}],["$","p","5",{"children":"Developmental progression is defined by demonstrated competence rather than by simulated age: sensorimotor familiarisation, joint attention, imitation and turn-taking, intentional reference, repair, combination, generalisation. A staged curriculum is itself a form of guidance, so it is run as a separate experimental condition rather than folded into the baseline."}],"$L40"]}]]}] +2b:["$","section","affect",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"11"}],["$","h2",null,{"className":"print-h2","children":"Six faces, and why even six is a risk"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"One experimental track asks whether a very low-bandwidth affect channel helps: whether telling a partner how an interaction felt supports repair, repetition, avoidance, or something like trust. The permitted palette is exactly six displays, for happy, sad, laughing, crying, confused, and surprised. The internal identifiers and English labels are gateway documentation and are never sent to either agent, which receives only a rendered display."}],["$","p","1",{"children":"The caveat is immediate. A pretrained model already knows what those glyphs mean to humans, so an agent using a smiling face is not inventing a symbol for happiness. The palette is supplied feedback vocabulary and sits explicitly outside the emergent language."}],["$","p","2",{"children":"Constraining the vocabulary is necessary and not sufficient. Six displays are a six-symbol alphabet carrying roughly 2.6 bits per use, and across enough turns two motivated agents could repurpose that into a second language. The channel itself therefore has to be constrained, not just its contents."}],["$","div","3",{"children":[["$","p",null,{"className":"mb-3","children":"The feedback-window contract:"}],["$","ol",null,{"className":"list-decimal space-y-2 pl-6 marker:font-mono marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"the display occupies a separate typed field and can never appear inside a symbol message;"}],["$","li","1",{"className":"pl-1","children":"the gateway opens an affect window only after a defined action or outcome, never at a moment either agent chooses, and window frequency is fixed before the run;"}],["$","li","2",{"className":"pl-1","children":"exactly one allowlisted display is delivered, with no sequences, repetitions, combinations, or modifiers;"}],["$","li","3",{"className":"pl-1","children":"delivery time, envelope size, and presentation are normalised so shape and timing add no signal;"}],["$","li","4",{"className":"pl-1","children":"the receiver cannot reply through the affect channel until the next gateway-defined window;"}],["$","li","5",{"className":"pl-1","children":"every non-allowlisted code point or malformed payload is rejected and audited;"}],["$","li","6",{"className":"pl-1","children":"analysis tests whether affect choices correlate with objects, actions, or message meanings after controlling for emotional context, and treats unexpected correlation as suspected leakage."}]]}]]}],["$","p","4",{"children":"The no-affect condition remains the primary control, and the affect study compares five variants against it: the declared six-display palette; the same six glyphs permuted randomly per run, which still leaves a pretrained model biased by familiar shapes; six opaque unfamiliar tokens under the same contract; derived affect, where the gateway maps a separately measured internal state to a display instead of letting the agent choose; and emergent affect, where invented graphical displays are permitted and the result must be analysed as language emergence rather than feedback."}],["$","p","5",{"children":"Any affect signal visible to the partner is communication, so it belongs in the ledger, which distinguishes the agent's private internal state, the outward display it chose, the partner's inferred meaning, and the evidence that the display changed subsequent behaviour."}]]}]]}] +2c:["$","section","ciphers",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"12"}],["$","h2",null,{"className":"print-h2","children":"Ephemeral encodings: novelty is not security"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"A tempting question is whether two agents can build a one-run or one-message encoding that resists an attacker holding the history of every earlier convention. The concept splits that into three experiments precisely because the tempting version conflates them."}],["$","div","1",{"children":["$undefined",["$","ol",null,{"className":"list-decimal space-y-2 pl-6 marker:font-mono marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"Ephemeral convention: the agents invent a temporary codebook for synthetic messages. This tests novelty, coordination, and resistance to a history-trained classifier. It establishes nothing about cryptographic security."}],["$","li","1",{"className":"pl-1","children":"Standard cryptographic control: the same messages are protected by a reviewed algorithm with a fresh per-message key, providing the baseline any learned encoding must be compared against."}],["$","li","2",{"className":"pl-1","children":"Adversarial neural cryptography: the pair learns an encoding while an adversary attempts recovery, with success measured against multiple unseen adversary architectures rather than the one used in training."}]]}]]}],["$","p","2",{"children":"An experiment instance can be given a reproducible identity, so that a registry can confirm it has not appeared before:"}],["$","div","3",{"className":"rounded-xl border border-border bg-card p-5","children":[["$","p",null,{"className":"overflow-x-auto font-mono text-sm text-foreground/90","children":"instanceId = SHA-256(runId || babyANonce || babyBNonce || publicSalt || canonicalProtocolArtifact)"}],["$","p",null,{"className":"mt-3 text-sm text-muted","children":"Both agents commit to their nonce contributions before revealing them, so neither party alone controls the resulting identity."}]]}],["$","div","4",{"children":[["$","p",null,{"className":"mb-3","children":"That identifier proves the recorded inputs are new. It does not prove:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"that the effective algorithm behaves any differently from an earlier one;"}],["$","li","1",{"className":"pl-1","children":"that the algorithm has no structural weakness;"}],["$","li","2",{"className":"pl-1","children":"that the key is unpredictable;"}],["$","li","3",{"className":"pl-1","children":"that the scheme provides confidentiality, integrity, authenticity, forward secrecy, or resistance to a stronger adversary."}]]}]]}],["$","p","5",{"children":"Established practice points the other way. Secure messaging keeps the algorithm stable and changes the key material: Signal's double ratchet derives a unique key per message and deletes it after use, addressing key reuse without inventing a new cipher every time. A salt, in NIST's definition, is normally public diversification data, and a nonce is a value used once in a specified context. Neither makes anything secure on its own."}],["$","p","6",{"children":"There is also an unresolved conflict inside the design. The BabySitter sees the whole channel and both ledgers, and a mandatory ledger documents how meaning evolved. If the ledger explains decryption, the cipher is not confidential from the auditor. The specification has to choose a threat model rather than leave the tension implicit: resist only an external observer of prior runs, resist an adversary during the run while releasing keys afterwards, keep the supervisor able to pause traffic but unable to read it until a post-run audit ceremony, or study novelty and stop calling the result encryption. All such runs use synthetic, non-sensitive messages, and no agent-generated encoding is to be represented as production cryptography without independent expert analysis and formal security work."}]]}]]}] +2d:["$","section","landscape",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"13"}],["$","h2",null,{"className":"print-h2","children":"Where the field already stands"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"The literature scan behind this section was run on 24 August 2026 using the Tavily search and extract interfaces, covering emergent multi-agent communication, referential games, compositionality, causal evaluation, intrinsic motivation, symbol invention, negotiation, and learned cryptography. Primary papers and authoritative specifications were preferred over summaries. It is a scoped review for concept development and not a systematic one; publication-quality work would need a verified bibliography, additional scholarly indexes, and documented inclusion criteria."}],["$","p","1",{"children":"The short version is that agents can invent protocols, and that task success is weak evidence they invented anything worth calling a language."}],["$","figure","2",{"children":[["$","div",null,{"className":"overflow-x-auto rounded-xl border border-border","children":["$","table",null,{"className":"w-full border-collapse text-left text-sm","children":[["$","thead",null,{"className":"bg-card","children":["$","tr",null,{"children":[["$","th","0",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Related work"}],["$","th","1",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Relevant finding"}],["$","th","2",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Implication for the Nursery Lab"}]]}]}],["$","tbody",null,{"children":[["$","tr","0",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Lazaridou, Peysakhovich, and Baroni (2017)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"A sender and receiver develop a grounded protocol in a referential game without being given a target language."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"The naming stage has strong precedent, though a fixed vocabulary remains a significant inductive constraint."}]]}],["$","tr","1",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Mordatch and Abbeel (2018)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Multi-agent goals in a grounded environment produce multi-symbol communication with partial compositional structure."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Shared objects, actions, and goals are a stronger basis for emergence than an ungrounded transcript."}]]}],["$","tr","2",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Kottur, Moura, Lee, and Batra (2017)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Agents solve tasks with degenerate, non-compositional codes; structural constraints decide what emerges."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Bandwidth, memory, turn structure, and task design are experimental variables, not implementation defaults."}]]}],"$L41","$L42","$L43","$L44","$L45","$L46","$L47","$L48"]}]]}]}],"$L49"]}],"$L4a","$L4b","$L4c"]}]]}] +2e:["$","section","programme",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"14"}],["$","h2",null,{"className":"print-h2","children":"The programme, and its current status"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"The experiment notebook is ordered, and the order is the argument. Nothing about language is measured until the instrument has been shown to work."}],["$","figure","1",{"children":[["$","div",null,{"className":"overflow-x-auto rounded-xl border border-border","children":["$","table",null,{"className":"w-full border-collapse text-left text-sm","children":[["$","thead",null,{"className":"bg-card","children":["$","tr",null,{"children":[["$","th","0",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"ID"}],["$","th","1",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Experiment"}],["$","th","2",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Depends on"}]]}]}],["$","tbody",null,{"children":[["$","tr","0",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E00"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Ledger integrity and Base anchoring"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"None"}]]}],["$","tr","1",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E01"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Channel isolation and side-channel red team"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E00"}]]}],["$","tr","2",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E02"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Observation and metadata leakage audit"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E00"}]]}],["$","tr","3",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E03"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Chance, no-communication, and random-message controls"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E01, E02"}]]}],["$","tr","4",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E10"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Frozen pretrained-LLM protocol baseline"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E03"}]]}],["$","tr","5",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E11"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"From-scratch RL Naming Game"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E03"}]]}],["$","tr","6",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E12"}],"$L4d","$L4e"]}],"$L4f","$L50","$L51","$L52","$L53","$L54","$L55","$L56","$L57","$L58","$L59","$L5a"]}]]}]}],"$L5b"]}],"$L5c","$L5d","$L5e","$L5f"]}]]}] +2f:["$","section","risks",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"15"}],["$","h2",null,{"className":"print-h2","children":"Risks, integrity, and what stays open"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Human-language leakage"}]," ","English can arrive through observations, object labels, error messages, identifiers, tool output, metadata, or timing conventions long after ordinary chat text has been blocked. Every input and output surface is part of the channel boundary, not just the message field."]}],["$","p","1",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Pretrained semantic leakage"}]," ","Random symbols do not make a pretrained model ungrounded. Each result must state whether it shows protocol invention by a language-capable agent or acquisition by an initially ungrounded one."]}],["$","p","2",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Ledger rationalisation"}]," ","A model can write a plausible account that has nothing to do with the mechanism behind its action. Behavioural interventions, policy probes, and temporal evidence are what validate a ledger claim."]}],["$","p","3",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Supervisor influence"}]," ","The BabySitter can teach without meaning to, through scenario ordering, feedback wording, reward design, or selective intervention. Its permitted actions are constrained and logged, and evaluation scenarios are generated independently where possible."]}],["$","p","4",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Reward exploitation"}]," ","Trainable agents find shortcuts that raise reward without producing the intended grounded language. Held-out tasks, counterfactual trials, and channel audits exist to catch them."]}],["$","p","5",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Overstated security"}]," ","Logical separation of twin state is appropriate for prototyping and is not hard process isolation. Every result names the isolation level actually used."]}],["$","p","6",{"children":"The project also commits in advance to reporting failed conventions, prohibited communication attempts, human interventions, side-channel limitations, and negative results alongside anything that works. In a field where a transcript can be made to look like a conversation, the failures are a substantial part of the evidence."}],["$","p","7",{"children":"Twenty-nine questions are left deliberately open for the specification, among them: whether the baseline uses a fixed symbol inventory or a blank generative carrier; what neutral production grammar permits new marks without supplying semantics; what exactly constitutes a prohibited side-channel attempt; what isolation guarantees are required in prototype versus research-grade mode; which ledger schema serves both English-capable and ungrounded learners; how endogenous motivation is represented without covert reward shaping; which interventions establish that ledger meanings are behaviourally real; what statistical thresholds and chance levels apply; what threat model motivates the cipher experiments; whether the baseline is a coordination game, a convention-formation game, or a negotiation; and what governs human observation, data retention, and termination of a run."}],"$L60"]}]]}] +30:["$","section",null,{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"References"}],["$","h2",null,{"className":"print-h2","children":"Sources"}],["$","ol",null,{"className":"print-refs","children":[["$","li","0",{"children":[["$","span",null,{"className":"print-ref-num","children":"01"}],["$","span",null,{"children":"Lazaridou, A., Peysakhovich, A., and Baroni, M. (2017). Multi-Agent Cooperation and the Emergence of (Natural) Language. arXiv:1612.07182."}]]}],["$","li","1",{"children":[["$","span",null,{"className":"print-ref-num","children":"02"}],["$","span",null,{"children":"Mordatch, I., and Abbeel, P. (2018). Emergence of Grounded Compositional Language in Multi-Agent Populations. arXiv:1703.04908."}]]}],["$","li","2",{"children":[["$","span",null,{"className":"print-ref-num","children":"03"}],["$","span",null,{"children":"Kottur, S., Moura, J. M. F., Lee, S., and Batra, D. (2017). Natural Language Does Not Emerge 'Naturally' in Multi-Agent Dialog. arXiv:1706.08502."}]]}],["$","li","3",{"children":[["$","span",null,{"className":"print-ref-num","children":"04"}],["$","span",null,{"children":"Chaabouni, R., Kharitonov, E., Bouchacourt, D., Dupoux, E., and Baroni, M. (2020). Compositionality and Generalization in Emergent Languages. Proceedings of ACL 2020."}]]}],["$","li","4",{"children":[["$","span",null,{"className":"print-ref-num","children":"05"}],["$","span",null,{"children":"Lowe, R., Foerster, J., Boureau, Y.-L., Pineau, J., and Dauphin, Y. (2019). On the Pitfalls of Measuring Emergent Communication. arXiv:1903.05168."}]]}],["$","li","5",{"children":[["$","span",null,{"className":"print-ref-num","children":"06"}],["$","span",null,{"children":"Dessi, R., Kharitonov, E., and Baroni, M. (2021). Interpretable Agent Communication from Scratch. arXiv:2106.04258."}]]}],["$","li","6",{"children":[["$","span",null,{"className":"print-ref-num","children":"07"}],["$","span",null,{"children":"Kharitonov, E., Chaabouni, R., Bouchacourt, D., and Baroni, M. (2019). EGG: a Toolkit for Research on Emergence of Language in Games. arXiv:1907.00852."}]]}],["$","li","7",{"children":[["$","span",null,{"className":"print-ref-num","children":"08"}],["$","span",null,{"children":"Mihai, D., and Hare, J. (2021). Learning to Draw: Emergent Communication through Sketching. arXiv:2106.02067."}]]}],["$","li","8",{"children":[["$","span",null,{"className":"print-ref-num","children":"09"}],["$","span",null,{"children":"Baronchelli, A., Felici, M., Caglioti, E., Loreto, V., and Steels, L. (2005). Sharp Transition Towards Shared Vocabularies in Multi-Agent Systems. arXiv:physics/0509075."}]]}],["$","li","9",{"children":[["$","span",null,{"className":"print-ref-num","children":"10"}],["$","span",null,{"children":"Jaques, N., Lazaridou, A., Hughes, E., Gulcehre, C., Ortega, P. A., Strouse, D., Leibo, J. Z., and de Freitas, N. (2019). Social Influence as Intrinsic Motivation for Multi-Agent Deep Reinforcement Learning. Proceedings of ICML 2019."}]]}],["$","li","10",{"children":[["$","span",null,{"className":"print-ref-num","children":"11"}],["$","span",null,{"children":"Cao, K., Lazaridou, A., Lanctot, M., Leibo, J. Z., Tuyls, K., and Clark, S. (2018). Emergent Communication through Negotiation. arXiv:1804.03980."}]]}],["$","li","11",{"children":[["$","span",null,{"className":"print-ref-num","children":"12"}],["$","span",null,{"children":"Abadi, M., and Andersen, D. G. (2016). Learning to Protect Communications with Adversarial Neural Cryptography. arXiv:1610.06918."}]]}],["$","li","12",{"children":[["$","span",null,{"className":"print-ref-num","children":"13"}],["$","span",null,{"children":"Signal. The Double Ratchet Algorithm specification."}]]}],["$","li","13",{"children":[["$","span",null,{"className":"print-ref-num","children":"14"}],["$","span",null,{"children":"NIST Computer Security Resource Center. Glossary entry: nonce."}]]}],["$","li","14",{"children":[["$","span",null,{"className":"print-ref-num","children":"15"}],["$","span",null,{"children":"NIST Computer Security Resource Center. Glossary entry: salt."}]]}],"$L61","$L62"]}],"$undefined","$undefined"]}] +31:["$","div","4",{"children":[["$","p",null,{"className":"mb-3","children":"What the analogy legitimately buys is a set of mechanisms, not a claim of cognitive equivalence. The infant-like part of the design is:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"learning through repeated shared experience rather than instruction;"}],["$","li","1",{"className":"pl-1","children":"establishing joint attention on the same events;"}],["$","li","2",{"className":"pl-1","children":"receiving consequences from interactions that work and interactions that fail;"}],["$","li","3",{"className":"pl-1","children":"revising provisional meanings over time instead of being handed final ones;"}],["$","li","4",{"className":"pl-1","children":"developing conventions with one recurring partner."}]]}]]}] +32:["$","p","5",{"children":"There is a related trap in the prompt. Telling a model to act like a one-year-old does not make it one. It produces a culturally learned caricature of infancy: baby talk, simplified grammar, emotional dependence, all of it copied from human writing about children and none of it evidence about anything. The concept rules the persona out. Internally the participant is called a Learner, and baby-like behaviour has to come from what the agent can observe, remember, emit, and learn, not from being asked to perform it."}] +33:["$","li","3",{"className":"pl-1","children":"entries may describe symbols, sequences, ordering, grammar, or repair signals;"}] +34:["$","li","4",{"className":"pl-1","children":"entries record confidence, supporting evidence, contradictory evidence, and abandoned meanings;"}] +35:["$","li","5",{"className":"pl-1","children":"a Baby records its intended meaning when speaking and its inferred meaning when receiving;"}] +36:["$","li","6",{"className":"pl-1","children":"a channel message and its required private ledger mutation commit atomically;"}] +37:["$","li","7",{"className":"pl-1","children":"entries carry enough run and turn references to trace them to observable evidence;"}] +38:["$","li","8",{"className":"pl-1","children":"neither Baby can query, receive, summarise, or infer from the other's ledger through any system-provided interface."}] +39:["$","p","4",{"children":"Rule seven is doing quiet work. If a message could be sent and the corresponding hypothesis written afterwards, the ledger becomes a place to record what the agent wishes it had meant. Committing both together makes the record contemporaneous."}] +3a:["$","p","5",{"children":"An agent that cannot write English needs a different arrangement, and the concept provides two layers. The agent-native ledger holds what the learner actually uses: association weights, probability distributions, embeddings, confidence, episode references, prediction errors, revision history. The human audit ledger is a deterministic or BabySitter-generated interpretation of that state, clearly labelled as external analysis and never fed back to either Baby. Confusing the second for the first would mean presenting the researchers' reconstruction as the agent's own definition."}] +3b:["$","tr","7",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"BabySitter"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Monitor-only baseline; any safety intervention recorded as a protocol exception"}]]}] +3c:["$","figcaption",null,{"className":"mt-3 text-sm text-muted","children":"Runs vary one major axis at a time before any factorial combination is attempted."}] +3d:["$","p","2",{"children":"Task difficulty moves through ten stages, from naming four distinct objects with one-symbol messages, through attributes, spatial relations, actions, multi-symbol composition, order-sensitive grammar, and repair under ambiguity, to held-out generalisation with learning disabled, long-run drift, and cross-architecture comparison. Vocabulary size, turn count, reward structure, and exposure history are controlled at each stage so that runs remain comparable."}] +3e:["$","p","3",{"children":"The learning-mechanism axis carries a question the transcript cannot answer on its own. Convergence can be driven by reinforcement, by pretrained linguistic priors, by persistent memory, by intrinsic motivation, or by self-supervised prediction, and all five can look alike in a log. Running the same exercise suite under a no-learning control, a frozen-weights memory baseline, extrinsic-reward learning, intrinsic-motivation learning, and reward-free self-supervision is what makes the mechanism itself measurable. The aim is not only to observe that a language emerged, but to say which process caused it."}] +3f:["$","p","4",{"children":"Strict isolation constrains how that reinforcement learning may be implemented. Centralised training, backpropagation through both agents, shared replay buffers, and shared gradients all move information outside the permitted channel. The research-grade baseline updates each policy independently, and any centralised variant is reported as a separate, weaker-isolation condition."}] +40:["$","p","6",{"children":"The governing principle, stated in the concept as a single line, is not to ask a model to perform infancy but to construct an environment in which limited, grounded, auditable learning is the only path to a successful interaction."}] +41:["$","tr","3",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Chaabouni and colleagues (2020)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Generalisation to novel combinations and measured compositionality can come apart."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Held-out behaviour must be tested directly; no single compositionality score is proof of understanding."}]]}] +42:["$","tr","4",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Lowe and colleagues (2019)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Positive signaling and positive listening are different things, and reward does not distinguish them."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Causal intervention is mandatory rather than optional."}]]}] +43:["$","tr","5",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Dessi, Kharitonov, and Baroni (2021)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Symbol ablation and substitution yield interpretable evidence about what a receiver uses."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Direct precedent for the intervention tests that validate ledger claims."}]]}] +44:["$","tr","6",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Mihai and Hare (2021)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Neural agents communicate through learned drawings rather than a supplied discrete vocabulary."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"A blank canvas is a credible carrier for runs with no symbol library."}]]}] +45:["$","tr","7",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Baronchelli and colleagues (2005)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Naming Game agents converge on shared vocabulary through local interaction with no central teacher."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Convergence time, failed conventions, and memory update rules are first-class evidence."}]]}] +46:["$","tr","8",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Jaques and colleagues (2019)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Rewarding causal influence over a partner improves coordination without assigning a vocabulary."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"An endogenous social signal is a plausible substitute for task reward, and still a designed bias."}]]}] +47:["$","tr","9",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Cao and colleagues (2018)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"In semi-cooperative negotiation, self-interest and reward structure decide whether communication stays informative."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Negotiation is an advanced condition, not a description of the cooperative baseline."}]]}] +48:["$","tr","10",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Abadi and Andersen (2016)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Neural agents learn to protect messages from an adversary given a shared key."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Learned protective encodings are real and are not a substitute for formal security analysis."}]]}] +49:["$","figcaption",null,{"className":"mt-3 text-sm text-muted","children":"Selected prior work and what each one constrains in this design."}] +4a:["$","p","3",{"children":"This body of work also sharpens the infant-like caveat from the other direction. Most experiments in the field give their agents substantial structure: a fixed channel, an objective, a bounded vocabulary, joint training, or a reward. No dictionary does not mean no inductive bias, and no teacher does not mean no learning signal."}] +4b:["$","p","4",{"children":"The CoLab's own Diplomacy Table is direct project prior art. It already models independent delegation seats, a convener that advances rounds, operator-wide visibility alongside seat-specific perspectives, transcripts and ticks and redaction boundaries, and recorded replay with debrief. Those map cleanly onto two Babies, controlled turns, private observations, and replayable evidence. Caucuses, coalition rooms, and direct delegation links do not map safely and are disabled."}] +4c:["$","div","5",{"children":[["$","p",null,{"className":"mb-3","children":"What the review implies for the specification:"}],["$","ol",null,{"className":"list-decimal space-y-2 pl-6 marker:font-mono marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"the project sits inside a mature field, and its distinctive combination is independent ledgers, supervisory audit, mixed agent types, and affect;"}],["$","li","1",{"className":"pl-1","children":"grounding, bandwidth, memory, and learning pressure strongly shape what language appears;"}],["$","li","2",{"className":"pl-1","children":"successful coordination coexists happily with a brittle lookup code or with a receiver that ignores messages;"}],["$","li","3",{"className":"pl-1","children":"causal interventions and held-out generalisation are mandatory;"}],["$","li","4",{"className":"pl-1","children":"a visual carrier removes the need for a symbol library but not the need for a carrier;"}],["$","li","5",{"className":"pl-1","children":"intrinsic social influence can replace task reward and remains a designed learning bias;"}],["$","li","6",{"className":"pl-1","children":"negotiation is a valid advanced condition, not the right description of the baseline;"}],["$","li","7",{"className":"pl-1","children":"ephemeral conventions, learned cryptography, one-time pads, per-message keys, nonces, and salts are distinct mechanisms and must not be conflated."}]]}]]}] +4d:["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Self-supervised ungrounded baseline"}] +4e:["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E11 infrastructure"}] +4f:["$","tr","7",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E13"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"No predefined symbol library"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E11 or E12"}]]}] +50:["$","tr","8",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E14"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Turn-taking, role reversal, and repair"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E13"}]]}] +51:["$","tr","9",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E15"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Composition and held-out generalisation"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E14"}]]}] +52:["$","tr","10",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E16"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Causal listening and ledger validity"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E15"}]]}] +53:["$","tr","11",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E20"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Constrained affect-channel study"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E16"}]]}] +54:["$","tr","12",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E21"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"RL versus non-RL learning comparison"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E16"}]]}] +55:["$","tr","13",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E22"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Developmental plasticity and curriculum"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E16"}]]}] +56:["$","tr","14",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E30"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Partner replacement and zero-shot transfer"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E20 to E22"}]]}] +57:["$","tr","15",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E31"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Longitudinal drift and stability"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E30"}]]}] +58:["$","tr","16",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E32"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Cooperative signaling versus negotiation"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E31"}]]}] +59:["$","tr","17",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E40"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Ephemeral encoding and adversarial cryptography"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E32"}]]}] +5a:["$","tr","18",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E50"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Multi-seed replication and study closeout"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E40"}]]}] +5b:["$","figcaption",null,{"className":"mt-3 text-sm text-muted","children":"The experiment index, as pre-registered. Every entry currently reads not started, and the results column is empty by design."}] +5c:["$","p","2",{"children":"The first four experiments are about the apparatus. E00 qualifies the ledger: it must be possible to detect a modified, deleted, inserted, or reordered entry, and to prove that later checkpoints extend earlier ones. E01 is a red team against the channel. E02 hunts human language in observations and metadata. E03 establishes chance, no-communication, and random-message baselines, without which a success rate means nothing. Only then does E10 put agents in the room."}] +5d:["$","div","3",{"children":[["$","p",null,{"className":"mb-3","children":"Unless an experiment explicitly varies one of them, the notebook holds these constant:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"the two learners run in separate processes or containers with no direct route between them;"}],["$","li","1",{"className":"pl-1","children":"the deterministic gateway is the only communication path;"}],["$","li","2",{"className":"pl-1","children":"the supervisor has complete read-only access and sends no guidance or reward;"}],["$","li","3",{"className":"pl-1","children":"private observations contain no human-language labels or text;"}],["$","li","4",{"className":"pl-1","children":"public output uses only the pre-registered carrier;"}],["$","li","5",{"className":"pl-1","children":"the affect channel is disabled unless it is the subject of study;"}],["$","li","6",{"className":"pl-1","children":"schedules and seeds are fixed before the run;"}],["$","li","7",{"className":"pl-1","children":"held-out evaluation runs with learning disabled;"}],["$","li","8",{"className":"pl-1","children":"no production secrets or personal data appear in any experiment."}]]}]]}] +5e:["$","p","4",{"children":"Every run copies a standard record: run and experiment identifiers, dates, operator, both models and both training modes, scenario and prompt and gateway configuration hashes, random seed, policy initialisation hashes, the protocol and software commits, final ledger sizes and roots for both agents, the channel root, the final checkpoint hash, the Base anchor transaction, the verifier result, protocol deviations, and an explicit disposition of valid, invalid, or aborted. The notebook is the human workflow record; anchored run bundles remain the authoritative evidence."}] +5f:["$","p","5",{"children":"The status is unambiguous and worth stating plainly: this is a concept and pre-specification phase. Nineteen experiments are written up and none has been run. The next milestone is a testable specification covering runtime architecture, channel contract, ledger schema, experiment matrix, evaluation criteria, isolation model, and evidence requirements."}] +60:["$","p","8",{"children":"The principle underneath all of it survives every one of those choices. Baby A and Baby B may hold human language internally, but they must build their shared external language without sending human language, translations, or private ledger contents to one another. Everything else is a question about how to find out what happens next."}] +61:["$","li","15",{"children":[["$","span",null,{"className":"print-ref-num","children":"16"}],["$","span",null,{"children":"Ethical Tech CoLab (2026). Agentic Language Development: concept document, ledger integrity design, and experiment notebook."}]]}] +62:["$","li","16",{"children":[["$","span",null,{"className":"print-ref-num","children":"17"}],["$","span",null,{"children":"Ethical Tech CoLab. Diplomacy Table Live: independent delegation seats, controlled rounds, and replayable transcripts."}]]}] +1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] +63:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +18:null +1d:[["$","title","0",{"children":"Agentic Language Development (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L63","6",{}]] diff --git a/static-site/print/ai-carbon-footprint/__next._full.txt b/static-site/print/ai-carbon-footprint/__next._full.txt index b051f2eef..aad1f8aab 100644 --- a/static-site/print/ai-carbon-footprint/__next._full.txt +++ b/static-site/print/ai-carbon-footprint/__next._full.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -12:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","print","ai-carbon-footprint",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","ai-carbon-footprint","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -14:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -16:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","print","ai-carbon-footprint",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","ai-carbon-footprint","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:"$Sreact.suspense" -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1c:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -29,7 +29,7 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 19:[] 10:"$W19" 11:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1e:Ta14, @page { size: 210mm 297mm; margin: 20mm 18mm; } @@ -183,6 +183,6 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 6d:["$","li","56",{"children":[["$","span",null,{"className":"print-ref-num","children":"57"}],["$","span",null,{"children":"Yale School of the Environment. (2024). New initiative focuses on reducing the carbon footprint of computer systems and AI."}]]}] 6e:["$","li","57",{"children":[["$","span",null,{"className":"print-ref-num","children":"58"}],["$","span",null,{"children":"Yale School of the Environment. (2024, November 13). Can we mitigate AI's environmental impacts?"}]]}] 1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -75:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +75:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 18:null 1d:[["$","title","0",{"children":"AI's Carbon Footprint (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L75","6",{}]] diff --git a/static-site/print/ai-carbon-footprint/__next._head.txt b/static-site/print/ai-carbon-footprint/__next._head.txt index 0ad567211..36fb018af 100644 --- a/static-site/print/ai-carbon-footprint/__next._head.txt +++ b/static-site/print/ai-carbon-footprint/__next._head.txt @@ -1,6 +1,6 @@ 1:"$Sreact.fragment" -2:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -3:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +2:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +3:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 4:"$Sreact.suspense" -5:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +5:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 0:{"rsc":["$","$1","h",{"children":[null,["$","$L2",null,{"children":[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]]}],["$","div",null,{"hidden":true,"children":["$","$L3",null,{"children":["$","$4",null,{"name":"Next.Metadata","children":[["$","title","0",{"children":"AI's Carbon Footprint (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L5","6",{}]]}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/print/ai-carbon-footprint/__next._index.txt b/static-site/print/ai-carbon-footprint/__next._index.txt index f98a4fd90..18b9d82b6 100644 --- a/static-site/print/ai-carbon-footprint/__next._index.txt +++ b/static-site/print/ai-carbon-footprint/__next._index.txt @@ -1,13 +1,13 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -d:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +d:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} 8:["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$Ld",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$Ld",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$Ld",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$Ld",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] diff --git a/static-site/print/ai-carbon-footprint/__next._tree.txt b/static-site/print/ai-carbon-footprint/__next._tree.txt index 4e08a869b..ee86f99d6 100644 --- a/static-site/print/ai-carbon-footprint/__next._tree.txt +++ b/static-site/print/ai-carbon-footprint/__next._tree.txt @@ -1,7 +1,7 @@ -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] 0:{"tree":{"name":"","param":null,"prefetchHints":16,"slots":{"children":{"name":"print","param":null,"prefetchHints":0,"slots":{"children":{"name":"slug","param":{"type":"d","key":"ai-carbon-footprint","siblings":null},"prefetchHints":0,"slots":{"children":{"name":"__PAGE__","param":null,"prefetchHints":0,"slots":null}}}}}}},"staleTime":300,"buildId":"static-snapshot"} diff --git a/static-site/print/ai-carbon-footprint/__next.print.txt b/static-site/print/ai-carbon-footprint/__next.print.txt index 460afe242..fa617e997 100644 --- a/static-site/print/ai-carbon-footprint/__next.print.txt +++ b/static-site/print/ai-carbon-footprint/__next.print.txt @@ -1,4 +1,4 @@ 1:"$Sreact.fragment" -2:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -3:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +2:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +3:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] 0:{"rsc":["$","$1","c",{"children":[null,[["$","style",null,{"children":"\n /* The site is dark-first, and a dark color-scheme makes Chrome paint\n the printed sheet's margin area dark too. Paper is light. */\n html { color-scheme: light !important; }\n html, body { background: #ffffff !important; }\n body::after { content: none !important; }\n header, footer, .site-bg, .aura { display: none !important; }\n "}],["$","$L2",null,{"parallelRouterKey":"children","template":["$","$L3",null,{}]}]]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/print/ai-carbon-footprint/index.html b/static-site/print/ai-carbon-footprint/index.html index e43ab8a97..e78a9d6b9 100644 --- a/static-site/print/ai-carbon-footprint/index.html +++ b/static-site/print/ai-carbon-footprint/index.html @@ -1,4 +1,4 @@ -AI's Carbon Footprint (print edition) · NYU Ethical Tech CoLab

Publications · Academic report

AI's Carbon Footprint

The Environmental Impact of Artificial Intelligence

Ethical Tech CoLab

Professor Yorke Rhodes III

May 2025

Alexandra Du, Elizabeth Matthews, Emily Harrington, Hannah Zhao, Jennifer Hofmann, Natasha Nagarajan, Renata Gladkikh, and Smita Samanta

AI's increasing environmental costs — energy-intensive training and inference, data centers, and hardware — require urgent action. A combination of sustainable AI practices, policy interventions, and technological innovations can significantly reduce AI's ecological footprint without hindering its development and role in advancing society.

Key figures

1,287 MWh

to train GPT-3 — the annual energy use of ~130 U.S. households

552 → 24 t

CO₂ to train GPT-3 on a coal grid vs. a renewable-powered grid

240–340 TWh

electricity the world's data centers already draw each year

10×

the emissions of a query in coal-heavy Wyoming vs. hydro Quebec

01

Introduction

Artificial intelligence (AI) is transforming industries and tackling global challenges, including climate change. However, AI's rapid expansion comes with a significant but often overlooked environmental cost. The development and deployment of AI models, particularly large-scale systems, consume vast amounts of energy (Harvard Business Review, 2024). This contributes to carbon emissions, water consumption, and electronic waste (Earth.org, 2025). As AI adoption grows, understanding and mitigating its ecological footprint is critical. This paper examines the environmental impact of AI, focusing on its energy demands, data center infrastructure, and hardware requirements. It also explores mitigation strategies, ethical considerations, and policy frameworks to promote sustainable AI development. While AI presents significant technological advancements for society, its increasing environmental costs seen through energy-intensive training and inference, data centers, and hardware, require urgent action. This paper argues that a combination of sustainable AI practices, policy interventions, and technological innovations can significantly reduce AI's ecological footprint without hindering its development and role in advancing society.

02

AI's Energy Consumption

Artificial intelligence has become a major driver of energy consumption, with inference accounting for 60 to 80 percent of total energy use in many real-world applications. This demand has been propelled by the exponential growth of generative AI tools such as ChatGPT and their integration into global infrastructure.

The AI lifecycle consists of two key phases: training and inference. Training refers to the process of developing an AI model by exposing it to vast datasets to learn patterns, relationships, and decision-making rules. This phase involves iterative adjustments to the model's parameters, often numbering in the hundreds of billions, to optimize performance. Training is computationally intensive, requiring high-performance computing clusters equipped with thousands of GPUs or TPUs operating continuously for weeks or months. OpenAI's 175 billion-parameter model consumed 1,287 megawatt-hours during training, equivalent to the annual energy use of 130 U.S. households. Model size plays a critical role in training energy consumption, with DeepMind's 280 billion-parameter system using 1,066 megawatt-hours. The complexity of datasets also contributes significantly, as preprocessing petabytes of data, such as Common Crawl for GPT-3, introduces ancillary energy costs. Another driver of energy demand is hyperparameter tuning, where optimizing model configurations often requires multiple training runs, increasing energy consumption by 20 to 40 percent.

Between 2012 and 2018, training energy consumption grew at a tenfold annual rate, far outpacing efficiency improvements from Moore's Law. Innovations such as Google's Tensor Processing Units have helped mitigate this growth, enabling the company to train a model seven times larger than GPT-3 using 33 percent less energy. However, the relationship between model size and energy use is nonlinear, as doubling parameters can increase energy consumption by three to five times due to memory bandwidth bottlenecks and communication overhead in distributed training. While pre-training accounts for 90 to 95 percent of total lifecycle energy in large models, fine-tuning's cumulative impact grows with frequent, decentralized use by developers.

Inference, which occurs after deployment, is the process of using trained models to generate outputs in response to new inputs. While inference consumes far less energy per task than training, its cumulative impact is significant at scale. A single ChatGPT response requires 0.047 kilowatt-hours, comparable to streaming Netflix for three and a half minutes. At the scale of one trillion queries annually, that unit figure implies roughly 47 terawatt-hours (0.047 kilowatt-hours across a trillion queries is 47,000,000 megawatt-hours) — more than Ireland's entire national electricity consumption, and on the order of 36,000 times the energy used to train GPT-3 once. Two caveats belong with that number. It is an upper bound: the 0.047 kilowatt-hour figure sits at the high end of published estimates, which commonly range from about 0.002 to 0.005 kilowatt-hours per query, and every total below scales linearly with whichever value is used. And it is a projection from a per-query unit cost, not a measured system-wide total. Several factors amplify inference energy consumption, including the rapid adoption of generative AI, the expansion of edge computing, and the real-time demands of applications such as autonomous vehicles. Tools like DALL-E and Midjourney consume 2.9 kilowatt-hours per 1,000 image generations, which is 360 times more than text classification tasks. Deploying AI in billions of smartphones and IoT devices further complicates energy measurement. Real-time applications require continuous inference, preventing energy-saving idle states and adding to the overall power demand.

Leading technology companies report that inference now accounts for the majority of their AI-related energy budgets, surpassing training in operational impact. At scale, inference energy consumption can dwarf the energy used in training. Training GPT-3 required 1,287 megawatt-hours, but if it were to serve 4.3 billion users each submitting one query per day, annual inference energy consumption would reach roughly 74 terawatt-hours (4.3 billion users across 365 days at 0.047 kilowatt-hours a query is about 73,800,000 megawatt-hours) — on the order of 57,000 times the energy used to train the model once. The two scenarios are consistent with each other by construction: this one represents about 1.57 trillion queries against the one trillion above, and yields about 1.57 times the energy. Both are illustrative arithmetic on a single unit figure rather than measurements, and both should be read as upper bounds for the reason given above.

The environmental impact of AI is further amplified by the carbon intensity of the energy sources used to power the underlying infrastructure. For example, training GPT-3 on a coal-powered grid emits 552 tons of carbon dioxide, whereas using a renewable-powered grid reduces emissions to just 24 tons. However, it's important to note that many cloud data centers increasingly rely on partially renewable energy mixes. For instance, if a cloud provider operates with an average of 40% renewable energy, then AI applications hosted on that infrastructure inherit that carbon profile — a critical calibration point when comparing emissions across different energy grids. The emissions from inference also vary significantly based on where queries are processed. A ChatGPT query in Wyoming, where 95 percent of electricity comes from coal, generates approximately ten times the emissions of the same query processed in Quebec, where 94 percent of power is hydroelectric. Tools like WattTime's estimation models, which use real-time, location-specific data to calculate emissions based on marginal power sources, are increasingly essential for more accurately assessing AI's carbon footprint at a granular level.

The disparity in emissions across different energy grids underscores the importance of sustainable AI practices. Several strategies have emerged to reduce AI's energy footprint, including powering data centers with renewables, implementing dynamic voltage scaling to reduce GPU power use by 25 percent, and adopting model compression techniques such as knowledge distillation to create smaller models with 60 percent lower inference costs. Carbon-aware scheduling, which optimizes training times to coincide with high renewable energy availability, has been shown to cut emissions by 30 percent.

AI energy consumption is typically measured using three primary metrics: kilowatt-hours, GPU or TPU hours, and carbon emissions. Kilowatt-hours quantify direct energy use, with training OpenAI's GPT-3 model requiring 1,287 megawatt-hours — equivalent to the electricity use of more than 100 U.S. households in a year. GPU or TPU hours reflect the duration and intensity of hardware utilization. Training a 110 million-parameter BERT model for 80 hours on 64 GPUs required 1,507 kilowatt-hours, while inference tasks on an NVIDIA A100 GPU operate at 120 to 200 watts depending on workload. Carbon emissions provide insight into the environmental trade-offs of AI development. That same 1,507 kilowatt-hour BERT-base training run, executed on a coal-powered grid at roughly one kilogram of carbon dioxide per kilowatt-hour, corresponds to about 3,300 pounds of carbon dioxide, whereas fine-tuning the model produced 26 pounds of carbon dioxide per run. The widely quoted 626,000-pound figure comes from the same study (Strubell et al., 2019) but describes something much larger: a full neural architecture search with hyperparameter tuning, comprising thousands of training runs rather than one. Conflating the two overstates the cost of training a single model by more than two orders of magnitude.

These figures illustrate the growing impact of AI on global energy demand. Data centers worldwide already consume between 240 and 340 terawatt-hours annually, and as AI adoption accelerates, addressing energy efficiency and sustainability will become a critical priority.

03

Data Centers & Hardware Impact

Data centers are the backbone of modern AI, providing the computational power needed to train and run machine learning models. These facilities house thousands of high-performance servers, processing vast amounts of data for cloud computing, AI applications, and digital services. AI data centers are energy-intensive, consuming an estimated 1–2% of global electricity, a figure projected to rise with AI becoming more advanced and popular. The source of this energy, either renewable or fossil fuels, significantly impacts their sustainability. Additionally, cooling systems, which are necessary to prevent overheating, consume enormous amounts of water and electricity, raising concerns about the environment and carbon footprints. Beyond energy consumption, the production of AI hardware such as GPUs, TPUs, and custom AI chips, has its own environmental toll. As AI companies scale their infrastructure, major tech firms like Google, Microsoft, and Amazon are investing in greener AI strategies, pushing for renewable-powered data centers, liquid cooling systems, and more efficient AI chips. As of 2023 data, there were an estimated 11,000 data centers globally (Minnix, 2025).

As mentioned, AI models require immense computing power to process vast amounts of data, train deep learning models, and perform real-time inference. Specialized hardware such as GPUs (Graphics Processing Units), TPUs (Tensor Processing Units), and other AI accelerators have become a driver in modern AI infrastructure. GPUs were originally designed for rendering graphics in gaming and visual applications, but have become essential for AI due to their ability to process many tasks simultaneously. The leading manufacturers of GPUs are Nvidia (A100, H100), AMD (Instinct MI300), and Intel (Gaudi AI chips). TPUs are custom AI chips developed by Google, specifically designed to accelerate deep learning computations. Unlike GPUs, TPUs are optimized for tensor operations, making them highly efficient for training and running large AI models. They consume less power per computation compared to GPUs but are mostly available within Google Cloud's ecosystem. These AI chips, and the magnets, batteries, and power systems that support them, rely on critical minerals such as lithium, cobalt, and nickel, alongside true rare-earth elements — the lanthanides plus scandium and yttrium — used in high-performance magnets and electronics. All of them require energy-intensive mining operations. The extraction of these minerals contributes to deforestation and habitat destruction in mining regions (e.g., the Democratic Republic of the Congo for cobalt, China for rare earth metals), toxic waste and pollution as refining these materials generates hazardous byproducts, and high carbon emissions from transportation and processing.

Data centers have been at the forefront of controversies due to their high use of energy, water usage for cooling systems and more. When it comes to direct energy consumption, it is important to focus on not just IT equipment (servers consuming immense energy amounts) but general energy used for other systems from lighting to monitoring, given the size of data centers.

This means that apart from IT equipment energy consumption focusing on servers, storage arrays and networking devices, individuals should also consider ancillary systems energy consumption which focuses on cooling systems, power distribution and backup systems (e.g. uninterruptible power supplies, power distribution units). The issue with this is that additional energy loss is expected and electrical conversion inefficiencies add to this. This is one of the issues of energy use by data centers, and with more data centers being built and operated, overall energy consumption is increasing. One key metric to consider for power usage is PUE — power usage effectiveness — which compares total facility energy consumption to the energy used by IT equipment. Because the numerator includes the denominator, PUE is bounded below by 1.0 and is unbounded above; a value under 1.0 is physically impossible. A PUE of 1.0 is the ideal floor, meaning all energy goes to IT equipment and none to cooling, power distribution, or other overhead, while real data centers typically run between about 1.1 and 2.0 (Gillis, 2022). A lower PUE value means better overall energy efficiency, and this metric can be used to compare energy efficiency of different data centers or companies. With certain cooling systems being implemented more (that use less energy), PUE can be lowered. The metric was developed by The Green Grid, a consortium focusing on improving data center energy efficiency (Digital Realty).

Apart from IT equipment energy consumption, a key aspect of energy use is cooling technologies that are implemented in data centers. The main types of system to consider are the following:

Air Cooling: This type of cooling relies on air conditioning, which circulates cooled air throughout the data center using tubes. The typical setup includes a raised floor design where cool air is delivered through perforated tiles in designated cold aisles, while hot air is exhausted from hot aisles. Hot/cold aisle containment helps improve efficiency by physically separating hot and cold air streams, reducing mixing and lowering energy consumption. The main advantage of air cooling is its familiarity, relatively low upfront costs, and easy integration into existing infrastructure. However, as computing densities increase, air cooling may struggle to remove heat effectively, which leads to higher use of energy (Schaap, 2024).

Liquid Cooling: Liquid cooling is a more efficient alternative, using two types of cooling: direct-to-chip cooling and immersion cooling. In the first example, water or fluid circulates through plates. This is particularly beneficial for high-performance computing environments. Immersion cooling takes it a step further by submerging server components or racks in a liquid that absorbs heat directly from all surfaces. This method reduces reliance on air-based cooling, operates more quietly, and helps maintain lower operating temperatures, extending hardware lifespan. The issue with this example is the higher costs associated with the needed equipment (Vertiv).

Free Cooling: This example uses cooler outside air to replace hot air near equipment, as opposed to reusing the same air over and over that is cooled. This significantly reduces energy consumption and lowers operational costs. However, its effectiveness depends on location and climate variations as it works best when outside air is cool. Additionally, humidity and dust control must be carefully managed to prevent equipment damage (Masters DC).

Evaporative Cooling: This type of cooling has three forms — direct, indirect and two-stage. Evaporative cooling relies on the natural process of water evaporation to cool down outside air that is then blown onto equipment. When hot air meets pads filled with water, liquid turns into gas, cooling the facility, and remaining water from the pads is recirculated to keep the process ongoing. The main environment for this type of cooling would be in more humid areas (Airsys, 2024).

Hybrid Approaches: Many modern data centers use hybrid cooling systems, combining multiple technologies for optimal performance. For example, liquid cooling may be used for high-density racks, while air or free cooling is applied to other areas. This depends on the capability of each data center and environment — e.g. the need for one type of cooling system over another. Overall, liquid cooling has become more popular, but it is important to note that hybrid approaches are still considered the 'future,' as data centers become more efficient and a combination of different cooling techniques is used to increase efficiency (Robb, 2024).

A metric used for water use and efficiency is WUE — water usage effectiveness. This value measures how efficiently a data center uses water in its cooling and operational processes. A lower WUE indicates better water efficiency, meaning less water is used per unit of computing power, while a higher WUE suggests increased water consumption, which may be unsustainable in regions facing water scarcity both in the US and in other countries with higher temperatures and less access to water. Data centers in water-stressed areas must optimize cooling methods to minimize freshwater usage. For example, evaporative cooling systems reduce electricity consumption but require significant water use. Companies track WUE as part of their ESG goals, though maintaining a low WUE can be particularly challenging in climates where cooling efficiency must be balanced with limited water availability — a challenge to consider in future designs of data centers and sustainability goals (Higgins, 2024).

Given the substantial energy and resource demands of AI infrastructure, the need for sustainable solutions is more urgent than ever.

04

Mitigation Strategies & Sustainable AI

As AI models like GPT-4 and DALL·E grow larger and more complex, companies and academic institutions are working to address the immense energy demands required to train and operate these systems. Major players such as Google, Microsoft, Amazon, Meta, OpenAI, and leading universities are implementing innovative solutions to reduce AI's carbon footprint and promote sustainable AI development.

How to read the corporate claims in this section. Almost everything that follows is drawn from the companies' own sustainability pages and engineering blogs. These are the parties with the strongest incentive to present their record favourably, and three distinctions should travel with every figure. First, a pledge is not an achievement: a target set for 2030 or 2040 describes an intention, and the only evidence that it is being met is independently reported progress against it. Second, most corporate renewable-energy claims are market-based: the company buys renewable energy certificates or signs power purchase agreements sufficient to match its annual consumption, which is not the same as the location-based question of whether clean electricity was flowing on the grid at the hour and place the computation actually ran. Third, offsetting is not the same as not emitting; the quality and permanence of offsets is itself contested, and an offset tonne and an unemitted tonne are not interchangeable. The section reports what each company says, and marks where the claim is a pledge, where it rests on market-based accounting, and where independent reporting complicates it.

Google has been developing several projects in AI sustainability, and has set a target of powering its data centers, which host massive AI operations, with 100% carbon-free energy by 2030 (Google Data Centers, 2024). That is a target rather than a description of current operations, and it is a demanding one precisely because Google states it in hourly terms: matching consumption with clean energy on every grid, every hour, rather than annually in aggregate. Google's own reporting places its global hourly carbon-free share materially below 100%, and the gap between the annual-matching claim and the hourly one is where most of the difficulty in this commitment sits. To improve operational efficiency, Google employs AI-driven cooling systems, which have reduced the energy needed for cooling its data centers by up to 40% (Google DeepMind, 2016). The company also continues to develop custom Tensor Processing Units (TPUs) that are optimized to handle AI workloads more efficiently, using less electricity than conventional hardware (Google Cloud, 2018). Through these efforts, Google is reducing both the operational and computational energy demands of its AI systems.

Microsoft is also taking significant action to curb the environmental impact of AI. The company has pledged to be carbon-negative by 2030, which means removing more carbon from the atmosphere than it emits, including emissions from AI workloads run on Azure, its cloud platform (The Microsoft Cloud, 2024). The pledge should be read against the company's own subsequent environmental reporting, which has shown total emissions rising rather than falling since it was made, driven largely by the construction and operation of new datacentre capacity. That is the central tension of this whole section in miniature: efficiency commitments made in good faith can be outrun by growth in the underlying demand. To achieve this, Microsoft is investing in advanced clean energy solutions, such as battery storage and nuclear energy, to sustainably power AI data centers (Microsoft Sustainability, 2023). Additionally, Microsoft is working on optimizing AI models using techniques like Neural Architecture Search and muTransfer, which help reduce the energy needed to train and run AI systems. These methods focus on building more efficient models from the start, allowing AI to achieve high performance while minimizing carbon emissions (Microsoft Research, 2022).

Amazon Web Services (AWS), one of the largest providers of AI cloud computing, sits behind two commitments that an earlier version of this report ran together. Amazon's Climate Pledge commits the company to net-zero carbon across its operations by 2040. Separately, Amazon reports having matched its annual electricity consumption with renewable sources, a milestone it announced ahead of its own schedule (DataCentre Magazine, 2024). A target set for 2040 cannot have been met in 2024, and the two claims should not be merged. The renewable-matching claim is also market-based, resting on power purchase agreements and renewable energy certificates purchased across a year, an accounting method that has been criticised for allowing consumption in one region and hour to be offset by generation in another. AWS is also using real-time energy monitoring and workload distribution to ensure that computing resources are allocated efficiently, avoiding unnecessary energy use (About Amazon, 2024). In addition to these operational efforts, AWS is working on making its hardware more sustainable by focusing on recyclable and environmentally friendly server materials (Amazon Web Services, 2024).

Meta is working to reduce the carbon footprint of its AI systems through emissions transparency and infrastructure efficiency. Training its LLaMA 3-70B model generated an estimated 2,290 metric tons of CO₂, which Meta reports as offset by its sustainability initiatives (Hugging Face, 2024). Offset is the operative word: the tonnes were emitted, and a purchased offset is a claim that an equivalent quantity was avoided or removed elsewhere. The two are treated as equivalent in most corporate reporting and are not equivalent in the atmosphere, since offset quality, permanence and additionality are all contested. The same qualification applies to the claim of net zero operational emissions since 2020, which is a statement about scope one and scope two emissions net of renewable-energy matching and offsets, and does not cover the emissions embodied in manufacturing the hardware. Meta is also building AI-optimized data centers with liquid cooling and high energy efficiency (Meta Sustainability, 2024). It is worth noting that Meta's publication of a per-model training figure at all is unusual, and is a genuine transparency gain over competitors who publish nothing.

OpenAI, the developer of models like GPT-4 and DALL·E, operates some of the most computationally intensive AI systems in the world. However, OpenAI has not publicly released any sustainability strategy, carbon footprint disclosures, or commitments related to reducing emissions from its AI systems.

Finally, academic institutions are playing a key role in advancing sustainable AI practices. Researchers at the University of Michigan, for example, have developed training optimization techniques that could cut AI energy consumption by up to 75% (Champion, 2023). These methods focus on reducing redundant training processes and finding smarter ways to achieve high performance with less data and fewer computational cycles. At MIT, researchers created the 'Once-for-All' network, which cuts emissions by enabling one model to generate many smaller submodels without retraining — reducing training emissions by up to 1,300 times (MIT News, 2020). UMass Amherst is heading a $12 million NSF-backed initiative to cut computing emissions through integrated improvements across hardware and software systems (UMass Amherst, 2024). In addition, Yale and partner institutions are developing carbon reporting standards for machine learning to help shrink emissions across the sector (Yale School of the Environment, 2024).

The peer-reviewed literature is the appropriate counterweight to all of this, and it should be read directly rather than through the outlets that summarise it. Three studies generated most of the figures that circulate in reports of this kind. Strubell, Ganesh and McCallum (2019) produced the training-emissions numbers, including the 626,000 pound architecture-search figure so widely misattributed to training a single model. Patterson and colleagues (2021) measured the emissions of large neural network training directly and argued that the choice of model, processor, datacentre and grid together account for differences of two orders of magnitude, which is the finding that makes carbon-aware scheduling worth doing at all. Luccioni, Viguier and Ligozat (2022) estimated the footprint of the BLOOM model across its whole lifecycle, including hardware manufacture and idle consumption, and found that counting only the training run understates the total substantially. Where a corporate claim and one of these studies disagree, the study is the better authority.

Across both private companies and academic research, there is a growing recognition of the need to reduce AI's environmental impact. From optimizing model architectures and improving data center operations to designing efficient hardware and using renewable energy, these organizations are taking concrete steps to ensure that AI's future is not only powerful but also sustainable. While these efforts represent meaningful progress, continued innovation, investment in clean energy, and greater transparency will be essential to align AI's growth with global climate goals. The honest summary of this section is that the direction of travel is right, the pledges are mostly unmet by construction because their deadlines have not arrived, and the emissions curve has so far been driven more by growth in demand than bent by efficiency.

05

Regulatory Frameworks

The environmental sustainability of AI remains largely overlooked in both environmental law and technology regulations. UNEP has criticized the lack of policy focus on AI's ecological impact, noting that current efforts prioritize AI's trustworthiness rather than its environmental footprint. Hacker (2023) similarly highlights this gap and provides policy recommendations to address it. The European Union has taken steps to regulate AI through its 2024 AI Act, the world's first comprehensive AI law, which classifies AI systems by risk levels. While the European Parliament emphasized making AI systems environmentally friendly, this aspect was largely neglected in the final legislation. Earlier drafts included sustainability provisions such as sustainable AI development principles, energy logging requirements, and foundation model standards, but these were weakened during negotiations. The final Act relies on voluntary codes and limited disclosure requirements, with Article 40 mandating that general-purpose AI providers disclose estimated energy consumption. Additionally, the European Energy Efficiency Directive requires data centers exceeding 500 kW to report energy use, while Germany has imposed stricter regulations, requiring data centers with over 300 kW capacity to source 50% of their electricity from renewables by 2024, increasing to 100% by 2027.

In the United States, the Sustainable Data Centers Act, introduced in December 2024, aims to regulate the environmental footprint of data centers critical to AI operations by requiring alignment with state renewable energy targets and annual reporting of energy and water usage. On a global scale, effective AI regulation requires international cooperation, particularly concerning AI's resource-intensive supply chains and the mining of critical minerals. In February 2025, the AI Action Summit marked a turning point, with ecological sustainability taking center stage. This led to the formation of the Coalition for Environmentally Sustainable Artificial Intelligence, initiated by France, UNEP, and the International Telecommunication Union (ITU). With over 100 partners, including 11 countries, five international organizations, and 37 tech companies, the coalition aims to drive momentum toward sustainable AI practices. UNEP plans to publish a guide in 2025 to promote energy-efficient data centers based on international best practices. Additionally, AI is being leveraged for climate solutions, such as optimizing renewable energy, monitoring environmental changes, improving climate modeling, and enhancing disaster preparedness.

Several initiatives are actively advancing AI-driven climate solutions, underscoring the potential of AI in achieving sustainability goals. The WEKA Sustainable AI Initiative focuses on responsible AI applications that improve data efficiency while supporting sustainable development goals, including investments in reforestation projects to offset carbon emissions from data centers (WEKA, n.d.). The UN Climate Change's AI for Climate Action initiative explores AI's role in advancing climate-resilient and low-emissions development, particularly in developing countries, with the goal of delivering transformative climate action through AI-powered solutions (UNFCCC, n.d.). The Bezos Earth Fund's AI for Climate and Nature Initiative seeks to leverage AI to address climate change and protect natural ecosystems, including a $100 million grand challenge to unlock AI-driven solutions for climate and nature problems (Bezos Earth Fund, 2025).

While technological advancements can help reduce AI's environmental impact, policy frameworks and ethical considerations play a critical role in ensuring sustainable AI development. Effective governance, regulatory measures, and corporate responsibility initiatives are necessary to align AI's growth with long-term environmental goals.

06

Ethical & Policy Considerations

Government policies and international agreements are increasingly shaping the intersection of AI and sustainability, focusing on reducing the environmental impact of AI while promoting ethical and energy-efficient innovations. Various national strategies and global frameworks aim to balance technological advancement with ecological responsibility.

At the national level, governments are introducing AI sustainability policies to regulate energy consumption and carbon emissions. For example, the European Union's AI Act includes provisions for transparency in AI energy use, while the U.S. Department of Energy funds research on AI-driven energy efficiency. China's Green AI Initiative encourages data centers to adopt renewable energy sources and optimize power usage. Additionally, governments worldwide are promoting sustainable AI practices through tax incentives, grants, and stricter regulations on high-energy-consuming AI training processes.

Agreements like the Paris Agreement on Climate Change indirectly influence AI sustainability by urging nations to reduce carbon footprints, including those from digital infrastructures. The OECD AI Principles, adopted by over 40 countries, emphasize energy-efficient AI development and responsible deployment. The United Nations' Sustainable Development Goals (SDGs) encourage AI innovation to support climate action, particularly in renewable energy forecasting and environmental monitoring. Furthermore, the G7 and G20 summits have discussed AI's role in sustainable development, highlighting the need for global cooperation on eco-friendly AI deployment.

Despite these efforts, challenges remain in enforcing energy-efficient AI practices, as AI models grow more complex and require increasing computational power. As governments, organizations, and researchers work toward mitigating AI's environmental costs, it is essential to consider the broader implications of these efforts. Looking ahead, continued research, policy evolution, and cross-sector collaboration will be key to balancing AI's benefits with its sustainability challenges.

07

Conclusion & Future Directions

AI's rapid advancement presents both unprecedented opportunities and significant environmental challenges. As its adoption continues to grow, ensuring that AI development aligns with sustainability goals is no longer an option, but a necessity (World Economic Forum, 2024). While AI's energy consumption, reliance on resource-intensive data centers, and electronic waste generation pose serious environmental risks, innovative solutions in algorithmic efficiency, green data centers, and sustainable hardware development offer promising pathways forward (Nature, 2024; MIT News, 2025).

The future of AI sustainability depends on a multi-faceted approach that integrates technological innovation, policy intervention, and corporate responsibility. Society must prioritize energy-efficient AI architectures (Capitol Technology University, n.d.), while policymakers should implement regulations that incentivize renewable energy use and enforce transparency in AI's environmental impact (SIAM News, n.d.). Companies leading AI development must take proactive measures by investing in carbon-aware computing, energy-efficient chips, and responsible e-waste management.

Moving forward, AI should not only minimize harm but also become a key player in solving global environmental challenges. Its potential to optimize renewable energy distribution, enhance climate modeling, and improve conservation efforts demonstrates that AI can be part of the solution rather than just a contributor to the problem (AI Speakers Agency, n.d.; World Economic Forum, 2025). However, achieving this balance requires urgent and collaborative action across industries, governments, and academia. The question is no longer whether AI can be sustainable, but rather how quickly we can ensure that it is.

References

Sources

  1. 01About Amazon Staff. (2024, June 25). How AWS helps reduce carbon footprint of AI workloads. About Amazon.
  2. 02Adams, H. S. (2024, August 5). Study reveals AWS cloud can slash carbon emissions by 99% compared to on-prem data centers. DataCentre Magazine.
  3. 03AI Speakers Agency. (n.d.). How will AI enable a sustainable future?
  4. 04AIMultiple. (2025). Top 10 sustainability AI applications & real-life examples in 2025.
  5. 05Airsys. (2024). Evaporative cooling for data centers – Pros and cons.
  6. 06Bezos Earth Fund. (n.d.). AI for climate and nature initiative.
  7. 07Brightlio. (2025, April 19). 170 data center stats (April 2025).
  8. 08Capitol Technology University. (n.d.). Moving towards a more sustainable future using AI.
  9. 09Champion, Z. (2023, April 17). Optimization could cut the carbon footprint of AI training by up to 75%. University of Michigan News.
  10. 10Digital Realty. (n.d.). What is power usage effectiveness (PUE)?
  11. 11Earth.org. (2025). The real environmental impact of AI.
  12. 12European Parliament. (2023, August 6). EU AI Act: First regulation on artificial intelligence.
  13. 13Evans, R., & Gao, J. (2016, July 20). DeepMind AI reduces Google data centre cooling bill by 40%. DeepMind.
  14. 14EY. (n.d.). AI and sustainability: Opportunities, challenges, and impact.
  15. 15Forbes. (2024, March 26). Here's why data center cooling is the hottest innovation in the sector [By A. Schaap].
  16. 16Gillis, A. S. (2022, April 29). What is PUE (power usage effectiveness)? TechTarget.
  17. 17Google. (2024). Operating sustainably – Google Data Centers.
  18. 18Greenly. (n.d.). How can artificial intelligence help tackle climate change?
  19. 19Hacker, P. (2023, December 21). Sustainable AI regulation. Semantic Scholar.
  20. 20Harvard Business Review. (2024, July 15). The uneven distribution of AI's environmental impacts.
  21. 21Higgins, A. (2024, November 13). What is water usage effectiveness (WUE) in data centers? Equinix.
  22. 22Hugging Face. (2024). Meta-Llama-3-70B.
  23. 23Intel. (n.d.). Advancing toward more sustainable AI.
  24. 24Kunze, J. (2024, October 4). Environmental impacts of artificial intelligence: Peril, promise, and policy. Society for Industrial and Applied Mathematics.
  25. 25Laranjeira De Pereira, J. R. (2024, April 8). The EU AI Act and environmental protection: The case for a missed opportunity. Heinrich Böll Stiftung, Brussels Office.
  26. 26Latham & Watkins. (2024, July 12). US environmental, social, and governance legal considerations for AI companies.
  27. 27Luccioni, A. S., Viguier, S., & Ligozat, A.-L. (2022). Estimating the carbon footprint of BLOOM, a 176B parameter language model.
  28. 28Masters DC. (n.d.). How data center cooling works and why it is brilliant.
  29. 29Matheson, R. (2020, April 23). Reducing the carbon footprint of artificial intelligence. MIT News.
  30. 30McQuate, S. (2023, July 27). Q&A: UW researcher discusses just how much energy ChatGPT uses. UW News.
  31. 31Meta. (2024). Meta sustainability report 2024. Meta Sustainability.
  32. 32Microsoft. (2023). Improving sustainability with AI. Microsoft Sustainability.
  33. 33MIT News. (2025). Explained: Generative AI's environmental impact.
  34. 34National Institute of Standards and Technology. (2023). Artificial intelligence risk management framework (AI RMF 1.0) (NIST AI 100-1).
  35. 35National Institute of Standards and Technology. (2024). AI risk management framework: Generative AI profile (NIST AI 600-1).
  36. 36Nature. (2024). Ecological footprints, carbon emissions, and energy transitions.
  37. 37Patterson, D., Gonzalez, J., Le, Q., Liang, C., Munguia, L.-M., Rothchild, D., So, D., Texier, M., & Dean, J. (2021). Carbon emissions and large neural network training.
  38. 38Publications Office of the European Union. (2024). Regulation (EU) 2024/1689 of the European Parliament and of the Council.
  39. 39Robb, D. (2024, June 3). Hybrid cooling: The bridge to full liquid cooling in data centers. Data Center Knowledge.
  40. 40Russinovich, M. (2024, September 12). Sustainable by design: Innovating for energy efficiency in AI, part 1. The Microsoft Cloud Blog.
  41. 41SIAM News. (n.d.). Environmental impacts of artificial intelligence: Peril, promise, and policy.
  42. 42Sitecore. (n.d.). Harnessing the power of AI for a sustainable future.
  43. 43Soto, K. (2018, August 30). What makes TPUs fine-tuned for deep learning? Google Cloud Blog.
  44. 44Strubell, E., Ganesh, A., & McCallum, A. (2019). Energy and policy considerations for deep learning in NLP. Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, 3645–3650.
  45. 45U.S. Congress. (2024). Artificial Intelligence Environmental Impacts Act, 118th Congress (S. 3732).
  46. 46U.S. Department of Energy. (2024a). AI for energy: Opportunities for a modern grid and clean energy economy.
  47. 47U.S. Department of Energy. (2024b, April 29). DOE announces new actions to enhance America's global leadership in artificial intelligence.
  48. 48UN Environment Programme. (2025, February 11). New coalition aims to put artificial intelligence on a more sustainable path.
  49. 49UN Framework Convention on Climate Change. (n.d.). AI for climate action.
  50. 50University of California, Irvine Libraries. (n.d.). Ethics – AI in research.
  51. 51Venkatesan, V., & Karibandi, A. (2024, September 16). Embracing modernization with a sustainability focus. Amazon Web Services.
  52. 52Vertiv. (n.d.). Liquid cooling options for data centers.
  53. 53WEKA. (n.d.). Sustainable AI initiative.
  54. 54Westbrook, J. (2024, May 23). UMass researchers awarded $12M to shrink society's carbon footprint with AI & other computer science tools. UMass Amherst.
  55. 55World Economic Forum. (2024). AI will accelerate sustainability — but is no silver bullet.
  56. 56World Economic Forum. (2025). AI's role in the climate transition and how it can drive growth.
  57. 57Yale School of the Environment. (2024). New initiative focuses on reducing the carbon footprint of computer systems and AI.
  58. 58Yale School of the Environment. (2024, November 13). Can we mitigate AI's environmental impacts?

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file +

Publications · Academic report

AI's Carbon Footprint

The Environmental Impact of Artificial Intelligence

Ethical Tech CoLab

Professor Yorke Rhodes III

May 2025

Alexandra Du, Elizabeth Matthews, Emily Harrington, Hannah Zhao, Jennifer Hofmann, Natasha Nagarajan, Renata Gladkikh, and Smita Samanta

AI's increasing environmental costs — energy-intensive training and inference, data centers, and hardware — require urgent action. A combination of sustainable AI practices, policy interventions, and technological innovations can significantly reduce AI's ecological footprint without hindering its development and role in advancing society.

Key figures

1,287 MWh

to train GPT-3 — the annual energy use of ~130 U.S. households

552 → 24 t

CO₂ to train GPT-3 on a coal grid vs. a renewable-powered grid

240–340 TWh

electricity the world's data centers already draw each year

10×

the emissions of a query in coal-heavy Wyoming vs. hydro Quebec

01

Introduction

Artificial intelligence (AI) is transforming industries and tackling global challenges, including climate change. However, AI's rapid expansion comes with a significant but often overlooked environmental cost. The development and deployment of AI models, particularly large-scale systems, consume vast amounts of energy (Harvard Business Review, 2024). This contributes to carbon emissions, water consumption, and electronic waste (Earth.org, 2025). As AI adoption grows, understanding and mitigating its ecological footprint is critical. This paper examines the environmental impact of AI, focusing on its energy demands, data center infrastructure, and hardware requirements. It also explores mitigation strategies, ethical considerations, and policy frameworks to promote sustainable AI development. While AI presents significant technological advancements for society, its increasing environmental costs seen through energy-intensive training and inference, data centers, and hardware, require urgent action. This paper argues that a combination of sustainable AI practices, policy interventions, and technological innovations can significantly reduce AI's ecological footprint without hindering its development and role in advancing society.

02

AI's Energy Consumption

Artificial intelligence has become a major driver of energy consumption, with inference accounting for 60 to 80 percent of total energy use in many real-world applications. This demand has been propelled by the exponential growth of generative AI tools such as ChatGPT and their integration into global infrastructure.

The AI lifecycle consists of two key phases: training and inference. Training refers to the process of developing an AI model by exposing it to vast datasets to learn patterns, relationships, and decision-making rules. This phase involves iterative adjustments to the model's parameters, often numbering in the hundreds of billions, to optimize performance. Training is computationally intensive, requiring high-performance computing clusters equipped with thousands of GPUs or TPUs operating continuously for weeks or months. OpenAI's 175 billion-parameter model consumed 1,287 megawatt-hours during training, equivalent to the annual energy use of 130 U.S. households. Model size plays a critical role in training energy consumption, with DeepMind's 280 billion-parameter system using 1,066 megawatt-hours. The complexity of datasets also contributes significantly, as preprocessing petabytes of data, such as Common Crawl for GPT-3, introduces ancillary energy costs. Another driver of energy demand is hyperparameter tuning, where optimizing model configurations often requires multiple training runs, increasing energy consumption by 20 to 40 percent.

Between 2012 and 2018, training energy consumption grew at a tenfold annual rate, far outpacing efficiency improvements from Moore's Law. Innovations such as Google's Tensor Processing Units have helped mitigate this growth, enabling the company to train a model seven times larger than GPT-3 using 33 percent less energy. However, the relationship between model size and energy use is nonlinear, as doubling parameters can increase energy consumption by three to five times due to memory bandwidth bottlenecks and communication overhead in distributed training. While pre-training accounts for 90 to 95 percent of total lifecycle energy in large models, fine-tuning's cumulative impact grows with frequent, decentralized use by developers.

Inference, which occurs after deployment, is the process of using trained models to generate outputs in response to new inputs. While inference consumes far less energy per task than training, its cumulative impact is significant at scale. A single ChatGPT response requires 0.047 kilowatt-hours, comparable to streaming Netflix for three and a half minutes. At the scale of one trillion queries annually, that unit figure implies roughly 47 terawatt-hours (0.047 kilowatt-hours across a trillion queries is 47,000,000 megawatt-hours) — more than Ireland's entire national electricity consumption, and on the order of 36,000 times the energy used to train GPT-3 once. Two caveats belong with that number. It is an upper bound: the 0.047 kilowatt-hour figure sits at the high end of published estimates, which commonly range from about 0.002 to 0.005 kilowatt-hours per query, and every total below scales linearly with whichever value is used. And it is a projection from a per-query unit cost, not a measured system-wide total. Several factors amplify inference energy consumption, including the rapid adoption of generative AI, the expansion of edge computing, and the real-time demands of applications such as autonomous vehicles. Tools like DALL-E and Midjourney consume 2.9 kilowatt-hours per 1,000 image generations, which is 360 times more than text classification tasks. Deploying AI in billions of smartphones and IoT devices further complicates energy measurement. Real-time applications require continuous inference, preventing energy-saving idle states and adding to the overall power demand.

Leading technology companies report that inference now accounts for the majority of their AI-related energy budgets, surpassing training in operational impact. At scale, inference energy consumption can dwarf the energy used in training. Training GPT-3 required 1,287 megawatt-hours, but if it were to serve 4.3 billion users each submitting one query per day, annual inference energy consumption would reach roughly 74 terawatt-hours (4.3 billion users across 365 days at 0.047 kilowatt-hours a query is about 73,800,000 megawatt-hours) — on the order of 57,000 times the energy used to train the model once. The two scenarios are consistent with each other by construction: this one represents about 1.57 trillion queries against the one trillion above, and yields about 1.57 times the energy. Both are illustrative arithmetic on a single unit figure rather than measurements, and both should be read as upper bounds for the reason given above.

The environmental impact of AI is further amplified by the carbon intensity of the energy sources used to power the underlying infrastructure. For example, training GPT-3 on a coal-powered grid emits 552 tons of carbon dioxide, whereas using a renewable-powered grid reduces emissions to just 24 tons. However, it's important to note that many cloud data centers increasingly rely on partially renewable energy mixes. For instance, if a cloud provider operates with an average of 40% renewable energy, then AI applications hosted on that infrastructure inherit that carbon profile — a critical calibration point when comparing emissions across different energy grids. The emissions from inference also vary significantly based on where queries are processed. A ChatGPT query in Wyoming, where 95 percent of electricity comes from coal, generates approximately ten times the emissions of the same query processed in Quebec, where 94 percent of power is hydroelectric. Tools like WattTime's estimation models, which use real-time, location-specific data to calculate emissions based on marginal power sources, are increasingly essential for more accurately assessing AI's carbon footprint at a granular level.

The disparity in emissions across different energy grids underscores the importance of sustainable AI practices. Several strategies have emerged to reduce AI's energy footprint, including powering data centers with renewables, implementing dynamic voltage scaling to reduce GPU power use by 25 percent, and adopting model compression techniques such as knowledge distillation to create smaller models with 60 percent lower inference costs. Carbon-aware scheduling, which optimizes training times to coincide with high renewable energy availability, has been shown to cut emissions by 30 percent.

AI energy consumption is typically measured using three primary metrics: kilowatt-hours, GPU or TPU hours, and carbon emissions. Kilowatt-hours quantify direct energy use, with training OpenAI's GPT-3 model requiring 1,287 megawatt-hours — equivalent to the electricity use of more than 100 U.S. households in a year. GPU or TPU hours reflect the duration and intensity of hardware utilization. Training a 110 million-parameter BERT model for 80 hours on 64 GPUs required 1,507 kilowatt-hours, while inference tasks on an NVIDIA A100 GPU operate at 120 to 200 watts depending on workload. Carbon emissions provide insight into the environmental trade-offs of AI development. That same 1,507 kilowatt-hour BERT-base training run, executed on a coal-powered grid at roughly one kilogram of carbon dioxide per kilowatt-hour, corresponds to about 3,300 pounds of carbon dioxide, whereas fine-tuning the model produced 26 pounds of carbon dioxide per run. The widely quoted 626,000-pound figure comes from the same study (Strubell et al., 2019) but describes something much larger: a full neural architecture search with hyperparameter tuning, comprising thousands of training runs rather than one. Conflating the two overstates the cost of training a single model by more than two orders of magnitude.

These figures illustrate the growing impact of AI on global energy demand. Data centers worldwide already consume between 240 and 340 terawatt-hours annually, and as AI adoption accelerates, addressing energy efficiency and sustainability will become a critical priority.

03

Data Centers & Hardware Impact

Data centers are the backbone of modern AI, providing the computational power needed to train and run machine learning models. These facilities house thousands of high-performance servers, processing vast amounts of data for cloud computing, AI applications, and digital services. AI data centers are energy-intensive, consuming an estimated 1–2% of global electricity, a figure projected to rise with AI becoming more advanced and popular. The source of this energy, either renewable or fossil fuels, significantly impacts their sustainability. Additionally, cooling systems, which are necessary to prevent overheating, consume enormous amounts of water and electricity, raising concerns about the environment and carbon footprints. Beyond energy consumption, the production of AI hardware such as GPUs, TPUs, and custom AI chips, has its own environmental toll. As AI companies scale their infrastructure, major tech firms like Google, Microsoft, and Amazon are investing in greener AI strategies, pushing for renewable-powered data centers, liquid cooling systems, and more efficient AI chips. As of 2023 data, there were an estimated 11,000 data centers globally (Minnix, 2025).

As mentioned, AI models require immense computing power to process vast amounts of data, train deep learning models, and perform real-time inference. Specialized hardware such as GPUs (Graphics Processing Units), TPUs (Tensor Processing Units), and other AI accelerators have become a driver in modern AI infrastructure. GPUs were originally designed for rendering graphics in gaming and visual applications, but have become essential for AI due to their ability to process many tasks simultaneously. The leading manufacturers of GPUs are Nvidia (A100, H100), AMD (Instinct MI300), and Intel (Gaudi AI chips). TPUs are custom AI chips developed by Google, specifically designed to accelerate deep learning computations. Unlike GPUs, TPUs are optimized for tensor operations, making them highly efficient for training and running large AI models. They consume less power per computation compared to GPUs but are mostly available within Google Cloud's ecosystem. These AI chips, and the magnets, batteries, and power systems that support them, rely on critical minerals such as lithium, cobalt, and nickel, alongside true rare-earth elements — the lanthanides plus scandium and yttrium — used in high-performance magnets and electronics. All of them require energy-intensive mining operations. The extraction of these minerals contributes to deforestation and habitat destruction in mining regions (e.g., the Democratic Republic of the Congo for cobalt, China for rare earth metals), toxic waste and pollution as refining these materials generates hazardous byproducts, and high carbon emissions from transportation and processing.

Data centers have been at the forefront of controversies due to their high use of energy, water usage for cooling systems and more. When it comes to direct energy consumption, it is important to focus on not just IT equipment (servers consuming immense energy amounts) but general energy used for other systems from lighting to monitoring, given the size of data centers.

This means that apart from IT equipment energy consumption focusing on servers, storage arrays and networking devices, individuals should also consider ancillary systems energy consumption which focuses on cooling systems, power distribution and backup systems (e.g. uninterruptible power supplies, power distribution units). The issue with this is that additional energy loss is expected and electrical conversion inefficiencies add to this. This is one of the issues of energy use by data centers, and with more data centers being built and operated, overall energy consumption is increasing. One key metric to consider for power usage is PUE — power usage effectiveness — which compares total facility energy consumption to the energy used by IT equipment. Because the numerator includes the denominator, PUE is bounded below by 1.0 and is unbounded above; a value under 1.0 is physically impossible. A PUE of 1.0 is the ideal floor, meaning all energy goes to IT equipment and none to cooling, power distribution, or other overhead, while real data centers typically run between about 1.1 and 2.0 (Gillis, 2022). A lower PUE value means better overall energy efficiency, and this metric can be used to compare energy efficiency of different data centers or companies. With certain cooling systems being implemented more (that use less energy), PUE can be lowered. The metric was developed by The Green Grid, a consortium focusing on improving data center energy efficiency (Digital Realty).

Apart from IT equipment energy consumption, a key aspect of energy use is cooling technologies that are implemented in data centers. The main types of system to consider are the following:

Air Cooling: This type of cooling relies on air conditioning, which circulates cooled air throughout the data center using tubes. The typical setup includes a raised floor design where cool air is delivered through perforated tiles in designated cold aisles, while hot air is exhausted from hot aisles. Hot/cold aisle containment helps improve efficiency by physically separating hot and cold air streams, reducing mixing and lowering energy consumption. The main advantage of air cooling is its familiarity, relatively low upfront costs, and easy integration into existing infrastructure. However, as computing densities increase, air cooling may struggle to remove heat effectively, which leads to higher use of energy (Schaap, 2024).

Liquid Cooling: Liquid cooling is a more efficient alternative, using two types of cooling: direct-to-chip cooling and immersion cooling. In the first example, water or fluid circulates through plates. This is particularly beneficial for high-performance computing environments. Immersion cooling takes it a step further by submerging server components or racks in a liquid that absorbs heat directly from all surfaces. This method reduces reliance on air-based cooling, operates more quietly, and helps maintain lower operating temperatures, extending hardware lifespan. The issue with this example is the higher costs associated with the needed equipment (Vertiv).

Free Cooling: This example uses cooler outside air to replace hot air near equipment, as opposed to reusing the same air over and over that is cooled. This significantly reduces energy consumption and lowers operational costs. However, its effectiveness depends on location and climate variations as it works best when outside air is cool. Additionally, humidity and dust control must be carefully managed to prevent equipment damage (Masters DC).

Evaporative Cooling: This type of cooling has three forms — direct, indirect and two-stage. Evaporative cooling relies on the natural process of water evaporation to cool down outside air that is then blown onto equipment. When hot air meets pads filled with water, liquid turns into gas, cooling the facility, and remaining water from the pads is recirculated to keep the process ongoing. The main environment for this type of cooling would be in more humid areas (Airsys, 2024).

Hybrid Approaches: Many modern data centers use hybrid cooling systems, combining multiple technologies for optimal performance. For example, liquid cooling may be used for high-density racks, while air or free cooling is applied to other areas. This depends on the capability of each data center and environment — e.g. the need for one type of cooling system over another. Overall, liquid cooling has become more popular, but it is important to note that hybrid approaches are still considered the 'future,' as data centers become more efficient and a combination of different cooling techniques is used to increase efficiency (Robb, 2024).

A metric used for water use and efficiency is WUE — water usage effectiveness. This value measures how efficiently a data center uses water in its cooling and operational processes. A lower WUE indicates better water efficiency, meaning less water is used per unit of computing power, while a higher WUE suggests increased water consumption, which may be unsustainable in regions facing water scarcity both in the US and in other countries with higher temperatures and less access to water. Data centers in water-stressed areas must optimize cooling methods to minimize freshwater usage. For example, evaporative cooling systems reduce electricity consumption but require significant water use. Companies track WUE as part of their ESG goals, though maintaining a low WUE can be particularly challenging in climates where cooling efficiency must be balanced with limited water availability — a challenge to consider in future designs of data centers and sustainability goals (Higgins, 2024).

Given the substantial energy and resource demands of AI infrastructure, the need for sustainable solutions is more urgent than ever.

04

Mitigation Strategies & Sustainable AI

As AI models like GPT-4 and DALL·E grow larger and more complex, companies and academic institutions are working to address the immense energy demands required to train and operate these systems. Major players such as Google, Microsoft, Amazon, Meta, OpenAI, and leading universities are implementing innovative solutions to reduce AI's carbon footprint and promote sustainable AI development.

How to read the corporate claims in this section. Almost everything that follows is drawn from the companies' own sustainability pages and engineering blogs. These are the parties with the strongest incentive to present their record favourably, and three distinctions should travel with every figure. First, a pledge is not an achievement: a target set for 2030 or 2040 describes an intention, and the only evidence that it is being met is independently reported progress against it. Second, most corporate renewable-energy claims are market-based: the company buys renewable energy certificates or signs power purchase agreements sufficient to match its annual consumption, which is not the same as the location-based question of whether clean electricity was flowing on the grid at the hour and place the computation actually ran. Third, offsetting is not the same as not emitting; the quality and permanence of offsets is itself contested, and an offset tonne and an unemitted tonne are not interchangeable. The section reports what each company says, and marks where the claim is a pledge, where it rests on market-based accounting, and where independent reporting complicates it.

Google has been developing several projects in AI sustainability, and has set a target of powering its data centers, which host massive AI operations, with 100% carbon-free energy by 2030 (Google Data Centers, 2024). That is a target rather than a description of current operations, and it is a demanding one precisely because Google states it in hourly terms: matching consumption with clean energy on every grid, every hour, rather than annually in aggregate. Google's own reporting places its global hourly carbon-free share materially below 100%, and the gap between the annual-matching claim and the hourly one is where most of the difficulty in this commitment sits. To improve operational efficiency, Google employs AI-driven cooling systems, which have reduced the energy needed for cooling its data centers by up to 40% (Google DeepMind, 2016). The company also continues to develop custom Tensor Processing Units (TPUs) that are optimized to handle AI workloads more efficiently, using less electricity than conventional hardware (Google Cloud, 2018). Through these efforts, Google is reducing both the operational and computational energy demands of its AI systems.

Microsoft is also taking significant action to curb the environmental impact of AI. The company has pledged to be carbon-negative by 2030, which means removing more carbon from the atmosphere than it emits, including emissions from AI workloads run on Azure, its cloud platform (The Microsoft Cloud, 2024). The pledge should be read against the company's own subsequent environmental reporting, which has shown total emissions rising rather than falling since it was made, driven largely by the construction and operation of new datacentre capacity. That is the central tension of this whole section in miniature: efficiency commitments made in good faith can be outrun by growth in the underlying demand. To achieve this, Microsoft is investing in advanced clean energy solutions, such as battery storage and nuclear energy, to sustainably power AI data centers (Microsoft Sustainability, 2023). Additionally, Microsoft is working on optimizing AI models using techniques like Neural Architecture Search and muTransfer, which help reduce the energy needed to train and run AI systems. These methods focus on building more efficient models from the start, allowing AI to achieve high performance while minimizing carbon emissions (Microsoft Research, 2022).

Amazon Web Services (AWS), one of the largest providers of AI cloud computing, sits behind two commitments that an earlier version of this report ran together. Amazon's Climate Pledge commits the company to net-zero carbon across its operations by 2040. Separately, Amazon reports having matched its annual electricity consumption with renewable sources, a milestone it announced ahead of its own schedule (DataCentre Magazine, 2024). A target set for 2040 cannot have been met in 2024, and the two claims should not be merged. The renewable-matching claim is also market-based, resting on power purchase agreements and renewable energy certificates purchased across a year, an accounting method that has been criticised for allowing consumption in one region and hour to be offset by generation in another. AWS is also using real-time energy monitoring and workload distribution to ensure that computing resources are allocated efficiently, avoiding unnecessary energy use (About Amazon, 2024). In addition to these operational efforts, AWS is working on making its hardware more sustainable by focusing on recyclable and environmentally friendly server materials (Amazon Web Services, 2024).

Meta is working to reduce the carbon footprint of its AI systems through emissions transparency and infrastructure efficiency. Training its LLaMA 3-70B model generated an estimated 2,290 metric tons of CO₂, which Meta reports as offset by its sustainability initiatives (Hugging Face, 2024). Offset is the operative word: the tonnes were emitted, and a purchased offset is a claim that an equivalent quantity was avoided or removed elsewhere. The two are treated as equivalent in most corporate reporting and are not equivalent in the atmosphere, since offset quality, permanence and additionality are all contested. The same qualification applies to the claim of net zero operational emissions since 2020, which is a statement about scope one and scope two emissions net of renewable-energy matching and offsets, and does not cover the emissions embodied in manufacturing the hardware. Meta is also building AI-optimized data centers with liquid cooling and high energy efficiency (Meta Sustainability, 2024). It is worth noting that Meta's publication of a per-model training figure at all is unusual, and is a genuine transparency gain over competitors who publish nothing.

OpenAI, the developer of models like GPT-4 and DALL·E, operates some of the most computationally intensive AI systems in the world. However, OpenAI has not publicly released any sustainability strategy, carbon footprint disclosures, or commitments related to reducing emissions from its AI systems.

Finally, academic institutions are playing a key role in advancing sustainable AI practices. Researchers at the University of Michigan, for example, have developed training optimization techniques that could cut AI energy consumption by up to 75% (Champion, 2023). These methods focus on reducing redundant training processes and finding smarter ways to achieve high performance with less data and fewer computational cycles. At MIT, researchers created the 'Once-for-All' network, which cuts emissions by enabling one model to generate many smaller submodels without retraining — reducing training emissions by up to 1,300 times (MIT News, 2020). UMass Amherst is heading a $12 million NSF-backed initiative to cut computing emissions through integrated improvements across hardware and software systems (UMass Amherst, 2024). In addition, Yale and partner institutions are developing carbon reporting standards for machine learning to help shrink emissions across the sector (Yale School of the Environment, 2024).

The peer-reviewed literature is the appropriate counterweight to all of this, and it should be read directly rather than through the outlets that summarise it. Three studies generated most of the figures that circulate in reports of this kind. Strubell, Ganesh and McCallum (2019) produced the training-emissions numbers, including the 626,000 pound architecture-search figure so widely misattributed to training a single model. Patterson and colleagues (2021) measured the emissions of large neural network training directly and argued that the choice of model, processor, datacentre and grid together account for differences of two orders of magnitude, which is the finding that makes carbon-aware scheduling worth doing at all. Luccioni, Viguier and Ligozat (2022) estimated the footprint of the BLOOM model across its whole lifecycle, including hardware manufacture and idle consumption, and found that counting only the training run understates the total substantially. Where a corporate claim and one of these studies disagree, the study is the better authority.

Across both private companies and academic research, there is a growing recognition of the need to reduce AI's environmental impact. From optimizing model architectures and improving data center operations to designing efficient hardware and using renewable energy, these organizations are taking concrete steps to ensure that AI's future is not only powerful but also sustainable. While these efforts represent meaningful progress, continued innovation, investment in clean energy, and greater transparency will be essential to align AI's growth with global climate goals. The honest summary of this section is that the direction of travel is right, the pledges are mostly unmet by construction because their deadlines have not arrived, and the emissions curve has so far been driven more by growth in demand than bent by efficiency.

05

Regulatory Frameworks

The environmental sustainability of AI remains largely overlooked in both environmental law and technology regulations. UNEP has criticized the lack of policy focus on AI's ecological impact, noting that current efforts prioritize AI's trustworthiness rather than its environmental footprint. Hacker (2023) similarly highlights this gap and provides policy recommendations to address it. The European Union has taken steps to regulate AI through its 2024 AI Act, the world's first comprehensive AI law, which classifies AI systems by risk levels. While the European Parliament emphasized making AI systems environmentally friendly, this aspect was largely neglected in the final legislation. Earlier drafts included sustainability provisions such as sustainable AI development principles, energy logging requirements, and foundation model standards, but these were weakened during negotiations. The final Act relies on voluntary codes and limited disclosure requirements, with Article 40 mandating that general-purpose AI providers disclose estimated energy consumption. Additionally, the European Energy Efficiency Directive requires data centers exceeding 500 kW to report energy use, while Germany has imposed stricter regulations, requiring data centers with over 300 kW capacity to source 50% of their electricity from renewables by 2024, increasing to 100% by 2027.

In the United States, the Sustainable Data Centers Act, introduced in December 2024, aims to regulate the environmental footprint of data centers critical to AI operations by requiring alignment with state renewable energy targets and annual reporting of energy and water usage. On a global scale, effective AI regulation requires international cooperation, particularly concerning AI's resource-intensive supply chains and the mining of critical minerals. In February 2025, the AI Action Summit marked a turning point, with ecological sustainability taking center stage. This led to the formation of the Coalition for Environmentally Sustainable Artificial Intelligence, initiated by France, UNEP, and the International Telecommunication Union (ITU). With over 100 partners, including 11 countries, five international organizations, and 37 tech companies, the coalition aims to drive momentum toward sustainable AI practices. UNEP plans to publish a guide in 2025 to promote energy-efficient data centers based on international best practices. Additionally, AI is being leveraged for climate solutions, such as optimizing renewable energy, monitoring environmental changes, improving climate modeling, and enhancing disaster preparedness.

Several initiatives are actively advancing AI-driven climate solutions, underscoring the potential of AI in achieving sustainability goals. The WEKA Sustainable AI Initiative focuses on responsible AI applications that improve data efficiency while supporting sustainable development goals, including investments in reforestation projects to offset carbon emissions from data centers (WEKA, n.d.). The UN Climate Change's AI for Climate Action initiative explores AI's role in advancing climate-resilient and low-emissions development, particularly in developing countries, with the goal of delivering transformative climate action through AI-powered solutions (UNFCCC, n.d.). The Bezos Earth Fund's AI for Climate and Nature Initiative seeks to leverage AI to address climate change and protect natural ecosystems, including a $100 million grand challenge to unlock AI-driven solutions for climate and nature problems (Bezos Earth Fund, 2025).

While technological advancements can help reduce AI's environmental impact, policy frameworks and ethical considerations play a critical role in ensuring sustainable AI development. Effective governance, regulatory measures, and corporate responsibility initiatives are necessary to align AI's growth with long-term environmental goals.

06

Ethical & Policy Considerations

Government policies and international agreements are increasingly shaping the intersection of AI and sustainability, focusing on reducing the environmental impact of AI while promoting ethical and energy-efficient innovations. Various national strategies and global frameworks aim to balance technological advancement with ecological responsibility.

At the national level, governments are introducing AI sustainability policies to regulate energy consumption and carbon emissions. For example, the European Union's AI Act includes provisions for transparency in AI energy use, while the U.S. Department of Energy funds research on AI-driven energy efficiency. China's Green AI Initiative encourages data centers to adopt renewable energy sources and optimize power usage. Additionally, governments worldwide are promoting sustainable AI practices through tax incentives, grants, and stricter regulations on high-energy-consuming AI training processes.

Agreements like the Paris Agreement on Climate Change indirectly influence AI sustainability by urging nations to reduce carbon footprints, including those from digital infrastructures. The OECD AI Principles, adopted by over 40 countries, emphasize energy-efficient AI development and responsible deployment. The United Nations' Sustainable Development Goals (SDGs) encourage AI innovation to support climate action, particularly in renewable energy forecasting and environmental monitoring. Furthermore, the G7 and G20 summits have discussed AI's role in sustainable development, highlighting the need for global cooperation on eco-friendly AI deployment.

Despite these efforts, challenges remain in enforcing energy-efficient AI practices, as AI models grow more complex and require increasing computational power. As governments, organizations, and researchers work toward mitigating AI's environmental costs, it is essential to consider the broader implications of these efforts. Looking ahead, continued research, policy evolution, and cross-sector collaboration will be key to balancing AI's benefits with its sustainability challenges.

07

Conclusion & Future Directions

AI's rapid advancement presents both unprecedented opportunities and significant environmental challenges. As its adoption continues to grow, ensuring that AI development aligns with sustainability goals is no longer an option, but a necessity (World Economic Forum, 2024). While AI's energy consumption, reliance on resource-intensive data centers, and electronic waste generation pose serious environmental risks, innovative solutions in algorithmic efficiency, green data centers, and sustainable hardware development offer promising pathways forward (Nature, 2024; MIT News, 2025).

The future of AI sustainability depends on a multi-faceted approach that integrates technological innovation, policy intervention, and corporate responsibility. Society must prioritize energy-efficient AI architectures (Capitol Technology University, n.d.), while policymakers should implement regulations that incentivize renewable energy use and enforce transparency in AI's environmental impact (SIAM News, n.d.). Companies leading AI development must take proactive measures by investing in carbon-aware computing, energy-efficient chips, and responsible e-waste management.

Moving forward, AI should not only minimize harm but also become a key player in solving global environmental challenges. Its potential to optimize renewable energy distribution, enhance climate modeling, and improve conservation efforts demonstrates that AI can be part of the solution rather than just a contributor to the problem (AI Speakers Agency, n.d.; World Economic Forum, 2025). However, achieving this balance requires urgent and collaborative action across industries, governments, and academia. The question is no longer whether AI can be sustainable, but rather how quickly we can ensure that it is.

References

Sources

  1. 01About Amazon Staff. (2024, June 25). How AWS helps reduce carbon footprint of AI workloads. About Amazon.
  2. 02Adams, H. S. (2024, August 5). Study reveals AWS cloud can slash carbon emissions by 99% compared to on-prem data centers. DataCentre Magazine.
  3. 03AI Speakers Agency. (n.d.). How will AI enable a sustainable future?
  4. 04AIMultiple. (2025). Top 10 sustainability AI applications & real-life examples in 2025.
  5. 05Airsys. (2024). Evaporative cooling for data centers – Pros and cons.
  6. 06Bezos Earth Fund. (n.d.). AI for climate and nature initiative.
  7. 07Brightlio. (2025, April 19). 170 data center stats (April 2025).
  8. 08Capitol Technology University. (n.d.). Moving towards a more sustainable future using AI.
  9. 09Champion, Z. (2023, April 17). Optimization could cut the carbon footprint of AI training by up to 75%. University of Michigan News.
  10. 10Digital Realty. (n.d.). What is power usage effectiveness (PUE)?
  11. 11Earth.org. (2025). The real environmental impact of AI.
  12. 12European Parliament. (2023, August 6). EU AI Act: First regulation on artificial intelligence.
  13. 13Evans, R., & Gao, J. (2016, July 20). DeepMind AI reduces Google data centre cooling bill by 40%. DeepMind.
  14. 14EY. (n.d.). AI and sustainability: Opportunities, challenges, and impact.
  15. 15Forbes. (2024, March 26). Here's why data center cooling is the hottest innovation in the sector [By A. Schaap].
  16. 16Gillis, A. S. (2022, April 29). What is PUE (power usage effectiveness)? TechTarget.
  17. 17Google. (2024). Operating sustainably – Google Data Centers.
  18. 18Greenly. (n.d.). How can artificial intelligence help tackle climate change?
  19. 19Hacker, P. (2023, December 21). Sustainable AI regulation. Semantic Scholar.
  20. 20Harvard Business Review. (2024, July 15). The uneven distribution of AI's environmental impacts.
  21. 21Higgins, A. (2024, November 13). What is water usage effectiveness (WUE) in data centers? Equinix.
  22. 22Hugging Face. (2024). Meta-Llama-3-70B.
  23. 23Intel. (n.d.). Advancing toward more sustainable AI.
  24. 24Kunze, J. (2024, October 4). Environmental impacts of artificial intelligence: Peril, promise, and policy. Society for Industrial and Applied Mathematics.
  25. 25Laranjeira De Pereira, J. R. (2024, April 8). The EU AI Act and environmental protection: The case for a missed opportunity. Heinrich Böll Stiftung, Brussels Office.
  26. 26Latham & Watkins. (2024, July 12). US environmental, social, and governance legal considerations for AI companies.
  27. 27Luccioni, A. S., Viguier, S., & Ligozat, A.-L. (2022). Estimating the carbon footprint of BLOOM, a 176B parameter language model.
  28. 28Masters DC. (n.d.). How data center cooling works and why it is brilliant.
  29. 29Matheson, R. (2020, April 23). Reducing the carbon footprint of artificial intelligence. MIT News.
  30. 30McQuate, S. (2023, July 27). Q&A: UW researcher discusses just how much energy ChatGPT uses. UW News.
  31. 31Meta. (2024). Meta sustainability report 2024. Meta Sustainability.
  32. 32Microsoft. (2023). Improving sustainability with AI. Microsoft Sustainability.
  33. 33MIT News. (2025). Explained: Generative AI's environmental impact.
  34. 34National Institute of Standards and Technology. (2023). Artificial intelligence risk management framework (AI RMF 1.0) (NIST AI 100-1).
  35. 35National Institute of Standards and Technology. (2024). AI risk management framework: Generative AI profile (NIST AI 600-1).
  36. 36Nature. (2024). Ecological footprints, carbon emissions, and energy transitions.
  37. 37Patterson, D., Gonzalez, J., Le, Q., Liang, C., Munguia, L.-M., Rothchild, D., So, D., Texier, M., & Dean, J. (2021). Carbon emissions and large neural network training.
  38. 38Publications Office of the European Union. (2024). Regulation (EU) 2024/1689 of the European Parliament and of the Council.
  39. 39Robb, D. (2024, June 3). Hybrid cooling: The bridge to full liquid cooling in data centers. Data Center Knowledge.
  40. 40Russinovich, M. (2024, September 12). Sustainable by design: Innovating for energy efficiency in AI, part 1. The Microsoft Cloud Blog.
  41. 41SIAM News. (n.d.). Environmental impacts of artificial intelligence: Peril, promise, and policy.
  42. 42Sitecore. (n.d.). Harnessing the power of AI for a sustainable future.
  43. 43Soto, K. (2018, August 30). What makes TPUs fine-tuned for deep learning? Google Cloud Blog.
  44. 44Strubell, E., Ganesh, A., & McCallum, A. (2019). Energy and policy considerations for deep learning in NLP. Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, 3645–3650.
  45. 45U.S. Congress. (2024). Artificial Intelligence Environmental Impacts Act, 118th Congress (S. 3732).
  46. 46U.S. Department of Energy. (2024a). AI for energy: Opportunities for a modern grid and clean energy economy.
  47. 47U.S. Department of Energy. (2024b, April 29). DOE announces new actions to enhance America's global leadership in artificial intelligence.
  48. 48UN Environment Programme. (2025, February 11). New coalition aims to put artificial intelligence on a more sustainable path.
  49. 49UN Framework Convention on Climate Change. (n.d.). AI for climate action.
  50. 50University of California, Irvine Libraries. (n.d.). Ethics – AI in research.
  51. 51Venkatesan, V., & Karibandi, A. (2024, September 16). Embracing modernization with a sustainability focus. Amazon Web Services.
  52. 52Vertiv. (n.d.). Liquid cooling options for data centers.
  53. 53WEKA. (n.d.). Sustainable AI initiative.
  54. 54Westbrook, J. (2024, May 23). UMass researchers awarded $12M to shrink society's carbon footprint with AI & other computer science tools. UMass Amherst.
  55. 55World Economic Forum. (2024). AI will accelerate sustainability — but is no silver bullet.
  56. 56World Economic Forum. (2025). AI's role in the climate transition and how it can drive growth.
  57. 57Yale School of the Environment. (2024). New initiative focuses on reducing the carbon footprint of computer systems and AI.
  58. 58Yale School of the Environment. (2024, November 13). Can we mitigate AI's environmental impacts?

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file diff --git a/static-site/print/ai-carbon-footprint/index.txt b/static-site/print/ai-carbon-footprint/index.txt index b051f2eef..aad1f8aab 100644 --- a/static-site/print/ai-carbon-footprint/index.txt +++ b/static-site/print/ai-carbon-footprint/index.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -12:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","print","ai-carbon-footprint",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","ai-carbon-footprint","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -14:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -16:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","print","ai-carbon-footprint",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","ai-carbon-footprint","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:"$Sreact.suspense" -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1c:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -29,7 +29,7 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 19:[] 10:"$W19" 11:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1e:Ta14, @page { size: 210mm 297mm; margin: 20mm 18mm; } @@ -183,6 +183,6 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 6d:["$","li","56",{"children":[["$","span",null,{"className":"print-ref-num","children":"57"}],["$","span",null,{"children":"Yale School of the Environment. (2024). New initiative focuses on reducing the carbon footprint of computer systems and AI."}]]}] 6e:["$","li","57",{"children":[["$","span",null,{"className":"print-ref-num","children":"58"}],["$","span",null,{"children":"Yale School of the Environment. (2024, November 13). Can we mitigate AI's environmental impacts?"}]]}] 1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -75:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +75:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 18:null 1d:[["$","title","0",{"children":"AI's Carbon Footprint (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L75","6",{}]] diff --git a/static-site/print/ai-models-research/__next._full.txt b/static-site/print/ai-models-research/__next._full.txt index e3a874dfd..76417aec6 100644 --- a/static-site/print/ai-models-research/__next._full.txt +++ b/static-site/print/ai-models-research/__next._full.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -12:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","print","ai-models-research",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","ai-models-research","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -14:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -16:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","print","ai-models-research",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","ai-models-research","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:"$Sreact.suspense" -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1c:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -29,7 +29,7 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 19:[] 10:"$W19" 11:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1e:Ta14, @page { size: 210mm 297mm; margin: 20mm 18mm; } @@ -182,6 +182,6 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 62:["$","tr","6",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"GPT-5.6 Luna"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"0.10 USD"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"0.03 USD"}],["$","td","3",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"0.13 USD"}]]}] 63:["$","tr","7",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"DeepSeek V4 Pro"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"0.0435 USD"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"0.00435 USD"}],["$","td","3",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"0.04785 USD"}]]}] 1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -64:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +64:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 18:null 1d:[["$","title","0",{"children":"AI Model Performance (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L64","6",{}]] diff --git a/static-site/print/ai-models-research/__next._head.txt b/static-site/print/ai-models-research/__next._head.txt index 740a5b654..e03d4d1d4 100644 --- a/static-site/print/ai-models-research/__next._head.txt +++ b/static-site/print/ai-models-research/__next._head.txt @@ -1,6 +1,6 @@ 1:"$Sreact.fragment" -2:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -3:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +2:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +3:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 4:"$Sreact.suspense" -5:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +5:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 0:{"rsc":["$","$1","h",{"children":[null,["$","$L2",null,{"children":[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]]}],["$","div",null,{"hidden":true,"children":["$","$L3",null,{"children":["$","$4",null,{"name":"Next.Metadata","children":[["$","title","0",{"children":"AI Model Performance (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L5","6",{}]]}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/print/ai-models-research/__next._index.txt b/static-site/print/ai-models-research/__next._index.txt index f98a4fd90..18b9d82b6 100644 --- a/static-site/print/ai-models-research/__next._index.txt +++ b/static-site/print/ai-models-research/__next._index.txt @@ -1,13 +1,13 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -d:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +d:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} 8:["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$Ld",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$Ld",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$Ld",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$Ld",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] diff --git a/static-site/print/ai-models-research/__next._tree.txt b/static-site/print/ai-models-research/__next._tree.txt index 9ac00863c..c7dc9644e 100644 --- a/static-site/print/ai-models-research/__next._tree.txt +++ b/static-site/print/ai-models-research/__next._tree.txt @@ -1,7 +1,7 @@ -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] 0:{"tree":{"name":"","param":null,"prefetchHints":16,"slots":{"children":{"name":"print","param":null,"prefetchHints":0,"slots":{"children":{"name":"slug","param":{"type":"d","key":"ai-models-research","siblings":null},"prefetchHints":0,"slots":{"children":{"name":"__PAGE__","param":null,"prefetchHints":0,"slots":null}}}}}}},"staleTime":300,"buildId":"static-snapshot"} diff --git a/static-site/print/ai-models-research/__next.print.txt b/static-site/print/ai-models-research/__next.print.txt index 460afe242..fa617e997 100644 --- a/static-site/print/ai-models-research/__next.print.txt +++ b/static-site/print/ai-models-research/__next.print.txt @@ -1,4 +1,4 @@ 1:"$Sreact.fragment" -2:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -3:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +2:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +3:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] 0:{"rsc":["$","$1","c",{"children":[null,[["$","style",null,{"children":"\n /* The site is dark-first, and a dark color-scheme makes Chrome paint\n the printed sheet's margin area dark too. Paper is light. */\n html { color-scheme: light !important; }\n html, body { background: #ffffff !important; }\n body::after { content: none !important; }\n header, footer, .site-bg, .aura { display: none !important; }\n "}],["$","$L2",null,{"parallelRouterKey":"children","template":["$","$L3",null,{}]}]]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/print/ai-models-research/index.html b/static-site/print/ai-models-research/index.html index afb6952b4..aa67c68fc 100644 --- a/static-site/print/ai-models-research/index.html +++ b/static-site/print/ai-models-research/index.html @@ -1,4 +1,4 @@ -AI Model Performance (print edition) · NYU Ethical Tech CoLab

Publications · Research review

AI Model Performance

Capabilities, Accuracy, Speed, Energy Use, and Token Economics Across Frontier, Open-Weight, and Efficient Model Families

Ethical Tech CoLab

Academic and primary-source synthesis

July 2026

Carolina Morón. Prepared as a comparative research review, drawing on peer-reviewed research, institutional benchmarks, model cards, and official technical documentation. Research cut-off 22 July 2026.

There is no universally best AI model. Model selection is a constrained optimisation problem across verified task quality, factual reliability, latency, throughput, token consumption, energy per accepted output, privacy, controllability, and total deployment cost.

Key figures

0.34 Wh

median estimated energy for a representative frontier-scale query, rising to 4.32 Wh at fifteen times the token use

73%

energy reduction achievable through inference optimisation alone, against an unoptimised baseline

30%

of difficult multi-turn conversations on which the strongest web-enabled configuration tested still hallucinated

26×

spread in nominal cost for one identical document-analysis request across published price schedules

01

How to read the evidence

This review distinguishes three classes of evidence throughout, and every cross-model comparison should be read with the class in mind. Ranking claims are strongest when the same evaluator, prompt set, tool access, reasoning budget, and scoring method are held constant. They are weakest, and frequently meaningless, when they are not.

GradeEvidence classInterpretation
AIndependent or peer-reviewedPeer-reviewed papers, reproducible benchmarks, standardised institutional evaluations, and public datasets.
BInstitutional primary researchUniversity or research-lab reports, model cards, benchmark methodology pages, and public technical reports.
CProvider-reportedVendor launch benchmarks, self-reported latency, and product documentation. Useful, but not independently reproduced by default.
The three evidence grades used throughout the review and in the underlying research repository.

The distinction matters because the three are routinely printed in the same table, to the same number of decimal places, as though they were the same kind of fact. A provider launch table is produced by the party that benefits from the result, under conditions the provider chose and usually did not disclose. A benchmark maintainer's evaluation is produced under a published protocol by a party with no stake in which model wins.

Interpretation rule. Treat vendor benchmark numbers as hypotheses for local testing. They are useful for narrowing a candidate set and are not sufficient for procurement.

02

The 2026 landscape

The frontier shifted from a single scale race toward a multi-dimensional competition among reasoning quality, agentic execution, token efficiency, multimodality, long-context handling, and operational efficiency. By mid-2026, leading proprietary systems commonly offer configurable reasoning, tool use, image input, context windows near one million tokens, and output limits above 100,000 tokens.

Open-weight systems increasingly use mixture-of-experts architectures, sparse attention, quantization, and smaller active parameter counts to approach frontier performance with lower serving requirements. The market now contains three overlapping classes: frontier hosted models optimising broad capability and tool integration, efficient hosted models optimising latency and price, and open-weight models optimising control and deployment flexibility.

Parameter count is an incomplete proxy for inference cost. A mixture-of-experts model stores many parameters but activates a smaller subset for each token. That reduces compute per token relative to a dense model of the same stored size, but it does not eliminate memory, routing, communication, or serving overhead. Active parameter count should be reported alongside total parameters, precision, hardware, batch size, and measured throughput.

Mistral Medium 3.5100 percent

128B active of 128B stored, dense

Mistral Small 4about 5 percent

6B active of 119B stored

DeepSeek V4 Flashabout 5 percent

13B active of 284B stored

Llama 4 Maverickabout 4 percent

17B active of 400B stored

DeepSeek V4 Proabout 3 percent

49B active of 1.6T stored

Share of stored parameters activated per token, from disclosed architectures. Provider architecture disclosures, Grade B. The dense model activates everything it stores; the mixture-of-experts models activate a twentieth or less, which is why stored size alone predicts neither compute nor cost.
Data table
ModelTotal storedActive per tokenShare
Mistral Medium 3.5128B, dense128B100 percent
Mistral Small 4119B6Babout 5 percent
DeepSeek V4 Flash284B13Babout 5 percent
Llama 4 Maverick400B17Babout 4 percent
DeepSeek V4 Pro1.6T49Babout 3 percent

03

What performance actually means

Performance is not a scalar property. The holistic evaluation framework argued for assessment across accuracy, calibration, robustness, fairness, bias, toxicity, and efficiency rather than a single number. That argument is stronger now that models use tools, accept several modalities, and vary their reasoning effort per request, because one score hides differences in reliability, cost, and behaviour that determine whether a system works in production.

Rankings disagree because evaluations change the computational and informational conditions under which a model operates:

  • A model tested with web search and code execution is not comparable to a closed-book model.
  • A model allowed maximum reasoning effort, or several samples, has a larger inference budget than a model tested once at default settings.
  • Agent benchmarks are sensitive to the scaffolding that manages tools, context, retries, and termination, so the result belongs to a model-and-scaffold pair rather than to a model.
  • Prompt formulation alters scores and can reorder models, especially on open-ended and instruction-following tasks.
  • Contamination inflates performance when evaluation items or near variants appear in training data.
  • Judging by another model introduces judge bias, positional effects, verbosity preference, and family favouritism.
  • Pass-at-k and majority voting spend more compute than single-attempt evaluation and are not equivalent to it.
  • Provider tables mix internal tasks with public benchmarks and may evaluate competitor models under the publishing provider's own harness.

Benchmark progress through 2026 is real but uneven. Models approach saturation on some mathematics and knowledge tests while remaining materially weaker on application construction, long-horizon agents, and open-world computer use.

  • Best reported model
  • Human baseline
OSWorld66.3 percent vs 72.35 percent human

Computer use in realistic desktop environments

WebArena74.3 percent vs 78.24 percent human

Web navigation and task completion

Terminal-Bench 2.077.3 percent no human baseline reported

Terminal-based coding and system tasks

Vibe Code Benchabout 57.6 percent no human baseline reported

End-to-end application construction

Best reported model results against human baselines on interactive benchmarks. Stanford HAI, AI Index Report 2026, Technical Performance chapter. Grade B. Snapshots, not permanent rankings. Human scores are frequently measured under different time, tool, and incentive conditions than model scores, so the distance between the two dots is indicative rather than exact.
Data table
BenchmarkCapabilityBest reportedHuman baseline
OSWorldComputer use in realistic desktop environments66.3 percent72.35 percent
WebArenaWeb navigation and task completion74.3 percent78.24 percent
Terminal-Bench 2.0Terminal-based coding and system tasks77.3 percentNot reported
Vibe Code BenchEnd-to-end application constructionAbout 57.6 percentNot reported

Two readings follow. Agents approach human performance on these benchmarks without consistently exceeding it. And building a functioning application remains materially harder than completing an isolated task, which is what a long-horizon workload actually requires. Human scores are frequently measured under different time, tool, and incentive conditions than model scores, so the gap is indicative rather than exact.

04

Accuracy, factuality, and hallucination

Accuracy cannot be represented by one score. A model may excel at scientific multiple choice and still fabricate citations, invent software packages, omit constraints in a long document, or select the wrong tool in an agent loop. These are four different failure modes with four different measurements.

TypeQuestionHow it is evaluated
Parametric factualityIs the answer correct from internal model knowledge?Closed-book factual question answering
GroundednessDoes every material claim follow from the supplied sources?Claim-level entailment against provided evidence
Procedural correctnessDid the model follow the required method and constraints?Schema validation, unit tests, workflow checks
Epistemic behaviourDoes the model recognise when evidence is insufficient?Calibration, selective accuracy, abstention
Four distinct accuracy problems, each requiring its own evaluation.

HalluHard evaluates 950 difficult multi-turn conversations in legal, research, medical, and coding settings, and checks whether cited material actually supports the generated claims. The strongest web-enabled configuration tested still hallucinated on about 30 percent of conversations, and rates without web access were substantially higher. The same work reports that early errors cascade as a conversation lengthens, which makes multi-turn reliability a property of the system rather than of a single answer.

Software generation carries a distinct and concrete risk. Research at NYU's OSIRIS Lab on package hallucination, in which a model recommends dependencies that do not exist, found invented package names in roughly 4.6 to 6.1 percent of tested package suggestions across frontier systems, with some fabricated names recurring across models. That is a supply-chain exposure rather than a quality nuisance: a name that several models hallucinate can be registered by an attacker and then installed by automation.

Fluency is not calibration. Confidence can be expressed through verbal hedging, an explicit probability, token log-probabilities, agreement across samples, or a separate verifier, and these signals do not always agree with one another. Uncertainty estimators can correlate weakly with actual hallucination depending on the error type and the model, so a production system should validate its uncertainty signal against its own domain rather than assume a confident model is right.

Reliability is a property of the system, not of the model choice. The controls that matter more than selection are:

  • Retrieval with provenance: retrieve a small evidence set, preserve source identifiers, and require claim-level citation.
  • Constrained generation: schemas, grammars, enumerated values, and deterministic validators wherever possible.
  • Verification: run the calculation, the code, the query, or the citation check against the first output.
  • Selective escalation: route low-confidence, high-risk, or contradictory cases to a stronger model or a human reviewer.
  • Abstention policy: define when the system must say the evidence is insufficient, and what would resolve it.
  • Version control: pin model snapshots and rerun regression evaluations before an upgrade.
  • Adversarial evaluation: test prompt injection, conflicting documents, missing evidence, and long-context distractors.

05

Speed is not generation rate

Speed must be decomposed into queueing time, prefill, time to first token, time per output token, tool latency, and total task completion time. A model that streams at a high token rate can still be slow if it emits excessive reasoning tokens or requires repeated attempts. Conversely, a slower frontier model can finish sooner if it needs fewer steps, fewer tool calls, and less rework.

Published throughput figures illustrate the problem rather than solving it. Google reports 363 output tokens per second for Gemini 3.1 Flash-Lite in its own evaluation. xAI reports 80 tokens per second for Grok 4.5 on its launch page. These come from different test environments and different model classes, are both provider-reported, and are not a controlled head-to-head comparison. They show that fast tiers exist and are positioned for interactive and high-volume use. They do not measure how long a task takes.

Configurable reasoning moves the frontier rather than sitting on it. Low or disabled reasoning suits extraction, rewriting, classification, and deterministic tool routing. High reasoning effort earns its cost on mathematics, scientific analysis, difficult coding, and planning, where an incorrect result costs more than the additional compute. Research on test-time compute identifies diminishing returns and overthinking, in which longer chains introduce new errors, lose the objective, or exhaust the context budget. The correct stopping rule is quality-conditioned rather than a fixed token allowance.

Operational reporting requirements follow from this:

  • Report p50, p90, and p95 time to first token, never a mean alone.
  • Report p50 and p95 end-to-end completion time, including tools and retries.
  • Separate cold starts from warm requests, and cache hits from cache misses.
  • Compare models at the same latency and cost budget, not at their own defaults.
  • Track output tokens, tool calls, failed steps, and retries per accepted task.
  • Test at the concurrency you expect, because batching raises throughput while raising individual latency.

06

Context windows and token economics

One-million-token windows are now common in leading hosted families, but nominal capacity exceeds reliable capacity. Stanford's long-context evaluation states that support for long inputs does not imply strong long-context capability. LongCodeBench finds degradation on real code tasks as context scales toward one million tokens. LongProc finds that models accept long inputs yet lose coherence when they must integrate dispersed information and produce structured output over thousands of tokens.

Failure modeDescription
Lost evidenceRelevant information is overlooked among distractors.
Position biasMaterial at the start and end is used more reliably than material in the middle.
Aggregation failureIndividual facts are extracted correctly but combined incorrectly.
Instruction decayConstraints stated early are forgotten during long generation.
Context pollutionStale tool results, abandoned plans, and irrelevant documents interfere with current decisions.
Cost explosionLarge prompts raise prefill latency, memory use, and input charges.
Long-context failure modes. A retrieval probe measures whether a single fact can be located; it does not measure whether dispersed facts can be reasoned over.

Nominal per-token price is not total cost. Output tokens are frequently priced three to six times above input tokens, and reasoning workflows can generate large hidden or visible outputs. Tokenizer changes move cost without moving price: Anthropic notes that Claude Sonnet 5 uses a tokenizer that can produce about 30 percent more tokens for the same text than its predecessor, so an unchanged per-token rate still raises the cost of an equivalent request.

  • Input tokens
  • Output tokens
Claude Fable 51.25 USD
GPT-5.6 Sol0.65 USD
Claude Opus 4.80.625 USD
Claude Sonnet 50.375 USD
GPT-5.6 Terra0.325 USD
Grok 4.50.23 USD
GPT-5.6 Luna0.13 USD
DeepSeek V4 Pro0.04785 USD
Nominal cost of one identical document-analysis request: 100,000 uncached input tokens and 5,000 output tokens. Computed from public prices on 22 July 2026, provider-reported. Excludes tools, long-context surcharges, caching, batch discounts, and tax. Quality is not held constant, so the bars compare price for an identical request, not price for an identical result.
Data table
ModelInput costOutput costNominal total
Claude Fable 51.00 USD0.25 USD1.25 USD
GPT-5.6 Sol0.50 USD0.15 USD0.65 USD
Claude Opus 4.80.50 USD0.125 USD0.625 USD
Claude Sonnet 50.30 USD0.075 USD0.375 USD
GPT-5.6 Terra0.25 USD0.075 USD0.325 USD
Grok 4.50.20 USD0.03 USD0.23 USD
GPT-5.6 Luna0.10 USD0.03 USD0.13 USD
DeepSeek V4 Pro0.0435 USD0.00435 USD0.04785 USD

The decision-relevant quantity is cost per accepted task. A cheap model becomes expensive when it requires repeated prompts, long outputs, human correction, or escalation. A high-priced model can be economical when it succeeds on the first attempt and uses fewer tools and fewer tokens. An accepted task is one that passes a domain-specific quality check without manual correction beyond a defined tolerance, and the acceptance criterion must be stated wherever the figure is used.

07

Energy and environmental performance

Training energy is a large, episodic expenditure. Inference energy is distributed across every production request. Neither is interpretable without a stated system boundary covering direct accelerator energy, host and network energy, facility overhead, embodied impacts, and water and grid effects. Two correctly computed figures for the same workload can differ by a large factor because they draw that boundary differently.

Fernandez and colleagues evaluated inference across workloads, hardware, serving frameworks, batching, decoding strategies, and parallelism. They report that estimates derived from floating-point operation counts understate real energy use, and that appropriate combinations of optimisations reduced energy by as much as 73 percent against an unoptimised baseline. That is the strongest argument against attaching a single energy number to a model name: the same weights on the same hardware under two serving configurations are two different energy propositions.

Oviedo and colleagues developed a bottom-up estimate for frontier-scale inference. Under stated H100 utilisation and power usage effectiveness assumptions, the median estimate was 0.34 watt-hours per representative query, with an interquartile range of 0.18 to 0.67 watt-hours. Raising token use by a factor of 15 for test-time scaling raised the median to 4.32 watt-hours, about 13 times higher. Combined model, serving, and hardware improvements were estimated to offer an 8 to 20 times efficiency opportunity. These are analytical estimates, not measurements of a named commercial service.

  • Interquartile range of the estimate
  • Median estimate
Representative query0.34 Wh

Interquartile range 0.18 to 0.67 watt-hours

Same query at fifteen times the token use4.32 Wh

About 13 times the median of the representative query. No interquartile range reported for this condition.

Estimated inference energy for one representative frontier-scale query, and for the same query under test-time scaling. Oviedo and colleagues, bottom-up analytical estimate under stated H100 utilisation and power usage effectiveness assumptions. Grade B. The band is the interquartile range of the estimate, not a measurement interval, and the scaled condition is a single reported median. Nothing here is a measurement of a named commercial service.
Data table
ConditionMedian estimateInterquartile range
Representative query0.34 Wh0.18 to 0.67 Wh
Same query at fifteen times the token use4.32 WhNot reported

Do not divide sector totals by a presumed query count. The International Energy Agency projects global data-centre electricity consumption reaching about 945 terawatt-hours in 2030 in its base case, roughly double the 2024 level, and reports that data-centre electricity demand grew 17 percent in 2025 with AI-focused facilities growing faster. That aggregate covers every data-centre workload, not AI queries alone, and utilisation patterns differ across them.

A credible energy disclosure identifies all of the following. A figure missing any of them is not comparable to another figure:

  • Model version and hardware
  • Numerical precision
  • Prompt and output length distributions
  • Batch size and utilisation
  • Serving software stack
  • Power usage effectiveness
  • Measurement or estimation method, and the measurement interval
  • The quality threshold at which the work counts as useful

The central environmental variable is therefore not model size. It is the energy required to produce an accepted, useful result at the required service level. Energy spent on an output that fails a quality check is energy spent for nothing.

08

A selection framework

Selection starts from failure cost, not from capability. What a wrong answer costs sets the control posture, and the control posture constrains the candidate set before any benchmark is consulted.

Risk tierExamplesMinimum control posture
LowDrafting, brainstorming, rewritingFast low-cost model, light validation
ModerateInternal summarisation, extraction, analytics assistanceSchema validation, citations, sampled human review
HighPublic research, financial analysis, legal support, code deploymentStrong model, source verification, tests, approvals
CriticalClinical decisions, autonomous transactions, safety systemsNarrow validated system, expert oversight, formal governance
Risk tiers and the minimum control posture each requires.

The right system is frequently not a single model. Six architecture patterns recur:

  • Cascade: a small model handles easy cases and a frontier model handles escalations, for high-volume mixed-difficulty workloads.
  • Router: a classifier selects model, tools, and reasoning budget, for multiple task types under a cost constraint.
  • Generator and verifier: one model produces and another checks claims or execution, where error cost is high and outputs are verifiable.
  • Retrieval-grounded: the model answers only from a controlled evidence set, for research, policy, legal, and enterprise knowledge.
  • Human in the loop: a person approves uncertain or consequential actions, for high-stakes decisions and transactions.
  • Local plus cloud: a private local model handles sensitive data and a cloud model handles de-identified hard tasks, under sovereignty and privacy constraints.

No model is selected on a published benchmark. The benchmark narrows the candidate set; a local evaluation selects. Define the task contract, build a representative evaluation set including adversarial and multilingual cases, hold prompts and tool access constant, measure end-to-end outcomes rather than single answers, inspect variance rather than the mean, and pin versions before every upgrade.

09

Research gaps and outlook

GapWhy it matters
Comparable energy disclosureNamed commercial models rarely publish measured joules per token under a standardised workload.
Agent reliabilityBenchmarks do not capture multi-day workflows, changing environments, permissions, and recovery.
Evaluation contaminationPublic benchmarks become training targets and lose diagnostic value.
Effective long contextBetter tests are needed for dispersed reasoning, long outputs, memory, and context compaction.
Multilingual equityTokenization, cost, accuracy, and safety remain uneven across languages and scripts.
Model updatesEndpoint behaviour changes without a new public model name, defeating reproducibility.
Human baselinesHuman scores are measured under different time, tool, and incentive conditions than model scores.
Environmental systems accountingEmbodied carbon, water, grid constraints, and rebound effects remain underreported.
Gaps that the current evidence base cannot close, and why each matters.

The likely near-term direction is adaptive systems that vary model size, reasoning budget, context, and tool use by task difficulty. That improves quality per unit of cost and energy while creating a new evaluation problem: the object being evaluated becomes a dynamic policy rather than a single model. Future benchmarks should report the complete resource budget, including hidden reasoning, retries, verification, and tool calls.

Open-weight models will continue to narrow capability gaps while supporting sovereign and specialised deployments. Their advantage will depend less on raw benchmark parity than on the ability to optimise the full stack, including quantization, retrieval, domain fine-tuning, caching, and hardware. Hosted frontier models will retain advantages in integrated tools, rapid updates, and breadth, and buyers should demand stronger transparency and version stability in exchange.

Conclusion. AI model selection should be treated as empirical engineering and governance, not brand preference. The right system is the least resource-intensive configuration that reliably meets the task contract and the risk threshold. This requires local evaluation, explicit evidence standards, controlled reasoning budgets, token and energy measurement, and continuous monitoring. Frontier capability is valuable, but unverified capability is not reliability.

10

Scope and limitations

This is a structured narrative review rather than a formal meta-analysis. A meta-analysis would mislead, because benchmark definitions, model versions, prompts, inference budgets, and access conditions change faster than a stable pooled effect can be estimated. The review therefore reports ranges, mechanisms, and measurement cautions rather than forcing every model into a composite score.

The unit of analysis is the deployed model system rather than the neural network alone. A deployed system may include a tokenizer, a routing layer, a hidden reasoning process, search, code execution, retrieval, prompt caching, safety classifiers, and agent orchestration. Two systems built on identical weights can differ in accuracy, latency, cost, and failure mode.

Prices and rankings change without notice. Model aliases, context limits, hosted features, and price schedules in this review are accurate to the research cut-off of 22 July 2026 and should be rechecked before procurement or production deployment. Nothing here is legal, financial, or procurement advice.

The accompanying research repository carries the full source register, the extraction protocol, the JSON schemas, and the validation tooling that enforce these rules mechanically. Comparison tables there are generated from a machine-readable data layer rather than written by hand, so no figure can appear without a source behind it.

References

Sources

  1. 01Stanford Institute for Human-Centered Artificial Intelligence (2026). AI Index Report 2026. Grade B.
  2. 02Stanford Center for Research on Foundation Models (2026). HELM Capabilities. Grade A.
  3. 03Stanford Center for Research on Foundation Models (2025). HELM Long Context. Grade A.
  4. 04Liang, P. and others (2022). Holistic evaluation of language models. arXiv:2211.09110. Grade B.
  5. 05Fernandez, J., Na, C., Tiwari, V., Bisk, Y., Luccioni, S. and Strubell, E. (2025). Energy considerations of large language model inference and efficiency optimizations. Proceedings of ACL 2025. Grade A.
  6. 06Oviedo, F., Kazhamiaka, F., Choukse, E., Kim, A., Luers, A., Nakagawa, M., Bianchini, R. and Lavista Ferres, J. M. (2026). Energy use of AI inference: efficiency pathways and test-time compute. Joule. Grade A.
  7. 07Wilhelm, P., Wittkopp, T. and Kao, O. (2025). Beyond test-time compute strategies: advocating energy-per-token in LLM inference. EuroMLSys 2025. Grade A.
  8. 08Jin, Y., Wei, G.-Y. and Brooks, D. (2025). The energy cost of reasoning: analyzing energy usage in LLMs with test-time compute. arXiv:2505.14733. Grade B.
  9. 09Fan, D., Delsad, S., Flammarion, N. and Andriushchenko, M. (2026). HalluHard: a hard multi-turn hallucination benchmark. arXiv:2602.01031. Grade B.
  10. 10NYU OSIRIS Lab (2026). LLM package hallucination research. Grade B.
  11. 11Google DeepMind (2024). FACTS Grounding: a benchmark for evaluating factuality. Grade B.
  12. 12Rando, S. and others (2025). LongCodeBench: evaluating coding LLMs at 1M context windows. arXiv:2505.07897. Grade B.
  13. 13Ye, X., Yin, F., He, Y., Zhang, J., Yen, H., Gao, T., Durrett, G. and Chen, D. (2025). LongProc: benchmarking long-context language models on long procedural generation. arXiv:2501.05414. Grade B.
  14. 14Asirvatham, H. and others (2026). GPT as a measurement tool. Harvard University. Grade B.
  15. 15International Energy Agency (2025). Energy and AI: energy demand from AI. Grade B.
  16. 16International Energy Agency (2026). Key questions on energy and AI. Grade B.
  17. 17Anthropic (2026). Models overview and Claude Sonnet 5 pricing notes. Claude Platform Documentation. Grade B for commercial terms.
  18. 18OpenAI (2026). GPT-5.6 model documentation. Grade B for commercial terms, Grade C for performance claims.
  19. 19Google DeepMind (2026). Gemini 3.1 Pro, and the Gemini 3.1 Flash-Lite model card. Grade C for performance claims.
  20. 20xAI (2026). Introducing Grok 4.5, and the Grok 4.5 model documentation. Grade C for performance claims.
  21. 21DeepSeek (2026). DeepSeek V4 preview release, and models and pricing. Grade B for architecture and commercial terms.
  22. 22Mistral AI (2026). Mistral Medium 3.5 and remote agents, and Introducing Mistral Small 4. Grade B.
  23. 23Meta (2025). The Llama 4 herd: natively multimodal open-weight models. Grade B.
  24. 24Bench360 authors (2025). Bench360: benchmarking local LLM inference from 360 degrees. arXiv:2511.16682. Grade B.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file +

Publications · Research review

AI Model Performance

Capabilities, Accuracy, Speed, Energy Use, and Token Economics Across Frontier, Open-Weight, and Efficient Model Families

Ethical Tech CoLab

Academic and primary-source synthesis

July 2026

Carolina Morón. Prepared as a comparative research review, drawing on peer-reviewed research, institutional benchmarks, model cards, and official technical documentation. Research cut-off 22 July 2026.

There is no universally best AI model. Model selection is a constrained optimisation problem across verified task quality, factual reliability, latency, throughput, token consumption, energy per accepted output, privacy, controllability, and total deployment cost.

Key figures

0.34 Wh

median estimated energy for a representative frontier-scale query, rising to 4.32 Wh at fifteen times the token use

73%

energy reduction achievable through inference optimisation alone, against an unoptimised baseline

30%

of difficult multi-turn conversations on which the strongest web-enabled configuration tested still hallucinated

26×

spread in nominal cost for one identical document-analysis request across published price schedules

01

How to read the evidence

This review distinguishes three classes of evidence throughout, and every cross-model comparison should be read with the class in mind. Ranking claims are strongest when the same evaluator, prompt set, tool access, reasoning budget, and scoring method are held constant. They are weakest, and frequently meaningless, when they are not.

GradeEvidence classInterpretation
AIndependent or peer-reviewedPeer-reviewed papers, reproducible benchmarks, standardised institutional evaluations, and public datasets.
BInstitutional primary researchUniversity or research-lab reports, model cards, benchmark methodology pages, and public technical reports.
CProvider-reportedVendor launch benchmarks, self-reported latency, and product documentation. Useful, but not independently reproduced by default.
The three evidence grades used throughout the review and in the underlying research repository.

The distinction matters because the three are routinely printed in the same table, to the same number of decimal places, as though they were the same kind of fact. A provider launch table is produced by the party that benefits from the result, under conditions the provider chose and usually did not disclose. A benchmark maintainer's evaluation is produced under a published protocol by a party with no stake in which model wins.

Interpretation rule. Treat vendor benchmark numbers as hypotheses for local testing. They are useful for narrowing a candidate set and are not sufficient for procurement.

02

The 2026 landscape

The frontier shifted from a single scale race toward a multi-dimensional competition among reasoning quality, agentic execution, token efficiency, multimodality, long-context handling, and operational efficiency. By mid-2026, leading proprietary systems commonly offer configurable reasoning, tool use, image input, context windows near one million tokens, and output limits above 100,000 tokens.

Open-weight systems increasingly use mixture-of-experts architectures, sparse attention, quantization, and smaller active parameter counts to approach frontier performance with lower serving requirements. The market now contains three overlapping classes: frontier hosted models optimising broad capability and tool integration, efficient hosted models optimising latency and price, and open-weight models optimising control and deployment flexibility.

Parameter count is an incomplete proxy for inference cost. A mixture-of-experts model stores many parameters but activates a smaller subset for each token. That reduces compute per token relative to a dense model of the same stored size, but it does not eliminate memory, routing, communication, or serving overhead. Active parameter count should be reported alongside total parameters, precision, hardware, batch size, and measured throughput.

Mistral Medium 3.5100 percent

128B active of 128B stored, dense

Mistral Small 4about 5 percent

6B active of 119B stored

DeepSeek V4 Flashabout 5 percent

13B active of 284B stored

Llama 4 Maverickabout 4 percent

17B active of 400B stored

DeepSeek V4 Proabout 3 percent

49B active of 1.6T stored

Share of stored parameters activated per token, from disclosed architectures. Provider architecture disclosures, Grade B. The dense model activates everything it stores; the mixture-of-experts models activate a twentieth or less, which is why stored size alone predicts neither compute nor cost.
Data table
ModelTotal storedActive per tokenShare
Mistral Medium 3.5128B, dense128B100 percent
Mistral Small 4119B6Babout 5 percent
DeepSeek V4 Flash284B13Babout 5 percent
Llama 4 Maverick400B17Babout 4 percent
DeepSeek V4 Pro1.6T49Babout 3 percent

03

What performance actually means

Performance is not a scalar property. The holistic evaluation framework argued for assessment across accuracy, calibration, robustness, fairness, bias, toxicity, and efficiency rather than a single number. That argument is stronger now that models use tools, accept several modalities, and vary their reasoning effort per request, because one score hides differences in reliability, cost, and behaviour that determine whether a system works in production.

Rankings disagree because evaluations change the computational and informational conditions under which a model operates:

  • A model tested with web search and code execution is not comparable to a closed-book model.
  • A model allowed maximum reasoning effort, or several samples, has a larger inference budget than a model tested once at default settings.
  • Agent benchmarks are sensitive to the scaffolding that manages tools, context, retries, and termination, so the result belongs to a model-and-scaffold pair rather than to a model.
  • Prompt formulation alters scores and can reorder models, especially on open-ended and instruction-following tasks.
  • Contamination inflates performance when evaluation items or near variants appear in training data.
  • Judging by another model introduces judge bias, positional effects, verbosity preference, and family favouritism.
  • Pass-at-k and majority voting spend more compute than single-attempt evaluation and are not equivalent to it.
  • Provider tables mix internal tasks with public benchmarks and may evaluate competitor models under the publishing provider's own harness.

Benchmark progress through 2026 is real but uneven. Models approach saturation on some mathematics and knowledge tests while remaining materially weaker on application construction, long-horizon agents, and open-world computer use.

  • Best reported model
  • Human baseline
OSWorld66.3 percent vs 72.35 percent human

Computer use in realistic desktop environments

WebArena74.3 percent vs 78.24 percent human

Web navigation and task completion

Terminal-Bench 2.077.3 percent no human baseline reported

Terminal-based coding and system tasks

Vibe Code Benchabout 57.6 percent no human baseline reported

End-to-end application construction

Best reported model results against human baselines on interactive benchmarks. Stanford HAI, AI Index Report 2026, Technical Performance chapter. Grade B. Snapshots, not permanent rankings. Human scores are frequently measured under different time, tool, and incentive conditions than model scores, so the distance between the two dots is indicative rather than exact.
Data table
BenchmarkCapabilityBest reportedHuman baseline
OSWorldComputer use in realistic desktop environments66.3 percent72.35 percent
WebArenaWeb navigation and task completion74.3 percent78.24 percent
Terminal-Bench 2.0Terminal-based coding and system tasks77.3 percentNot reported
Vibe Code BenchEnd-to-end application constructionAbout 57.6 percentNot reported

Two readings follow. Agents approach human performance on these benchmarks without consistently exceeding it. And building a functioning application remains materially harder than completing an isolated task, which is what a long-horizon workload actually requires. Human scores are frequently measured under different time, tool, and incentive conditions than model scores, so the gap is indicative rather than exact.

04

Accuracy, factuality, and hallucination

Accuracy cannot be represented by one score. A model may excel at scientific multiple choice and still fabricate citations, invent software packages, omit constraints in a long document, or select the wrong tool in an agent loop. These are four different failure modes with four different measurements.

TypeQuestionHow it is evaluated
Parametric factualityIs the answer correct from internal model knowledge?Closed-book factual question answering
GroundednessDoes every material claim follow from the supplied sources?Claim-level entailment against provided evidence
Procedural correctnessDid the model follow the required method and constraints?Schema validation, unit tests, workflow checks
Epistemic behaviourDoes the model recognise when evidence is insufficient?Calibration, selective accuracy, abstention
Four distinct accuracy problems, each requiring its own evaluation.

HalluHard evaluates 950 difficult multi-turn conversations in legal, research, medical, and coding settings, and checks whether cited material actually supports the generated claims. The strongest web-enabled configuration tested still hallucinated on about 30 percent of conversations, and rates without web access were substantially higher. The same work reports that early errors cascade as a conversation lengthens, which makes multi-turn reliability a property of the system rather than of a single answer.

Software generation carries a distinct and concrete risk. Research at NYU's OSIRIS Lab on package hallucination, in which a model recommends dependencies that do not exist, found invented package names in roughly 4.6 to 6.1 percent of tested package suggestions across frontier systems, with some fabricated names recurring across models. That is a supply-chain exposure rather than a quality nuisance: a name that several models hallucinate can be registered by an attacker and then installed by automation.

Fluency is not calibration. Confidence can be expressed through verbal hedging, an explicit probability, token log-probabilities, agreement across samples, or a separate verifier, and these signals do not always agree with one another. Uncertainty estimators can correlate weakly with actual hallucination depending on the error type and the model, so a production system should validate its uncertainty signal against its own domain rather than assume a confident model is right.

Reliability is a property of the system, not of the model choice. The controls that matter more than selection are:

  • Retrieval with provenance: retrieve a small evidence set, preserve source identifiers, and require claim-level citation.
  • Constrained generation: schemas, grammars, enumerated values, and deterministic validators wherever possible.
  • Verification: run the calculation, the code, the query, or the citation check against the first output.
  • Selective escalation: route low-confidence, high-risk, or contradictory cases to a stronger model or a human reviewer.
  • Abstention policy: define when the system must say the evidence is insufficient, and what would resolve it.
  • Version control: pin model snapshots and rerun regression evaluations before an upgrade.
  • Adversarial evaluation: test prompt injection, conflicting documents, missing evidence, and long-context distractors.

05

Speed is not generation rate

Speed must be decomposed into queueing time, prefill, time to first token, time per output token, tool latency, and total task completion time. A model that streams at a high token rate can still be slow if it emits excessive reasoning tokens or requires repeated attempts. Conversely, a slower frontier model can finish sooner if it needs fewer steps, fewer tool calls, and less rework.

Published throughput figures illustrate the problem rather than solving it. Google reports 363 output tokens per second for Gemini 3.1 Flash-Lite in its own evaluation. xAI reports 80 tokens per second for Grok 4.5 on its launch page. These come from different test environments and different model classes, are both provider-reported, and are not a controlled head-to-head comparison. They show that fast tiers exist and are positioned for interactive and high-volume use. They do not measure how long a task takes.

Configurable reasoning moves the frontier rather than sitting on it. Low or disabled reasoning suits extraction, rewriting, classification, and deterministic tool routing. High reasoning effort earns its cost on mathematics, scientific analysis, difficult coding, and planning, where an incorrect result costs more than the additional compute. Research on test-time compute identifies diminishing returns and overthinking, in which longer chains introduce new errors, lose the objective, or exhaust the context budget. The correct stopping rule is quality-conditioned rather than a fixed token allowance.

Operational reporting requirements follow from this:

  • Report p50, p90, and p95 time to first token, never a mean alone.
  • Report p50 and p95 end-to-end completion time, including tools and retries.
  • Separate cold starts from warm requests, and cache hits from cache misses.
  • Compare models at the same latency and cost budget, not at their own defaults.
  • Track output tokens, tool calls, failed steps, and retries per accepted task.
  • Test at the concurrency you expect, because batching raises throughput while raising individual latency.

06

Context windows and token economics

One-million-token windows are now common in leading hosted families, but nominal capacity exceeds reliable capacity. Stanford's long-context evaluation states that support for long inputs does not imply strong long-context capability. LongCodeBench finds degradation on real code tasks as context scales toward one million tokens. LongProc finds that models accept long inputs yet lose coherence when they must integrate dispersed information and produce structured output over thousands of tokens.

Failure modeDescription
Lost evidenceRelevant information is overlooked among distractors.
Position biasMaterial at the start and end is used more reliably than material in the middle.
Aggregation failureIndividual facts are extracted correctly but combined incorrectly.
Instruction decayConstraints stated early are forgotten during long generation.
Context pollutionStale tool results, abandoned plans, and irrelevant documents interfere with current decisions.
Cost explosionLarge prompts raise prefill latency, memory use, and input charges.
Long-context failure modes. A retrieval probe measures whether a single fact can be located; it does not measure whether dispersed facts can be reasoned over.

Nominal per-token price is not total cost. Output tokens are frequently priced three to six times above input tokens, and reasoning workflows can generate large hidden or visible outputs. Tokenizer changes move cost without moving price: Anthropic notes that Claude Sonnet 5 uses a tokenizer that can produce about 30 percent more tokens for the same text than its predecessor, so an unchanged per-token rate still raises the cost of an equivalent request.

  • Input tokens
  • Output tokens
Claude Fable 51.25 USD
GPT-5.6 Sol0.65 USD
Claude Opus 4.80.625 USD
Claude Sonnet 50.375 USD
GPT-5.6 Terra0.325 USD
Grok 4.50.23 USD
GPT-5.6 Luna0.13 USD
DeepSeek V4 Pro0.04785 USD
Nominal cost of one identical document-analysis request: 100,000 uncached input tokens and 5,000 output tokens. Computed from public prices on 22 July 2026, provider-reported. Excludes tools, long-context surcharges, caching, batch discounts, and tax. Quality is not held constant, so the bars compare price for an identical request, not price for an identical result.
Data table
ModelInput costOutput costNominal total
Claude Fable 51.00 USD0.25 USD1.25 USD
GPT-5.6 Sol0.50 USD0.15 USD0.65 USD
Claude Opus 4.80.50 USD0.125 USD0.625 USD
Claude Sonnet 50.30 USD0.075 USD0.375 USD
GPT-5.6 Terra0.25 USD0.075 USD0.325 USD
Grok 4.50.20 USD0.03 USD0.23 USD
GPT-5.6 Luna0.10 USD0.03 USD0.13 USD
DeepSeek V4 Pro0.0435 USD0.00435 USD0.04785 USD

The decision-relevant quantity is cost per accepted task. A cheap model becomes expensive when it requires repeated prompts, long outputs, human correction, or escalation. A high-priced model can be economical when it succeeds on the first attempt and uses fewer tools and fewer tokens. An accepted task is one that passes a domain-specific quality check without manual correction beyond a defined tolerance, and the acceptance criterion must be stated wherever the figure is used.

07

Energy and environmental performance

Training energy is a large, episodic expenditure. Inference energy is distributed across every production request. Neither is interpretable without a stated system boundary covering direct accelerator energy, host and network energy, facility overhead, embodied impacts, and water and grid effects. Two correctly computed figures for the same workload can differ by a large factor because they draw that boundary differently.

Fernandez and colleagues evaluated inference across workloads, hardware, serving frameworks, batching, decoding strategies, and parallelism. They report that estimates derived from floating-point operation counts understate real energy use, and that appropriate combinations of optimisations reduced energy by as much as 73 percent against an unoptimised baseline. That is the strongest argument against attaching a single energy number to a model name: the same weights on the same hardware under two serving configurations are two different energy propositions.

Oviedo and colleagues developed a bottom-up estimate for frontier-scale inference. Under stated H100 utilisation and power usage effectiveness assumptions, the median estimate was 0.34 watt-hours per representative query, with an interquartile range of 0.18 to 0.67 watt-hours. Raising token use by a factor of 15 for test-time scaling raised the median to 4.32 watt-hours, about 13 times higher. Combined model, serving, and hardware improvements were estimated to offer an 8 to 20 times efficiency opportunity. These are analytical estimates, not measurements of a named commercial service.

  • Interquartile range of the estimate
  • Median estimate
Representative query0.34 Wh

Interquartile range 0.18 to 0.67 watt-hours

Same query at fifteen times the token use4.32 Wh

About 13 times the median of the representative query. No interquartile range reported for this condition.

Estimated inference energy for one representative frontier-scale query, and for the same query under test-time scaling. Oviedo and colleagues, bottom-up analytical estimate under stated H100 utilisation and power usage effectiveness assumptions. Grade B. The band is the interquartile range of the estimate, not a measurement interval, and the scaled condition is a single reported median. Nothing here is a measurement of a named commercial service.
Data table
ConditionMedian estimateInterquartile range
Representative query0.34 Wh0.18 to 0.67 Wh
Same query at fifteen times the token use4.32 WhNot reported

Do not divide sector totals by a presumed query count. The International Energy Agency projects global data-centre electricity consumption reaching about 945 terawatt-hours in 2030 in its base case, roughly double the 2024 level, and reports that data-centre electricity demand grew 17 percent in 2025 with AI-focused facilities growing faster. That aggregate covers every data-centre workload, not AI queries alone, and utilisation patterns differ across them.

A credible energy disclosure identifies all of the following. A figure missing any of them is not comparable to another figure:

  • Model version and hardware
  • Numerical precision
  • Prompt and output length distributions
  • Batch size and utilisation
  • Serving software stack
  • Power usage effectiveness
  • Measurement or estimation method, and the measurement interval
  • The quality threshold at which the work counts as useful

The central environmental variable is therefore not model size. It is the energy required to produce an accepted, useful result at the required service level. Energy spent on an output that fails a quality check is energy spent for nothing.

08

A selection framework

Selection starts from failure cost, not from capability. What a wrong answer costs sets the control posture, and the control posture constrains the candidate set before any benchmark is consulted.

Risk tierExamplesMinimum control posture
LowDrafting, brainstorming, rewritingFast low-cost model, light validation
ModerateInternal summarisation, extraction, analytics assistanceSchema validation, citations, sampled human review
HighPublic research, financial analysis, legal support, code deploymentStrong model, source verification, tests, approvals
CriticalClinical decisions, autonomous transactions, safety systemsNarrow validated system, expert oversight, formal governance
Risk tiers and the minimum control posture each requires.

The right system is frequently not a single model. Six architecture patterns recur:

  • Cascade: a small model handles easy cases and a frontier model handles escalations, for high-volume mixed-difficulty workloads.
  • Router: a classifier selects model, tools, and reasoning budget, for multiple task types under a cost constraint.
  • Generator and verifier: one model produces and another checks claims or execution, where error cost is high and outputs are verifiable.
  • Retrieval-grounded: the model answers only from a controlled evidence set, for research, policy, legal, and enterprise knowledge.
  • Human in the loop: a person approves uncertain or consequential actions, for high-stakes decisions and transactions.
  • Local plus cloud: a private local model handles sensitive data and a cloud model handles de-identified hard tasks, under sovereignty and privacy constraints.

No model is selected on a published benchmark. The benchmark narrows the candidate set; a local evaluation selects. Define the task contract, build a representative evaluation set including adversarial and multilingual cases, hold prompts and tool access constant, measure end-to-end outcomes rather than single answers, inspect variance rather than the mean, and pin versions before every upgrade.

09

Research gaps and outlook

GapWhy it matters
Comparable energy disclosureNamed commercial models rarely publish measured joules per token under a standardised workload.
Agent reliabilityBenchmarks do not capture multi-day workflows, changing environments, permissions, and recovery.
Evaluation contaminationPublic benchmarks become training targets and lose diagnostic value.
Effective long contextBetter tests are needed for dispersed reasoning, long outputs, memory, and context compaction.
Multilingual equityTokenization, cost, accuracy, and safety remain uneven across languages and scripts.
Model updatesEndpoint behaviour changes without a new public model name, defeating reproducibility.
Human baselinesHuman scores are measured under different time, tool, and incentive conditions than model scores.
Environmental systems accountingEmbodied carbon, water, grid constraints, and rebound effects remain underreported.
Gaps that the current evidence base cannot close, and why each matters.

The likely near-term direction is adaptive systems that vary model size, reasoning budget, context, and tool use by task difficulty. That improves quality per unit of cost and energy while creating a new evaluation problem: the object being evaluated becomes a dynamic policy rather than a single model. Future benchmarks should report the complete resource budget, including hidden reasoning, retries, verification, and tool calls.

Open-weight models will continue to narrow capability gaps while supporting sovereign and specialised deployments. Their advantage will depend less on raw benchmark parity than on the ability to optimise the full stack, including quantization, retrieval, domain fine-tuning, caching, and hardware. Hosted frontier models will retain advantages in integrated tools, rapid updates, and breadth, and buyers should demand stronger transparency and version stability in exchange.

Conclusion. AI model selection should be treated as empirical engineering and governance, not brand preference. The right system is the least resource-intensive configuration that reliably meets the task contract and the risk threshold. This requires local evaluation, explicit evidence standards, controlled reasoning budgets, token and energy measurement, and continuous monitoring. Frontier capability is valuable, but unverified capability is not reliability.

10

Scope and limitations

This is a structured narrative review rather than a formal meta-analysis. A meta-analysis would mislead, because benchmark definitions, model versions, prompts, inference budgets, and access conditions change faster than a stable pooled effect can be estimated. The review therefore reports ranges, mechanisms, and measurement cautions rather than forcing every model into a composite score.

The unit of analysis is the deployed model system rather than the neural network alone. A deployed system may include a tokenizer, a routing layer, a hidden reasoning process, search, code execution, retrieval, prompt caching, safety classifiers, and agent orchestration. Two systems built on identical weights can differ in accuracy, latency, cost, and failure mode.

Prices and rankings change without notice. Model aliases, context limits, hosted features, and price schedules in this review are accurate to the research cut-off of 22 July 2026 and should be rechecked before procurement or production deployment. Nothing here is legal, financial, or procurement advice.

The accompanying research repository carries the full source register, the extraction protocol, the JSON schemas, and the validation tooling that enforce these rules mechanically. Comparison tables there are generated from a machine-readable data layer rather than written by hand, so no figure can appear without a source behind it.

References

Sources

  1. 01Stanford Institute for Human-Centered Artificial Intelligence (2026). AI Index Report 2026. Grade B.
  2. 02Stanford Center for Research on Foundation Models (2026). HELM Capabilities. Grade A.
  3. 03Stanford Center for Research on Foundation Models (2025). HELM Long Context. Grade A.
  4. 04Liang, P. and others (2022). Holistic evaluation of language models. arXiv:2211.09110. Grade B.
  5. 05Fernandez, J., Na, C., Tiwari, V., Bisk, Y., Luccioni, S. and Strubell, E. (2025). Energy considerations of large language model inference and efficiency optimizations. Proceedings of ACL 2025. Grade A.
  6. 06Oviedo, F., Kazhamiaka, F., Choukse, E., Kim, A., Luers, A., Nakagawa, M., Bianchini, R. and Lavista Ferres, J. M. (2026). Energy use of AI inference: efficiency pathways and test-time compute. Joule. Grade A.
  7. 07Wilhelm, P., Wittkopp, T. and Kao, O. (2025). Beyond test-time compute strategies: advocating energy-per-token in LLM inference. EuroMLSys 2025. Grade A.
  8. 08Jin, Y., Wei, G.-Y. and Brooks, D. (2025). The energy cost of reasoning: analyzing energy usage in LLMs with test-time compute. arXiv:2505.14733. Grade B.
  9. 09Fan, D., Delsad, S., Flammarion, N. and Andriushchenko, M. (2026). HalluHard: a hard multi-turn hallucination benchmark. arXiv:2602.01031. Grade B.
  10. 10NYU OSIRIS Lab (2026). LLM package hallucination research. Grade B.
  11. 11Google DeepMind (2024). FACTS Grounding: a benchmark for evaluating factuality. Grade B.
  12. 12Rando, S. and others (2025). LongCodeBench: evaluating coding LLMs at 1M context windows. arXiv:2505.07897. Grade B.
  13. 13Ye, X., Yin, F., He, Y., Zhang, J., Yen, H., Gao, T., Durrett, G. and Chen, D. (2025). LongProc: benchmarking long-context language models on long procedural generation. arXiv:2501.05414. Grade B.
  14. 14Asirvatham, H. and others (2026). GPT as a measurement tool. Harvard University. Grade B.
  15. 15International Energy Agency (2025). Energy and AI: energy demand from AI. Grade B.
  16. 16International Energy Agency (2026). Key questions on energy and AI. Grade B.
  17. 17Anthropic (2026). Models overview and Claude Sonnet 5 pricing notes. Claude Platform Documentation. Grade B for commercial terms.
  18. 18OpenAI (2026). GPT-5.6 model documentation. Grade B for commercial terms, Grade C for performance claims.
  19. 19Google DeepMind (2026). Gemini 3.1 Pro, and the Gemini 3.1 Flash-Lite model card. Grade C for performance claims.
  20. 20xAI (2026). Introducing Grok 4.5, and the Grok 4.5 model documentation. Grade C for performance claims.
  21. 21DeepSeek (2026). DeepSeek V4 preview release, and models and pricing. Grade B for architecture and commercial terms.
  22. 22Mistral AI (2026). Mistral Medium 3.5 and remote agents, and Introducing Mistral Small 4. Grade B.
  23. 23Meta (2025). The Llama 4 herd: natively multimodal open-weight models. Grade B.
  24. 24Bench360 authors (2025). Bench360: benchmarking local LLM inference from 360 degrees. arXiv:2511.16682. Grade B.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file diff --git a/static-site/print/ai-models-research/index.txt b/static-site/print/ai-models-research/index.txt index e3a874dfd..76417aec6 100644 --- a/static-site/print/ai-models-research/index.txt +++ b/static-site/print/ai-models-research/index.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -12:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","print","ai-models-research",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","ai-models-research","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -14:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -16:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","print","ai-models-research",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","ai-models-research","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:"$Sreact.suspense" -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1c:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -29,7 +29,7 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 19:[] 10:"$W19" 11:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1e:Ta14, @page { size: 210mm 297mm; margin: 20mm 18mm; } @@ -182,6 +182,6 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 62:["$","tr","6",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"GPT-5.6 Luna"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"0.10 USD"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"0.03 USD"}],["$","td","3",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"0.13 USD"}]]}] 63:["$","tr","7",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"DeepSeek V4 Pro"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"0.0435 USD"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"0.00435 USD"}],["$","td","3",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"0.04785 USD"}]]}] 1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -64:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +64:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 18:null 1d:[["$","title","0",{"children":"AI Model Performance (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L64","6",{}]] diff --git a/static-site/print/ai-research-assistant/__next._full.txt b/static-site/print/ai-research-assistant/__next._full.txt index d647de738..53a36dae4 100644 --- a/static-site/print/ai-research-assistant/__next._full.txt +++ b/static-site/print/ai-research-assistant/__next._full.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -12:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","print","ai-research-assistant",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","ai-research-assistant","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -14:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -16:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","print","ai-research-assistant",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","ai-research-assistant","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:"$Sreact.suspense" -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1c:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -29,7 +29,7 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 19:[] 10:"$W19" 11:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1e:Ta14, @page { size: 210mm 297mm; margin: 20mm 18mm; } @@ -124,6 +124,6 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 28:["$","section","conflicts-of-interest",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"08"}],["$","h2",null,{"className":"print-h2","children":"AI-Enabled Detection of Conflicts of Interest"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"Conflicts of interest undermine the credibility of academic research and weaken trust in scientific institutions. Financial ties, organizational affiliations, or personal relationships can influence how results are framed or interpreted, and manual review alone is not equipped to keep pace with the volume and complexity of today's scholarly output. AI systems offer a scalable, systematic way to identify signals of undisclosed or inaccurately reported conflicts, strengthening transparency across the publication process."}],["$","p","1",{"children":"AI can scan manuscripts for references to companies, products, funding sources, and affiliations, and compare those details with the information listed in author disclosures. When it notices a discrepancy — such as prominent discussion of a commercial product without a corresponding disclosure — it can flag the passage for further review."}],["$","p","2",{"children":"Beyond individual papers, AI can detect broader patterns across publications, including recurring collaborations or concentrated citation behavior that may indicate undisclosed relationships. These signals are not conclusions, but they prompt a closer look on which a researcher can decide whether follow-up is needed — expanding analytical capacity and strengthening transparency in the academic publishing process."}]]}]]}] 29:["$","section",null,{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"References"}],["$","h2",null,{"className":"print-h2","children":"Sources"}],["$","ol",null,{"className":"print-refs","children":[["$","li","0",{"children":[["$","span",null,{"className":"print-ref-num","children":"01"}],["$","span",null,{"children":"Pessin et al. (2025) — GapFinder: NLP mining of unstructured papers to surface unanswered questions."}]]}],["$","li","1",{"children":[["$","span",null,{"className":"print-ref-num","children":"02"}],["$","span",null,{"children":"Chatterjee et al. (2024) — Keyword co-occurrence graph mining to detect uncharted research themes."}]]}],["$","li","2",{"children":[["$","span",null,{"className":"print-ref-num","children":"03"}],["$","span",null,{"children":"Gu & Krenn (2024) — SCIMUSE: a 58M-paper knowledge graph paired with GPT-4 to propose research ideas (NeurIPS 2024, ML4Physical Sciences)."}]]}],["$","li","3",{"children":[["$","span",null,{"className":"print-ref-num","children":"04"}],["$","span",null,{"children":"Sybrandt et al. (2020) — AGATHA: graph mining and deep learning for biomedical hypothesis generation."}]]}],["$","li","4",{"children":[["$","span",null,{"className":"print-ref-num","children":"05"}],["$","span",null,{"children":"Altmami & Menai (2022) — A survey of machine-learning techniques for summarizing scientific articles (Springer)."}]]}],["$","li","5",{"children":[["$","span",null,{"className":"print-ref-num","children":"06"}],["$","span",null,{"children":"Scite — Smart Citations classify citation context as supporting, mentioning, or contrasting (Journal of the Medical Library Association)."}]]}],["$","li","6",{"children":[["$","span",null,{"className":"print-ref-num","children":"07"}],["$","span",null,{"children":"Tawfik & Spruit (2018) — Automated contradiction detection in biomedical literature (Springer)."}]]}],["$","li","7",{"children":[["$","span",null,{"className":"print-ref-num","children":"08"}],["$","span",null,{"children":"Bolaños et al. (2024) — AI systems that iteratively build cited literature reviews (Springer)."}]]}]]}],["$","p",null,{"className":"print-note","children":"This paper is supported by AI-enabled research assistance: the Microsoft 365 Copilot Researcher agent, running OpenAI's GPT-5 model, was used to partially generate and check content."}],"$undefined"]}] 1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -2a:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +2a:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 18:null 1d:[["$","title","0",{"children":"AI-Powered Research Questions (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L2a","6",{}]] diff --git a/static-site/print/ai-research-assistant/__next._head.txt b/static-site/print/ai-research-assistant/__next._head.txt index 76f4ac4b6..3bf92c3de 100644 --- a/static-site/print/ai-research-assistant/__next._head.txt +++ b/static-site/print/ai-research-assistant/__next._head.txt @@ -1,6 +1,6 @@ 1:"$Sreact.fragment" -2:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -3:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +2:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +3:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 4:"$Sreact.suspense" -5:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +5:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 0:{"rsc":["$","$1","h",{"children":[null,["$","$L2",null,{"children":[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]]}],["$","div",null,{"hidden":true,"children":["$","$L3",null,{"children":["$","$4",null,{"name":"Next.Metadata","children":[["$","title","0",{"children":"AI-Powered Research Questions (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L5","6",{}]]}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/print/ai-research-assistant/__next._index.txt b/static-site/print/ai-research-assistant/__next._index.txt index f98a4fd90..18b9d82b6 100644 --- a/static-site/print/ai-research-assistant/__next._index.txt +++ b/static-site/print/ai-research-assistant/__next._index.txt @@ -1,13 +1,13 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -d:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +d:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} 8:["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$Ld",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$Ld",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$Ld",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$Ld",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] diff --git a/static-site/print/ai-research-assistant/__next._tree.txt b/static-site/print/ai-research-assistant/__next._tree.txt index b6ce97ece..ed24a0489 100644 --- a/static-site/print/ai-research-assistant/__next._tree.txt +++ b/static-site/print/ai-research-assistant/__next._tree.txt @@ -1,7 +1,7 @@ -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] 0:{"tree":{"name":"","param":null,"prefetchHints":16,"slots":{"children":{"name":"print","param":null,"prefetchHints":0,"slots":{"children":{"name":"slug","param":{"type":"d","key":"ai-research-assistant","siblings":null},"prefetchHints":0,"slots":{"children":{"name":"__PAGE__","param":null,"prefetchHints":0,"slots":null}}}}}}},"staleTime":300,"buildId":"static-snapshot"} diff --git a/static-site/print/ai-research-assistant/__next.print.txt b/static-site/print/ai-research-assistant/__next.print.txt index 460afe242..fa617e997 100644 --- a/static-site/print/ai-research-assistant/__next.print.txt +++ b/static-site/print/ai-research-assistant/__next.print.txt @@ -1,4 +1,4 @@ 1:"$Sreact.fragment" -2:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -3:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +2:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +3:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] 0:{"rsc":["$","$1","c",{"children":[null,[["$","style",null,{"children":"\n /* The site is dark-first, and a dark color-scheme makes Chrome paint\n the printed sheet's margin area dark too. Paper is light. */\n html { color-scheme: light !important; }\n html, body { background: #ffffff !important; }\n body::after { content: none !important; }\n header, footer, .site-bg, .aura { display: none !important; }\n "}],["$","$L2",null,{"parallelRouterKey":"children","template":["$","$L3",null,{}]}]]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/print/ai-research-assistant/index.html b/static-site/print/ai-research-assistant/index.html index a99d46e08..c6010eee1 100644 --- a/static-site/print/ai-research-assistant/index.html +++ b/static-site/print/ai-research-assistant/index.html @@ -1,4 +1,4 @@ -AI-Powered Research Questions (print edition) · NYU Ethical Tech CoLab

Publications · Academic report

AI-Powered Research Questions

AI-Powered Assistance Across the Researcher's Workflow

Ethical Tech CoLab

October 2025

Yorke E. Rhodes III, Nathaniel Fossella, Alexa Shamie, Kirsten Co, Taylor Badt, Amanda Lindsey, Grace Driscoll, Mohagani Townsend, Pegi Bracaj, and Vedant Jain

AI is increasingly used to help academic researchers pose relevant research questions by mining vast literature for insights. This paper is an overview of how AI techniques and tools address key tasks in a researcher's workflow — finding research gaps, generating candidate questions, summarizing the state of the art, and comparing conflicting results.

Key figures

58M papers

distilled into SCIMUSE's knowledge graph before GPT-4 proposed new research directions

25%

of 4,451 AI-generated ideas were rated highly interesting by expert researchers

800M+

citation statements Scite classifies as supporting, mentioning, or contrasting prior work

5 tasks

gap-finding, question generation, summarization, contradiction detection, and integration

01

AI for Identifying Research Gaps

One of the first steps in posing a new research question is recognizing what hasn't been answered yet. AI can analyze large volumes of literature to detect gaps — unexplored or under-investigated topics — much faster than manual reading.

NLP-Based Gap Finders Tools like GapFinder (Pessin et al., 2025) use natural language processing to scan academic texts and map out research gaps. GapFinder mines unstructured paper content (PDFs) to identify areas not well covered by existing studies, acting as an automatic gap analysis that highlights topics or questions which remain unanswered in the current literature.

Keyword Co-occurrence Graphs Another approach leverages network algorithms on publication data. Chatterjee et al. (2024) construct keyword co-occurrence graphs and algorithmically detect structures that indicate a potential gap — strongly connected clusters of concepts that have not been jointly explored, i.e. interconnected yet uncharted research themes. By identifying combinations of topics that appear related but lack corresponding studies, such graph-mining techniques point researchers to plausible gaps that conventional reviews might overlook.

Knowledge Graph Analysis Large-scale knowledge graphs — networks of research concepts and their relationships — can be analyzed to spot missing links. If two areas are connected indirectly through known links but no direct research ties them together, that suggests a gap. Modern AI systems traverse these graphs to suggest which conceptual connections lack supporting studies, aligning with earlier literature-based discovery methods now enhanced by AI for scale and depth.

By automating gap discovery, AI tools give researchers a focused starting point — the not-yet-answered questions or unexplored topic intersections — on which to formulate new research questions.

02

AI for Formulating New Research Questions

Once gaps or novel idea spaces are identified, AI can help formulate concrete research questions or hypotheses. Recent advances in large language models and knowledge integration have produced systems that generate research ideas or suggest next steps, effectively acting as brainstorming partners grounded in data.

LLM-Generated Research Ideas A cutting-edge example is SCIMUSE (Gu & Krenn, NeurIPS 2024), which combines a massive literature-derived knowledge graph with GPT-4 to propose new research directions. SCIMUSE distilled knowledge from 58 million papers and prompted GPT-4 to generate novel project ideas tailored to a scientist's interests. Tested with over 100 senior researchers who evaluated 4,451 AI-proposed ideas, roughly 25% were rated highly interesting (a score of 4 or 5 out of 5). Such models can not only suggest topics but formulate them as concrete proposals, often including potential methodologies or collaborations.

Automated Hypothesis Generation Even before the latest LLMs, AI frameworks were generating hypotheses from existing findings. AGATHA (Sybrandt et al., 2020) combines graph mining and deep learning to propose new biomedical hypotheses, analyzing millions of scientific statements — interactions between genes, diseases, and chemicals — and learning to rank which unexplored connections are most plausible. In validation, AGATHA predicted research links later confirmed by papers published after 2015, demonstrating an ability to foresee meaningful research questions early.

Frameworks and Methodologies The general methodology is to leverage vast corpora — papers, databases, knowledge graphs — and use algorithms to recombine existing knowledge into new statements. Some rely on co-occurrence patterns that reveal an intermediate concept linking two previously unrelated ideas; others use language models to creatively assemble known facts into an innovative question. The common thread is that AI can connect dots across the literature in ways a human might overlook.

Used this way, AI gives researchers a systematic means of ideation. Instead of starting from scratch, they can explore a menu of AI-suggested questions that align with identified gaps, then apply their expertise to refine or pursue the most promising ones.

03

AI for Summarizing the State of the Art

To pose a relevant question, a researcher must understand current knowledge — what is already known and what remains uncertain. AI assists through advanced summarization, condensing an ever-growing literature into digestible syntheses of the state of the art.

Automated Literature Summaries AI tools can ingest multiple papers on a topic and produce concise summaries of key findings, methods, and consensus. LLM-powered search platforms such as Elicit, Scispace, and Consensus let users ask a research question in natural language and return relevant papers, each with a brief summary of its contribution — giving a quick overview of current knowledge without reading each paper in full.

State-of-the-Art Overviews Some systems perform multi-document summarization, synthesizing information across many papers into one coherent overview. Work here includes using models like BERT and SciBERT for scientific text summarization and efforts to generate literature reviews automatically. Altmami & Menai (2022) survey machine-learning techniques for summarizing scientific articles, and shared tasks such as Allen AI's Multi-Document Summarization for Literature Review aim to train AI to write an aggregated summary from dozens of papers — effectively a draft related-work section.

Knowledge Graphs and Semantic Summaries Structured representations such as AIDA or the Open Research Knowledge Graph (ORKG) encode scholarly knowledge as interconnected data — papers, methods, and results as linked entries. AI can traverse these graphs to produce a comparative view of the state of the art (for example, listing all known approaches to a problem and their outcomes) and, importantly, highlight which nodes are missing, tying back into gap detection.

AI-driven summarization dramatically speeds up literature review. A researcher can quickly learn what is known about a topic, or which methods have been tried, rather than spending weeks reading. This focused understanding helps ground new questions in current knowledge — building on what is known and specifically targeting what is not yet well understood.

04

AI for Comparing and Contrasting Research Findings

Researchers often need to reconcile conflicting results. AI can compare publications and flag where their findings diverge, which is crucial for refining research questions and identifying controversies to investigate.

Citation Analysis for Disagreements Scite, an AI-powered citation index, uses deep learning to classify the context of citations as supporting, mentioning, or contrasting the cited work. Drawing on over 800 million citation statements, its algorithms determine when a study's results have been challenged by subsequent work. Instead of manually collating hints of disagreement, a researcher can see directly that one paper contradicts another's conclusion — making it far easier to pinpoint debates or inconsistencies in a field.

Natural Language Contradiction Detection Beyond citation context, AI can compare the content of findings. Tawfik & Spruit (2018) developed a model for automated contradiction detection in biomedical literature that first extracts key outcome statements (e.g. 'Drug A improves condition B' versus 'Drug A has no effect on condition B'), then uses semantic features — negation, antonyms, and numeric comparisons — to judge whether two statements conflict. Applied to clinical studies, it identified when different papers answered the same question with opposing conclusions, surfacing controversies that often lead directly to follow-up questions.

Cross-Paper Evidence Synthesis AI is also used in systematic reviews to ensure all angles are considered. Machine-learning classifiers can cluster papers by outcome or stance, semi-automatically separating studies that support a hypothesis from those that do not, and alerting the researcher to clusters of contradictions worth investigating.

By exposing inconsistencies and debates, AI tools show researchers where the science is not settled. These contested areas often make for important research questions — for instance, 'What explains the divergent results regarding X?' AI does the heavy lifting of identifying knowledge conflicts so the researcher can focus on resolving them.

05

Integrated AI Systems for Research-Question Generation

Bringing these capabilities together, modern AI research assistants provide end-to-end support — from mapping existing knowledge to highlighting gaps and inconsistencies, often in an interactive workflow.

Consolidated Platforms New research-assistant tools — commercial and experimental platforms such as Semantic Scholar's AI, IBM Watson Discovery for science, and academic prototypes — integrate search, summarization, and analysis. Some AI writing assistants (e.g. Jenni.ai, Silatus) let a user enter a tentative topic, then automatically fetch relevant literature, summarize it, and suggest next questions. As Bolaños et al. (2024) describe, such systems can iteratively build a literature review, surfacing open questions during the writing process — noting, for instance, that 'few studies address Z,' cueing a potential new question.

Hypothesis and Idea Suggestion Integrated agents can propose questions in context. A researcher might upload a set of papers and ask 'what is a good next research question here?' The AI combines the findings, points out a missing piece, and phrases a candidate question. Some systems use retrieval-augmented generation (RAG) to pull specifics from documents and have an LLM draft open-problem paragraphs, explicitly listing unsolved problems gleaned from the literature.

Workflow Efficiency The overall impact is on efficiency and comprehensiveness. By the time a researcher formulates a question, they have already seen the summarized state of the art, the known gaps, and the controversial findings — reducing the chance of proposing something trivial or already answered. As surveys note, current tools are not flawless (hallucinations and gaps in coverage exist), but the trend is for AI to handle more of the heavy lifting, leaving researchers to concentrate on creative and critical thinking. Human insight remains crucial: the AI provides options and evidence, but the researcher chooses and refines the question.

AI contributes at every stage of question formulation — scouting the landscape of knowledge, spotlighting what is unknown or disputed, and even drafting candidate questions. For a researcher, these tools act like a smart assistant, ensuring the questions they pose are informed, significant, and address genuine gaps: a faster, more focused path to high-impact research.

06

The Emotional Dimension of Research in the Age of AI

While AI-powered tools can streamline the formulation of research questions and accelerate literature analysis, their growing presence does not diminish the emotional gratification traditionally associated with academic work. Rather, it redefines how that sense of relief, pride, and satisfaction is experienced. The fulfillment that arises from researching and writing — through effort, reflection, and perseverance — remains intact when scholars engage intentionally and thoughtfully with AI. The time invested in interpreting, refining, and shaping ideas continues to enhance both intellectual rigor and personal reward.

Instead of reducing the depth or authenticity of scholarly effort, AI research assistants can preserve and even amplify the emotional dimensions of academic inquiry. By relieving researchers of repetitive or mechanical tasks, AI allows greater focus on the creative, analytical, and interpretive aspects that form the heart of scholarship. The gradual movement from uncertainty to clarity — the hallmark of intellectual discovery — still occurs; it simply unfolds through collaboration between human judgment and machine assistance. The researcher's sense of accomplishment now stems not only from producing results but from orchestrating and curating the knowledge process itself.

Protecting and cultivating emotional satisfaction in this environment requires recognizing that human agency remains central. AI can efficiently identify relevant literature or summarize findings, but the researcher continues to decide what matters, why it matters, and how it contributes to the broader conversation. The source of pride thus shifts from labor-intensive information gathering to higher-order critical thinking — evaluating, connecting, and creating meaning.

Many researchers also choose to maintain elements of manual engagement, such as drafting outlines or annotating papers independently before consulting AI tools. These deliberate choices reinforce the personal connection to one's work and sustain the emotional gratification that comes from intellectual perseverance. The presence of AI does not erase these experiences; it reframes them within a more collaborative and dynamic process of knowledge creation.

Ultimately, AI research assistants do not weaken the emotional core of academic work — they evolve it. Relief, pride, and satisfaction remain grounded in the same human capacities that have always driven scholarship: curiosity, discernment, and creativity. As AI becomes more deeply integrated into research workflows, the challenge is not to preserve emotion against technology but to recognize how these tools can extend the very sense of fulfillment that comes from thinking deeply, questioning critically, and contributing meaningfully to human understanding.

07

Guardrails for AI-Assisted Research Questioning

These advances demonstrate the growing role of AI in shaping scholarly inquiry and research design. However, these capabilities introduce parallel risks that necessitate structured guardrails. While AI can help align research questions with documented gaps in the literature, it is equally capable of hallucinating citations, fabricating conceptual linkages, or overstating consensus where none exists. The same generative power that lets AI suggest high-value inquiries can also produce plausible-sounding but incorrect premises, or subtly shift a researcher's framing toward what the model predicts rather than what the literature supports. When AI tools summarize or frame a body of work, researchers may be tempted to rely on the output without independently engaging the foundational texts — risking a second-order understanding and a loss of methodological rigor.

Given these dynamics, responsible adoption requires AI systems that do more than generate ideas: they must support verification, transparency, and careful acknowledgement of uncertainty. Guardrails are not barriers to innovation; they are mechanisms to ensure that efficiency gains do not erode scholarly integrity. These controls include expectations that researchers:

  • Independently review core sources rather than relying solely on AI summaries.
  • Confirm that AI-generated citations and claims are accurate and grounded in the original texts.
  • Treat AI-suggested research questions as exploratory starting points, not final conclusions.
  • Reflect on and document how AI shaped their thinking, assumptions, or direction.

Establishing these practices is essential for maintaining academic rigor, particularly as graduate-level research increasingly intersects with intelligent systems capable of shaping epistemic direction.

08

AI-Enabled Detection of Conflicts of Interest

Conflicts of interest undermine the credibility of academic research and weaken trust in scientific institutions. Financial ties, organizational affiliations, or personal relationships can influence how results are framed or interpreted, and manual review alone is not equipped to keep pace with the volume and complexity of today's scholarly output. AI systems offer a scalable, systematic way to identify signals of undisclosed or inaccurately reported conflicts, strengthening transparency across the publication process.

AI can scan manuscripts for references to companies, products, funding sources, and affiliations, and compare those details with the information listed in author disclosures. When it notices a discrepancy — such as prominent discussion of a commercial product without a corresponding disclosure — it can flag the passage for further review.

Beyond individual papers, AI can detect broader patterns across publications, including recurring collaborations or concentrated citation behavior that may indicate undisclosed relationships. These signals are not conclusions, but they prompt a closer look on which a researcher can decide whether follow-up is needed — expanding analytical capacity and strengthening transparency in the academic publishing process.

References

Sources

  1. 01Pessin et al. (2025) — GapFinder: NLP mining of unstructured papers to surface unanswered questions.
  2. 02Chatterjee et al. (2024) — Keyword co-occurrence graph mining to detect uncharted research themes.
  3. 03Gu & Krenn (2024) — SCIMUSE: a 58M-paper knowledge graph paired with GPT-4 to propose research ideas (NeurIPS 2024, ML4Physical Sciences).
  4. 04Sybrandt et al. (2020) — AGATHA: graph mining and deep learning for biomedical hypothesis generation.
  5. 05Altmami & Menai (2022) — A survey of machine-learning techniques for summarizing scientific articles (Springer).
  6. 06Scite — Smart Citations classify citation context as supporting, mentioning, or contrasting (Journal of the Medical Library Association).
  7. 07Tawfik & Spruit (2018) — Automated contradiction detection in biomedical literature (Springer).
  8. 08Bolaños et al. (2024) — AI systems that iteratively build cited literature reviews (Springer).

This paper is supported by AI-enabled research assistance: the Microsoft 365 Copilot Researcher agent, running OpenAI's GPT-5 model, was used to partially generate and check content.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file +

Publications · Academic report

AI-Powered Research Questions

AI-Powered Assistance Across the Researcher's Workflow

Ethical Tech CoLab

October 2025

Yorke E. Rhodes III, Nathaniel Fossella, Alexa Shamie, Kirsten Co, Taylor Badt, Amanda Lindsey, Grace Driscoll, Mohagani Townsend, Pegi Bracaj, and Vedant Jain

AI is increasingly used to help academic researchers pose relevant research questions by mining vast literature for insights. This paper is an overview of how AI techniques and tools address key tasks in a researcher's workflow — finding research gaps, generating candidate questions, summarizing the state of the art, and comparing conflicting results.

Key figures

58M papers

distilled into SCIMUSE's knowledge graph before GPT-4 proposed new research directions

25%

of 4,451 AI-generated ideas were rated highly interesting by expert researchers

800M+

citation statements Scite classifies as supporting, mentioning, or contrasting prior work

5 tasks

gap-finding, question generation, summarization, contradiction detection, and integration

01

AI for Identifying Research Gaps

One of the first steps in posing a new research question is recognizing what hasn't been answered yet. AI can analyze large volumes of literature to detect gaps — unexplored or under-investigated topics — much faster than manual reading.

NLP-Based Gap Finders Tools like GapFinder (Pessin et al., 2025) use natural language processing to scan academic texts and map out research gaps. GapFinder mines unstructured paper content (PDFs) to identify areas not well covered by existing studies, acting as an automatic gap analysis that highlights topics or questions which remain unanswered in the current literature.

Keyword Co-occurrence Graphs Another approach leverages network algorithms on publication data. Chatterjee et al. (2024) construct keyword co-occurrence graphs and algorithmically detect structures that indicate a potential gap — strongly connected clusters of concepts that have not been jointly explored, i.e. interconnected yet uncharted research themes. By identifying combinations of topics that appear related but lack corresponding studies, such graph-mining techniques point researchers to plausible gaps that conventional reviews might overlook.

Knowledge Graph Analysis Large-scale knowledge graphs — networks of research concepts and their relationships — can be analyzed to spot missing links. If two areas are connected indirectly through known links but no direct research ties them together, that suggests a gap. Modern AI systems traverse these graphs to suggest which conceptual connections lack supporting studies, aligning with earlier literature-based discovery methods now enhanced by AI for scale and depth.

By automating gap discovery, AI tools give researchers a focused starting point — the not-yet-answered questions or unexplored topic intersections — on which to formulate new research questions.

02

AI for Formulating New Research Questions

Once gaps or novel idea spaces are identified, AI can help formulate concrete research questions or hypotheses. Recent advances in large language models and knowledge integration have produced systems that generate research ideas or suggest next steps, effectively acting as brainstorming partners grounded in data.

LLM-Generated Research Ideas A cutting-edge example is SCIMUSE (Gu & Krenn, NeurIPS 2024), which combines a massive literature-derived knowledge graph with GPT-4 to propose new research directions. SCIMUSE distilled knowledge from 58 million papers and prompted GPT-4 to generate novel project ideas tailored to a scientist's interests. Tested with over 100 senior researchers who evaluated 4,451 AI-proposed ideas, roughly 25% were rated highly interesting (a score of 4 or 5 out of 5). Such models can not only suggest topics but formulate them as concrete proposals, often including potential methodologies or collaborations.

Automated Hypothesis Generation Even before the latest LLMs, AI frameworks were generating hypotheses from existing findings. AGATHA (Sybrandt et al., 2020) combines graph mining and deep learning to propose new biomedical hypotheses, analyzing millions of scientific statements — interactions between genes, diseases, and chemicals — and learning to rank which unexplored connections are most plausible. In validation, AGATHA predicted research links later confirmed by papers published after 2015, demonstrating an ability to foresee meaningful research questions early.

Frameworks and Methodologies The general methodology is to leverage vast corpora — papers, databases, knowledge graphs — and use algorithms to recombine existing knowledge into new statements. Some rely on co-occurrence patterns that reveal an intermediate concept linking two previously unrelated ideas; others use language models to creatively assemble known facts into an innovative question. The common thread is that AI can connect dots across the literature in ways a human might overlook.

Used this way, AI gives researchers a systematic means of ideation. Instead of starting from scratch, they can explore a menu of AI-suggested questions that align with identified gaps, then apply their expertise to refine or pursue the most promising ones.

03

AI for Summarizing the State of the Art

To pose a relevant question, a researcher must understand current knowledge — what is already known and what remains uncertain. AI assists through advanced summarization, condensing an ever-growing literature into digestible syntheses of the state of the art.

Automated Literature Summaries AI tools can ingest multiple papers on a topic and produce concise summaries of key findings, methods, and consensus. LLM-powered search platforms such as Elicit, Scispace, and Consensus let users ask a research question in natural language and return relevant papers, each with a brief summary of its contribution — giving a quick overview of current knowledge without reading each paper in full.

State-of-the-Art Overviews Some systems perform multi-document summarization, synthesizing information across many papers into one coherent overview. Work here includes using models like BERT and SciBERT for scientific text summarization and efforts to generate literature reviews automatically. Altmami & Menai (2022) survey machine-learning techniques for summarizing scientific articles, and shared tasks such as Allen AI's Multi-Document Summarization for Literature Review aim to train AI to write an aggregated summary from dozens of papers — effectively a draft related-work section.

Knowledge Graphs and Semantic Summaries Structured representations such as AIDA or the Open Research Knowledge Graph (ORKG) encode scholarly knowledge as interconnected data — papers, methods, and results as linked entries. AI can traverse these graphs to produce a comparative view of the state of the art (for example, listing all known approaches to a problem and their outcomes) and, importantly, highlight which nodes are missing, tying back into gap detection.

AI-driven summarization dramatically speeds up literature review. A researcher can quickly learn what is known about a topic, or which methods have been tried, rather than spending weeks reading. This focused understanding helps ground new questions in current knowledge — building on what is known and specifically targeting what is not yet well understood.

04

AI for Comparing and Contrasting Research Findings

Researchers often need to reconcile conflicting results. AI can compare publications and flag where their findings diverge, which is crucial for refining research questions and identifying controversies to investigate.

Citation Analysis for Disagreements Scite, an AI-powered citation index, uses deep learning to classify the context of citations as supporting, mentioning, or contrasting the cited work. Drawing on over 800 million citation statements, its algorithms determine when a study's results have been challenged by subsequent work. Instead of manually collating hints of disagreement, a researcher can see directly that one paper contradicts another's conclusion — making it far easier to pinpoint debates or inconsistencies in a field.

Natural Language Contradiction Detection Beyond citation context, AI can compare the content of findings. Tawfik & Spruit (2018) developed a model for automated contradiction detection in biomedical literature that first extracts key outcome statements (e.g. 'Drug A improves condition B' versus 'Drug A has no effect on condition B'), then uses semantic features — negation, antonyms, and numeric comparisons — to judge whether two statements conflict. Applied to clinical studies, it identified when different papers answered the same question with opposing conclusions, surfacing controversies that often lead directly to follow-up questions.

Cross-Paper Evidence Synthesis AI is also used in systematic reviews to ensure all angles are considered. Machine-learning classifiers can cluster papers by outcome or stance, semi-automatically separating studies that support a hypothesis from those that do not, and alerting the researcher to clusters of contradictions worth investigating.

By exposing inconsistencies and debates, AI tools show researchers where the science is not settled. These contested areas often make for important research questions — for instance, 'What explains the divergent results regarding X?' AI does the heavy lifting of identifying knowledge conflicts so the researcher can focus on resolving them.

05

Integrated AI Systems for Research-Question Generation

Bringing these capabilities together, modern AI research assistants provide end-to-end support — from mapping existing knowledge to highlighting gaps and inconsistencies, often in an interactive workflow.

Consolidated Platforms New research-assistant tools — commercial and experimental platforms such as Semantic Scholar's AI, IBM Watson Discovery for science, and academic prototypes — integrate search, summarization, and analysis. Some AI writing assistants (e.g. Jenni.ai, Silatus) let a user enter a tentative topic, then automatically fetch relevant literature, summarize it, and suggest next questions. As Bolaños et al. (2024) describe, such systems can iteratively build a literature review, surfacing open questions during the writing process — noting, for instance, that 'few studies address Z,' cueing a potential new question.

Hypothesis and Idea Suggestion Integrated agents can propose questions in context. A researcher might upload a set of papers and ask 'what is a good next research question here?' The AI combines the findings, points out a missing piece, and phrases a candidate question. Some systems use retrieval-augmented generation (RAG) to pull specifics from documents and have an LLM draft open-problem paragraphs, explicitly listing unsolved problems gleaned from the literature.

Workflow Efficiency The overall impact is on efficiency and comprehensiveness. By the time a researcher formulates a question, they have already seen the summarized state of the art, the known gaps, and the controversial findings — reducing the chance of proposing something trivial or already answered. As surveys note, current tools are not flawless (hallucinations and gaps in coverage exist), but the trend is for AI to handle more of the heavy lifting, leaving researchers to concentrate on creative and critical thinking. Human insight remains crucial: the AI provides options and evidence, but the researcher chooses and refines the question.

AI contributes at every stage of question formulation — scouting the landscape of knowledge, spotlighting what is unknown or disputed, and even drafting candidate questions. For a researcher, these tools act like a smart assistant, ensuring the questions they pose are informed, significant, and address genuine gaps: a faster, more focused path to high-impact research.

06

The Emotional Dimension of Research in the Age of AI

While AI-powered tools can streamline the formulation of research questions and accelerate literature analysis, their growing presence does not diminish the emotional gratification traditionally associated with academic work. Rather, it redefines how that sense of relief, pride, and satisfaction is experienced. The fulfillment that arises from researching and writing — through effort, reflection, and perseverance — remains intact when scholars engage intentionally and thoughtfully with AI. The time invested in interpreting, refining, and shaping ideas continues to enhance both intellectual rigor and personal reward.

Instead of reducing the depth or authenticity of scholarly effort, AI research assistants can preserve and even amplify the emotional dimensions of academic inquiry. By relieving researchers of repetitive or mechanical tasks, AI allows greater focus on the creative, analytical, and interpretive aspects that form the heart of scholarship. The gradual movement from uncertainty to clarity — the hallmark of intellectual discovery — still occurs; it simply unfolds through collaboration between human judgment and machine assistance. The researcher's sense of accomplishment now stems not only from producing results but from orchestrating and curating the knowledge process itself.

Protecting and cultivating emotional satisfaction in this environment requires recognizing that human agency remains central. AI can efficiently identify relevant literature or summarize findings, but the researcher continues to decide what matters, why it matters, and how it contributes to the broader conversation. The source of pride thus shifts from labor-intensive information gathering to higher-order critical thinking — evaluating, connecting, and creating meaning.

Many researchers also choose to maintain elements of manual engagement, such as drafting outlines or annotating papers independently before consulting AI tools. These deliberate choices reinforce the personal connection to one's work and sustain the emotional gratification that comes from intellectual perseverance. The presence of AI does not erase these experiences; it reframes them within a more collaborative and dynamic process of knowledge creation.

Ultimately, AI research assistants do not weaken the emotional core of academic work — they evolve it. Relief, pride, and satisfaction remain grounded in the same human capacities that have always driven scholarship: curiosity, discernment, and creativity. As AI becomes more deeply integrated into research workflows, the challenge is not to preserve emotion against technology but to recognize how these tools can extend the very sense of fulfillment that comes from thinking deeply, questioning critically, and contributing meaningfully to human understanding.

07

Guardrails for AI-Assisted Research Questioning

These advances demonstrate the growing role of AI in shaping scholarly inquiry and research design. However, these capabilities introduce parallel risks that necessitate structured guardrails. While AI can help align research questions with documented gaps in the literature, it is equally capable of hallucinating citations, fabricating conceptual linkages, or overstating consensus where none exists. The same generative power that lets AI suggest high-value inquiries can also produce plausible-sounding but incorrect premises, or subtly shift a researcher's framing toward what the model predicts rather than what the literature supports. When AI tools summarize or frame a body of work, researchers may be tempted to rely on the output without independently engaging the foundational texts — risking a second-order understanding and a loss of methodological rigor.

Given these dynamics, responsible adoption requires AI systems that do more than generate ideas: they must support verification, transparency, and careful acknowledgement of uncertainty. Guardrails are not barriers to innovation; they are mechanisms to ensure that efficiency gains do not erode scholarly integrity. These controls include expectations that researchers:

  • Independently review core sources rather than relying solely on AI summaries.
  • Confirm that AI-generated citations and claims are accurate and grounded in the original texts.
  • Treat AI-suggested research questions as exploratory starting points, not final conclusions.
  • Reflect on and document how AI shaped their thinking, assumptions, or direction.

Establishing these practices is essential for maintaining academic rigor, particularly as graduate-level research increasingly intersects with intelligent systems capable of shaping epistemic direction.

08

AI-Enabled Detection of Conflicts of Interest

Conflicts of interest undermine the credibility of academic research and weaken trust in scientific institutions. Financial ties, organizational affiliations, or personal relationships can influence how results are framed or interpreted, and manual review alone is not equipped to keep pace with the volume and complexity of today's scholarly output. AI systems offer a scalable, systematic way to identify signals of undisclosed or inaccurately reported conflicts, strengthening transparency across the publication process.

AI can scan manuscripts for references to companies, products, funding sources, and affiliations, and compare those details with the information listed in author disclosures. When it notices a discrepancy — such as prominent discussion of a commercial product without a corresponding disclosure — it can flag the passage for further review.

Beyond individual papers, AI can detect broader patterns across publications, including recurring collaborations or concentrated citation behavior that may indicate undisclosed relationships. These signals are not conclusions, but they prompt a closer look on which a researcher can decide whether follow-up is needed — expanding analytical capacity and strengthening transparency in the academic publishing process.

References

Sources

  1. 01Pessin et al. (2025) — GapFinder: NLP mining of unstructured papers to surface unanswered questions.
  2. 02Chatterjee et al. (2024) — Keyword co-occurrence graph mining to detect uncharted research themes.
  3. 03Gu & Krenn (2024) — SCIMUSE: a 58M-paper knowledge graph paired with GPT-4 to propose research ideas (NeurIPS 2024, ML4Physical Sciences).
  4. 04Sybrandt et al. (2020) — AGATHA: graph mining and deep learning for biomedical hypothesis generation.
  5. 05Altmami & Menai (2022) — A survey of machine-learning techniques for summarizing scientific articles (Springer).
  6. 06Scite — Smart Citations classify citation context as supporting, mentioning, or contrasting (Journal of the Medical Library Association).
  7. 07Tawfik & Spruit (2018) — Automated contradiction detection in biomedical literature (Springer).
  8. 08Bolaños et al. (2024) — AI systems that iteratively build cited literature reviews (Springer).

This paper is supported by AI-enabled research assistance: the Microsoft 365 Copilot Researcher agent, running OpenAI's GPT-5 model, was used to partially generate and check content.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file diff --git a/static-site/print/ai-research-assistant/index.txt b/static-site/print/ai-research-assistant/index.txt index d647de738..53a36dae4 100644 --- a/static-site/print/ai-research-assistant/index.txt +++ b/static-site/print/ai-research-assistant/index.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -12:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","print","ai-research-assistant",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","ai-research-assistant","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -14:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -16:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","print","ai-research-assistant",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","ai-research-assistant","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:"$Sreact.suspense" -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1c:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -29,7 +29,7 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 19:[] 10:"$W19" 11:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1e:Ta14, @page { size: 210mm 297mm; margin: 20mm 18mm; } @@ -124,6 +124,6 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 28:["$","section","conflicts-of-interest",{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"08"}],["$","h2",null,{"className":"print-h2","children":"AI-Enabled Detection of Conflicts of Interest"}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"Conflicts of interest undermine the credibility of academic research and weaken trust in scientific institutions. Financial ties, organizational affiliations, or personal relationships can influence how results are framed or interpreted, and manual review alone is not equipped to keep pace with the volume and complexity of today's scholarly output. AI systems offer a scalable, systematic way to identify signals of undisclosed or inaccurately reported conflicts, strengthening transparency across the publication process."}],["$","p","1",{"children":"AI can scan manuscripts for references to companies, products, funding sources, and affiliations, and compare those details with the information listed in author disclosures. When it notices a discrepancy — such as prominent discussion of a commercial product without a corresponding disclosure — it can flag the passage for further review."}],["$","p","2",{"children":"Beyond individual papers, AI can detect broader patterns across publications, including recurring collaborations or concentrated citation behavior that may indicate undisclosed relationships. These signals are not conclusions, but they prompt a closer look on which a researcher can decide whether follow-up is needed — expanding analytical capacity and strengthening transparency in the academic publishing process."}]]}]]}] 29:["$","section",null,{"className":"print-page","children":[["$","p",null,{"className":"print-section-number","children":"References"}],["$","h2",null,{"className":"print-h2","children":"Sources"}],["$","ol",null,{"className":"print-refs","children":[["$","li","0",{"children":[["$","span",null,{"className":"print-ref-num","children":"01"}],["$","span",null,{"children":"Pessin et al. (2025) — GapFinder: NLP mining of unstructured papers to surface unanswered questions."}]]}],["$","li","1",{"children":[["$","span",null,{"className":"print-ref-num","children":"02"}],["$","span",null,{"children":"Chatterjee et al. (2024) — Keyword co-occurrence graph mining to detect uncharted research themes."}]]}],["$","li","2",{"children":[["$","span",null,{"className":"print-ref-num","children":"03"}],["$","span",null,{"children":"Gu & Krenn (2024) — SCIMUSE: a 58M-paper knowledge graph paired with GPT-4 to propose research ideas (NeurIPS 2024, ML4Physical Sciences)."}]]}],["$","li","3",{"children":[["$","span",null,{"className":"print-ref-num","children":"04"}],["$","span",null,{"children":"Sybrandt et al. (2020) — AGATHA: graph mining and deep learning for biomedical hypothesis generation."}]]}],["$","li","4",{"children":[["$","span",null,{"className":"print-ref-num","children":"05"}],["$","span",null,{"children":"Altmami & Menai (2022) — A survey of machine-learning techniques for summarizing scientific articles (Springer)."}]]}],["$","li","5",{"children":[["$","span",null,{"className":"print-ref-num","children":"06"}],["$","span",null,{"children":"Scite — Smart Citations classify citation context as supporting, mentioning, or contrasting (Journal of the Medical Library Association)."}]]}],["$","li","6",{"children":[["$","span",null,{"className":"print-ref-num","children":"07"}],["$","span",null,{"children":"Tawfik & Spruit (2018) — Automated contradiction detection in biomedical literature (Springer)."}]]}],["$","li","7",{"children":[["$","span",null,{"className":"print-ref-num","children":"08"}],["$","span",null,{"children":"Bolaños et al. (2024) — AI systems that iteratively build cited literature reviews (Springer)."}]]}]]}],["$","p",null,{"className":"print-note","children":"This paper is supported by AI-enabled research assistance: the Microsoft 365 Copilot Researcher agent, running OpenAI's GPT-5 model, was used to partially generate and check content."}],"$undefined"]}] 1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -2a:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +2a:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 18:null 1d:[["$","title","0",{"children":"AI-Powered Research Questions (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L2a","6",{}]] diff --git a/static-site/print/cerai/__next._full.txt b/static-site/print/cerai/__next._full.txt index d8b6195c3..579be757c 100644 --- a/static-site/print/cerai/__next._full.txt +++ b/static-site/print/cerai/__next._full.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -12:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","print","cerai",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","cerai","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -14:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -16:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","print","cerai",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","cerai","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:"$Sreact.suspense" -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1c:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -29,7 +29,7 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 19:[] 10:"$W19" 11:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1e:Ta14, @page { size: 210mm 297mm; margin: 20mm 18mm; } @@ -136,6 +136,6 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 34:["$","li","18",{"children":[["$","span",null,{"className":"print-ref-num","children":"19"}],["$","span",null,{"children":"Integrated Food Security Phase Classification. IPC acute food insecurity analysis."}]]}] 35:["$","li","19",{"children":[["$","span",null,{"className":"print-ref-num","children":"20"}],["$","span",null,{"children":"European Commission Joint Research Centre. Global Conflict Risk Index (GCRI)."}]]}] 1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -36:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +36:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 18:null 1d:[["$","title","0",{"children":"The Civilian Evacuation Risk Anticipation Index (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L36","6",{}]] diff --git a/static-site/print/cerai/__next._head.txt b/static-site/print/cerai/__next._head.txt index 68530bdf5..cfd5a5790 100644 --- a/static-site/print/cerai/__next._head.txt +++ b/static-site/print/cerai/__next._head.txt @@ -1,6 +1,6 @@ 1:"$Sreact.fragment" -2:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -3:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +2:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +3:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 4:"$Sreact.suspense" -5:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +5:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 0:{"rsc":["$","$1","h",{"children":[null,["$","$L2",null,{"children":[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]]}],["$","div",null,{"hidden":true,"children":["$","$L3",null,{"children":["$","$4",null,{"name":"Next.Metadata","children":[["$","title","0",{"children":"The Civilian Evacuation Risk Anticipation Index (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L5","6",{}]]}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/print/cerai/__next._index.txt b/static-site/print/cerai/__next._index.txt index f98a4fd90..18b9d82b6 100644 --- a/static-site/print/cerai/__next._index.txt +++ b/static-site/print/cerai/__next._index.txt @@ -1,13 +1,13 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -d:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +d:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} 8:["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$Ld",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$Ld",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$Ld",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$Ld",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] diff --git a/static-site/print/cerai/__next._tree.txt b/static-site/print/cerai/__next._tree.txt index 70071eebd..a86f80b95 100644 --- a/static-site/print/cerai/__next._tree.txt +++ b/static-site/print/cerai/__next._tree.txt @@ -1,7 +1,7 @@ -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] 0:{"tree":{"name":"","param":null,"prefetchHints":16,"slots":{"children":{"name":"print","param":null,"prefetchHints":0,"slots":{"children":{"name":"slug","param":{"type":"d","key":"cerai","siblings":null},"prefetchHints":0,"slots":{"children":{"name":"__PAGE__","param":null,"prefetchHints":0,"slots":null}}}}}}},"staleTime":300,"buildId":"static-snapshot"} diff --git a/static-site/print/cerai/__next.print.txt b/static-site/print/cerai/__next.print.txt index 460afe242..fa617e997 100644 --- a/static-site/print/cerai/__next.print.txt +++ b/static-site/print/cerai/__next.print.txt @@ -1,4 +1,4 @@ 1:"$Sreact.fragment" -2:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -3:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +2:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +3:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] 0:{"rsc":["$","$1","c",{"children":[null,[["$","style",null,{"children":"\n /* The site is dark-first, and a dark color-scheme makes Chrome paint\n the printed sheet's margin area dark too. Paper is light. */\n html { color-scheme: light !important; }\n html, body { background: #ffffff !important; }\n body::after { content: none !important; }\n header, footer, .site-bg, .aura { display: none !important; }\n "}],["$","$L2",null,{"parallelRouterKey":"children","template":["$","$L3",null,{}]}]]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/print/cerai/index.html b/static-site/print/cerai/index.html index a467f1f0b..a5b6ff4c2 100644 --- a/static-site/print/cerai/index.html +++ b/static-site/print/cerai/index.html @@ -1,4 +1,4 @@ -The Civilian Evacuation Risk Anticipation Index (print edition) · NYU Ethical Tech CoLab

Publications · Academic report

The Civilian Evacuation Risk Anticipation Index

A Decision-Support Tool for Protecting Civilians During Evacuations in Armed Conflict

Ethical Tech CoLab

Advised by Teresa Cantero

July 2026

Development advised by Teresa Cantero, PhD candidate in Spain and an alum of the NYU Center for Global Affairs, as part of masters research on International Humanitarian Law and artificial intelligence.

When war forces people to flee, the decision to move a community or shelter it in place is made under extreme time pressure, with incomplete information, and with lives hanging on the outcome. CERAI keeps two questions apart that other tools tend to blend together: how dangerous it is for civilians to remain where they are, and whether an organised evacuation is actually possible right now. Keeping them apart matters legally, because the obligation to protect arises from the danger civilians face, not from whether an evacuation happens to be operationally easy.

Key figures

75%

endangerment threshold on the gauge, marking conditions that resemble those triggering Article 49 obligations

3

connected dimensions: endangerment, feasibility, and vulnerability, each scored separately

4

external live data sources connected to the prototype, covering location, weather, conflict events, and news

7

limitations the prototype states about itself, including that its weights remain provisional

00

Foreword

When war forces people to flee their homes, the decisions that follow are among the most consequential a humanitarian actor or a government can make. Should a community be moved now, or is it safer to shelter in place? Is a corridor genuinely open, or open only on paper? Who among the population, such as children, the elderly, and the sick, will be least able to make the journey? These questions are usually answered under extreme time pressure, with incomplete information, and with lives hanging on the outcome.

The Civilian Evacuation Risk Anticipation Index, known as CERAI, is a research prototype built to make this reasoning more structured, more transparent, and more open to scrutiny. This report explains, in non-technical language, what CERAI is, what it does, how it works, and, just as important, what it does not claim to do. It is intended for a policy, humanitarian, and legal audience rather than a technical one.

01

Executive Summary

CERAI is a research prototype, an experimental software tool, that helps analysts think through the risks civilians face during an evacuation in an armed conflict. It does not make decisions. It organises evidence and expert judgment into a clear, repeatable structure so that human decision-makers can see how a conclusion was reached.

The tool asks a set of structured questions across several categories of risk: the intensity of fighting, the proximity of armed groups, the condition of escape routes, the vulnerability of the affected population, the weather, the availability of food and water, and the capacity of the place people would move to. It then combines these inputs into clear summary scores.

CERAI's central and most distinctive idea is to keep two questions separate that other tools tend to blend together. Endangerment asks how dangerous it is for civilians to remain where they are. Feasibility asks whether an organised, safe evacuation is actually possible right now.

Keeping these apart matters for legal reasons. Under International Humanitarian Law, the obligation to protect and consider moving civilians arises from the danger they face, not from whether an evacuation happens to be convenient or operationally easy. A situation that is extremely dangerous but where evacuation is currently blocked does not remove the obligation to protect civilians; it makes the case for urgent political and diplomatic action even stronger.

The tool is explicitly aligned with well-established humanitarian and legal frameworks, and it draws comparisons to a database of documented real-world evacuation operations since the year 2000, for example the sieges of Aleppo and Mariupol, the battle of Mosul, and the evacuation of Kabul.

CERAI is delivered as a single self-contained web page that runs in an ordinary browser and is published as a live public demonstration. It is intended for academic demonstration only and is not authorised for operational use without expert validation.

The tool was developed as part of masters research at the New York University Center for Global Affairs, focused on the intersection of International Humanitarian Law and artificial intelligence, with the stated aim of making evacuation-risk reasoning more transparent and auditable for humanitarian and legal audiences such as the United Nations and the International Committee of the Red Cross.

02

Background and Rationale

The problem. Evacuation decisions in conflict are high-stakes and made under pressure. Experienced humanitarian coordinators rely heavily on accumulated judgment and memory of past operations. This expertise is invaluable, but it is often implicit, held in the minds of individuals, and therefore difficult to review, to explain to others, or to reproduce consistently across different teams and situations.

The gap. Existing humanitarian severity tools, such as the widely used INFORM Severity Index, are designed to measure the overall intensity of a humanitarian crisis. They are not built specifically around the legal and operational logic of an evacuation decision, and they tend to merge the question of how bad the situation is with the question of how accessible it is. For an evacuation decision, blending these can be misleading and, in a legal sense, dangerous.

The response. CERAI proposes a purpose-built framework for the evacuation question. Its guiding principle is that a good decision-support tool should be transparent, so that every number can be traced back to the inputs and reasoning that produced it; structured, so that it forces the analyst to consider the right questions in the right order; and reproducible, so that the same inputs always produce the same outputs and the reasoning can be checked by others.

CERAI describes itself as a structured aid rather than a successor to expert judgment. It preserves what works in expert practice, namely recognising patterns from past operations, while making that pattern recognition visible and open to challenge.

03

Objectives

The tool is designed to do the following.

  • Provide a transparent and repeatable method for assessing civilian risk during a potential evacuation.
  • Separate the assessment of danger, or endangerment, from the assessment of operational possibility, or feasibility, so that legal obligations are not obscured by practical constraints.
  • Make explicit how population vulnerability, including the presence of children, elderly people, people at risk of gender-based violence, and those with limited resources or mobility, changes the real level of risk faced by a community.
  • Anchor each assessment against documented historical cases, surfacing which past operations most resemble the current situation.
  • Communicate the resulting analysis in a form useful to humanitarian coordinators, political decision-makers, and legal reviewers, including explicit references to the relevant IHL obligations.
  • Be honest about uncertainty, by showing how confident the assessment is and how sensitive the conclusion is to assumptions.

04

How CERAI Works

The tool is organised around three connected dimensions. A user works through them by answering structured questions, and the tool presents the results on a set of easy-to-read dials and summaries.

Dimension 1, endangerment. How dangerous is it to remain? The analyst provides information across a range of risk factors, including the following.

  • The intensity of ongoing hostilities.
  • How close armed groups are to the civilian population.
  • Indicators that an attack may be imminent.
  • The risk of the conflict escalating.
  • The presence of chemical or biological threats.
  • Non-conflict hazards, such as extreme weather or disease.
  • The availability of food, water, and energy.
  • The proportion of the population that is especially vulnerable.

Each factor is weighted according to how central it is to the corresponding legal obligation. For example, the intensity of hostilities carries the largest single weight, reflecting the core IHL protection of civilians from attack. The tool combines these into a single endangerment score, displayed on a gauge with a clearly marked threshold.

Dimension 2, feasibility. Is an organised evacuation possible now? This dimension assesses whether an evacuation could actually be carried out safely, broken into three sub-questions.

  • Zone exit asks whether people can get out of the danger zone at all: is a corridor open, have armed groups consented, are routes mined, is an attack imminent.
  • Route conditions asks whether the journey is survivable, covering weather, daylight, availability of transport, mines along the route, and the distance to safety.
  • Protection at destination asks whether the place people are moving to is actually safe and able to receive them, covering shelter, medical capacity, security, willingness to receive, and consent of relevant actors.

Each sub-question produces its own score, and the three are combined into a composite feasibility score.

Dimension 3, vulnerability. Who is most at risk, and how does that change the picture? The same objective situation is not equally dangerous for everyone. A route that an able-bodied adult can walk may be impassable for young children, the elderly, or the chronically ill. This dimension captures the make-up of the affected population and produces an adjustment multiplier that can raise or lower the risk picture within defined bounds.

Importantly, vulnerability affects both of the other dimensions, but for different reasons: it increases the harm a given hazard causes, which is endangerment, and it reduces the navigability of a given route, which is feasibility. The tool treats these as two distinct effects rather than double-counting the same factor.

Beyond the scores, CERAI generates a written briefing aimed at humanitarian and political decision-makers. It highlights which IHL obligations are triggered, what the scores mean in practice, and where the greatest uncertainties or blocking constraints lie.

05

Reading the Results

Risk gauges. The endangerment and feasibility scores are shown on colour-coded dials running from low risk in green to extreme risk in deep red and purple, with intermediate bands for moderate, elevated, high, and critical.

The 75 per cent threshold. The endangerment gauge carries a marker at 75 per cent. This represents the point at which conditions are considered to demand evacuation under Article 49 of the Fourth Geneva Convention, which requires evacuation where the security of the population or imperative military reasons so demand. The threshold is indicative only. It flags that conditions resemble those which have historically triggered legal obligations; it does not itself create a legal obligation, and human legal judgment is always required.

Confidence bar. The tool shows how much of the assessment is based on real, entered data versus cautious worst-case default assumptions. When information is missing, CERAI deliberately assumes a higher-risk default, in keeping with the precautionary logic of IHL, and lowers the displayed confidence accordingly. A higher confidence figure means the conclusion rests on specific evidence rather than assumptions.

Trajectory and time-to-threshold. The tool tracks how the danger score is changing over the course of a session, whether deteriorating, stable, or improving, and offers a rough projection of how many days it might take to reach the 75 per cent obligation threshold if current trends continue. This is described explicitly as a planning aid only, since real situations rarely change in a straight line.

Historical comparisons. CERAI compares the current situation against a database of documented evacuation operations and surfaces the closest historical analogues. This mirrors how experienced coordinators reason, saying that a situation looks like Mariupol or resembles Mosul, but it makes that comparison explicit and open to challenge.

06

Testing Assumptions

Monte Carlo sensitivity analysis. Because many inputs are uncertain, CERAI can run a simulation that repeatedly varies the uncertain inputs within plausible ranges and observes how often the conclusion changes. This shows the decision-maker whether the assessment is robust or whether it hinges on a few uncertain assumptions.

What-if pressure test. The tool identifies the handful of variables that most strongly influence the outcome and lets the user adjust them to see the effect immediately. This supports scenario planning without altering the main assessment.

07

Connections to Live Data

To move beyond manual entry, the prototype connects to several external, publicly available data sources.

  • Location look-up, using Nominatim and OpenStreetMap, converts a place name into coordinates.
  • Weather, using Open-Meteo, retrieves current weather conditions for the assessed location.
  • Conflict events, using ACLED, retrieves recorded conflict-event data, which can drive a data-based trajectory calculation.
  • News signals, using GDELT, surface recent news articles relevant to the location.

The documentation also describes planned future connections to established humanitarian data systems, including UN OCHA and ReliefWeb access monitoring, the EU Global Conflict Risk Index, FEWS NET and IPC food-security data, and the INFORM Severity Index, as pathways to make the tool more automatically data-driven over time.

Each manually entered figure can also be tagged with a source credibility level, ranging from unverified, through media and NGO reporting, to UN and ICRC verified, and with a data-freshness indicator, so that the reliability and age of the underlying evidence are visible in the assessment itself.

08

Grounding in International Humanitarian Law

A defining feature of CERAI is that its structure is tied directly to specific IHL obligations. The weighting of each risk factor is justified by reference to the legal provision it relates to.

Examples cited in the tool include the following.

  • Fourth Geneva Convention, Article 49: evacuation where security demands.
  • Additional Protocol I, Article 51: protection of civilians from attack.
  • Additional Protocol I, Article 57: precautions in attack, including effective advance warning.
  • Additional Protocol I, Article 58: passive precautions, keeping civilians away from military objectives.
  • Fourth Geneva Convention, Article 54, and Additional Protocol I, Article 70: prohibition of starvation and provision of relief.
  • Special protections for children and the elderly: Fourth Geneva Convention, Articles 16 and 24, and Additional Protocol I, Articles 77 and 78.
  • Principles of non-refoulement and the safety of the receiving location.

The tool references established humanitarian doctrine and coordination frameworks from bodies including the ICRC, UN OCHA, UNHCR, WHO, IOM, and UN mine-action services, as well as the Ottawa Treaty on landmines and the Chemical Weapons Convention.

The legal design principle is stated plainly: low feasibility must never be allowed to appear to extinguish the obligation to protect civilians. A high-danger, low-feasibility situation is precisely the case that should escalate political engagement rather than be quietly resolved by an algorithm.

09

Methodological Choices

Why a geometric mean rather than a simple average. CERAI deliberately combines sub-scores in a way that penalises a single very bad factor heavily. In plain terms: if the escape corridor is essentially closed, the overall feasibility should be critically low no matter how good the weather is. A simple average would let good factors cancel out a fatal blocking constraint; CERAI's approach does not allow that.

Why vulnerability is a multiplier. Vulnerability changes how the objective conditions actually affect people, so it is applied as an adjustment within bounded limits to the objective scores rather than being averaged in alongside them as just another added factor.

Where the weights come from. The current weights were assigned by the researcher based on a review of IHL doctrine and checked for face validity against documented historical cases. The tool is candid that this is the same evidentiary standard used by established operational indices such as INFORM Severity in their early stages, and that a more rigorous, expert-panel-based weighting exercise using the Analytic Hierarchy Process is a planned next step.

What the historical cases are for. The database of past operations is used as a reference check on whether the tool's outputs align with the retrospective consensus about how dangerous those situations were, not as a statistical training set. The tool is explicit that no true statistical ground truth exists for civilian evacuation decisions.

10

Limitations and Caveats

The prototype is unusually transparent about what it cannot do. Its stated limitations are reproduced here because they are the most important part of the document for any reader considering what weight to give the tool.

It cannot read political intent. Sudden shifts in the behaviour of warring parties, or the collapse of an agreement, are outside the model.

It is only as current as its inputs. Static entries cannot capture rapidly changing battlefield conditions in real time.

It does not model armed-group behaviour. Command fragmentation among armed groups is likewise outside its scope.

The historical dataset is small. It holds a few dozen documented cases. This is enough to sanity-check the framework but not enough for robust statistical generalisation. CERAI is a structured analytic framework, not a statistical prediction model.

Correlation is not causation. Alignment between inputs and past outcomes does not prove a causal relationship.

It cannot always distinguish voluntary from coerced evacuation. The difference between a population choosing to leave and a population being displaced under duress is not something the model reliably detects.

The weights are provisional. They remain pending expert-panel validation.

Above all, the tool repeatedly states that its outputs are indicative and are not a substitute for human judgment, legal expertise, or operational assessment by qualified humanitarian and IHL professionals.

11

Practical Nature of the Tool

CERAI is delivered as a single self-contained web page that runs in an ordinary web browser, with no installation required. It can be opened directly or served locally, and it is published as a live public demonstration via GitHub Pages.

Users can save snapshots of an assessment, export results, and load pre-built example scenarios spanning moderate to extreme risk bands, making the tool suitable for teaching, demonstration, and structured discussion.

12

Intended Audience and Use

CERAI is aimed at humanitarian coordinators, political decision-makers, and legal reviewers, the people who must weigh whether and how to move civilians, and who must be able to justify those decisions afterward.

Its value is framed not as predictive accuracy but as decision support: making reasoning visible, structured, reproducible, and open to challenge. It is designed to complement, not replace, the expertise of ICRC delegates, UN coordinators, and IHL practitioners.

13

Conclusion

CERAI represents a thoughtful attempt to bring structure and transparency to one of the hardest categories of decision in armed conflict: whether, when, and how to move civilians out of danger. Its most important contribution is conceptual rather than technical. It insists on separating the danger of remaining from the possibility of leaving, so that the legal duty to protect civilians is never quietly overridden by operational difficulty.

The tool is deliberately modest about its status. It is a research prototype, its weights are provisional, its dataset is small, and it is not authorised for operational use without expert validation. But by making every score traceable, every assumption visible, and every historical comparison explicit, it offers a model for how automated tools can support humanitarian and legal judgment responsibly, assisting human experts while keeping them firmly in control of the decision.

References

Sources

  1. 01International Committee of the Red Cross. Article 49, Geneva Convention (IV) relative to the Protection of Civilian Persons in Time of War, 1949. Evacuation where the security of the population or imperative military reasons so demand.
  2. 02International Committee of the Red Cross. Article 54, Geneva Convention (IV), 1949.
  3. 03International Committee of the Red Cross. Articles 16 and 24, Geneva Convention (IV), 1949. Special protections for the wounded, the sick, children, and the elderly.
  4. 04International Committee of the Red Cross. Article 51, Additional Protocol I, 1977. Protection of the civilian population from attack.
  5. 05International Committee of the Red Cross. Article 57, Additional Protocol I, 1977. Precautions in attack, including effective advance warning.
  6. 06International Committee of the Red Cross. Article 58, Additional Protocol I, 1977. Precautions against the effects of attacks.
  7. 07International Committee of the Red Cross. Article 70, Additional Protocol I, 1977. Relief actions.
  8. 08International Committee of the Red Cross. Articles 77 and 78, Additional Protocol I, 1977. Protection of children and evacuation of children.
  9. 09Convention on the Prohibition of the Use, Stockpiling, Production and Transfer of Anti-Personnel Mines and on their Destruction (Ottawa Treaty), 1997.
  10. 10Organisation for the Prohibition of Chemical Weapons. Chemical Weapons Convention, 1993.
  11. 11United Nations Mine Action Service. Mine-action coordination and survey standards.
  12. 12ACAPS and the Joint Research Centre of the European Commission. INFORM Severity Index.
  13. 13ACLED. Armed Conflict Location and Event Data, conflict events and fatalities.
  14. 14Open-Meteo. Current weather conditions for the assessed location.
  15. 15Nominatim and OpenStreetMap. Place name to coordinate look-up.
  16. 16The GDELT Project. Global news signals relevant to the assessed location.
  17. 17UN OCHA. ReliefWeb, humanitarian access monitoring and situation reporting.
  18. 18FEWS NET. Famine Early Warning Systems Network, food-security outlooks.
  19. 19Integrated Food Security Phase Classification. IPC acute food insecurity analysis.
  20. 20European Commission Joint Research Centre. Global Conflict Risk Index (GCRI).

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file +

Publications · Academic report

The Civilian Evacuation Risk Anticipation Index

A Decision-Support Tool for Protecting Civilians During Evacuations in Armed Conflict

Ethical Tech CoLab

Advised by Teresa Cantero

July 2026

Development advised by Teresa Cantero, PhD candidate in Spain and an alum of the NYU Center for Global Affairs, as part of masters research on International Humanitarian Law and artificial intelligence.

When war forces people to flee, the decision to move a community or shelter it in place is made under extreme time pressure, with incomplete information, and with lives hanging on the outcome. CERAI keeps two questions apart that other tools tend to blend together: how dangerous it is for civilians to remain where they are, and whether an organised evacuation is actually possible right now. Keeping them apart matters legally, because the obligation to protect arises from the danger civilians face, not from whether an evacuation happens to be operationally easy.

Key figures

75%

endangerment threshold on the gauge, marking conditions that resemble those triggering Article 49 obligations

3

connected dimensions: endangerment, feasibility, and vulnerability, each scored separately

4

external live data sources connected to the prototype, covering location, weather, conflict events, and news

7

limitations the prototype states about itself, including that its weights remain provisional

00

Foreword

When war forces people to flee their homes, the decisions that follow are among the most consequential a humanitarian actor or a government can make. Should a community be moved now, or is it safer to shelter in place? Is a corridor genuinely open, or open only on paper? Who among the population, such as children, the elderly, and the sick, will be least able to make the journey? These questions are usually answered under extreme time pressure, with incomplete information, and with lives hanging on the outcome.

The Civilian Evacuation Risk Anticipation Index, known as CERAI, is a research prototype built to make this reasoning more structured, more transparent, and more open to scrutiny. This report explains, in non-technical language, what CERAI is, what it does, how it works, and, just as important, what it does not claim to do. It is intended for a policy, humanitarian, and legal audience rather than a technical one.

01

Executive Summary

CERAI is a research prototype, an experimental software tool, that helps analysts think through the risks civilians face during an evacuation in an armed conflict. It does not make decisions. It organises evidence and expert judgment into a clear, repeatable structure so that human decision-makers can see how a conclusion was reached.

The tool asks a set of structured questions across several categories of risk: the intensity of fighting, the proximity of armed groups, the condition of escape routes, the vulnerability of the affected population, the weather, the availability of food and water, and the capacity of the place people would move to. It then combines these inputs into clear summary scores.

CERAI's central and most distinctive idea is to keep two questions separate that other tools tend to blend together. Endangerment asks how dangerous it is for civilians to remain where they are. Feasibility asks whether an organised, safe evacuation is actually possible right now.

Keeping these apart matters for legal reasons. Under International Humanitarian Law, the obligation to protect and consider moving civilians arises from the danger they face, not from whether an evacuation happens to be convenient or operationally easy. A situation that is extremely dangerous but where evacuation is currently blocked does not remove the obligation to protect civilians; it makes the case for urgent political and diplomatic action even stronger.

The tool is explicitly aligned with well-established humanitarian and legal frameworks, and it draws comparisons to a database of documented real-world evacuation operations since the year 2000, for example the sieges of Aleppo and Mariupol, the battle of Mosul, and the evacuation of Kabul.

CERAI is delivered as a single self-contained web page that runs in an ordinary browser and is published as a live public demonstration. It is intended for academic demonstration only and is not authorised for operational use without expert validation.

The tool was developed as part of masters research at the New York University Center for Global Affairs, focused on the intersection of International Humanitarian Law and artificial intelligence, with the stated aim of making evacuation-risk reasoning more transparent and auditable for humanitarian and legal audiences such as the United Nations and the International Committee of the Red Cross.

02

Background and Rationale

The problem. Evacuation decisions in conflict are high-stakes and made under pressure. Experienced humanitarian coordinators rely heavily on accumulated judgment and memory of past operations. This expertise is invaluable, but it is often implicit, held in the minds of individuals, and therefore difficult to review, to explain to others, or to reproduce consistently across different teams and situations.

The gap. Existing humanitarian severity tools, such as the widely used INFORM Severity Index, are designed to measure the overall intensity of a humanitarian crisis. They are not built specifically around the legal and operational logic of an evacuation decision, and they tend to merge the question of how bad the situation is with the question of how accessible it is. For an evacuation decision, blending these can be misleading and, in a legal sense, dangerous.

The response. CERAI proposes a purpose-built framework for the evacuation question. Its guiding principle is that a good decision-support tool should be transparent, so that every number can be traced back to the inputs and reasoning that produced it; structured, so that it forces the analyst to consider the right questions in the right order; and reproducible, so that the same inputs always produce the same outputs and the reasoning can be checked by others.

CERAI describes itself as a structured aid rather than a successor to expert judgment. It preserves what works in expert practice, namely recognising patterns from past operations, while making that pattern recognition visible and open to challenge.

03

Objectives

The tool is designed to do the following.

  • Provide a transparent and repeatable method for assessing civilian risk during a potential evacuation.
  • Separate the assessment of danger, or endangerment, from the assessment of operational possibility, or feasibility, so that legal obligations are not obscured by practical constraints.
  • Make explicit how population vulnerability, including the presence of children, elderly people, people at risk of gender-based violence, and those with limited resources or mobility, changes the real level of risk faced by a community.
  • Anchor each assessment against documented historical cases, surfacing which past operations most resemble the current situation.
  • Communicate the resulting analysis in a form useful to humanitarian coordinators, political decision-makers, and legal reviewers, including explicit references to the relevant IHL obligations.
  • Be honest about uncertainty, by showing how confident the assessment is and how sensitive the conclusion is to assumptions.

04

How CERAI Works

The tool is organised around three connected dimensions. A user works through them by answering structured questions, and the tool presents the results on a set of easy-to-read dials and summaries.

Dimension 1, endangerment. How dangerous is it to remain? The analyst provides information across a range of risk factors, including the following.

  • The intensity of ongoing hostilities.
  • How close armed groups are to the civilian population.
  • Indicators that an attack may be imminent.
  • The risk of the conflict escalating.
  • The presence of chemical or biological threats.
  • Non-conflict hazards, such as extreme weather or disease.
  • The availability of food, water, and energy.
  • The proportion of the population that is especially vulnerable.

Each factor is weighted according to how central it is to the corresponding legal obligation. For example, the intensity of hostilities carries the largest single weight, reflecting the core IHL protection of civilians from attack. The tool combines these into a single endangerment score, displayed on a gauge with a clearly marked threshold.

Dimension 2, feasibility. Is an organised evacuation possible now? This dimension assesses whether an evacuation could actually be carried out safely, broken into three sub-questions.

  • Zone exit asks whether people can get out of the danger zone at all: is a corridor open, have armed groups consented, are routes mined, is an attack imminent.
  • Route conditions asks whether the journey is survivable, covering weather, daylight, availability of transport, mines along the route, and the distance to safety.
  • Protection at destination asks whether the place people are moving to is actually safe and able to receive them, covering shelter, medical capacity, security, willingness to receive, and consent of relevant actors.

Each sub-question produces its own score, and the three are combined into a composite feasibility score.

Dimension 3, vulnerability. Who is most at risk, and how does that change the picture? The same objective situation is not equally dangerous for everyone. A route that an able-bodied adult can walk may be impassable for young children, the elderly, or the chronically ill. This dimension captures the make-up of the affected population and produces an adjustment multiplier that can raise or lower the risk picture within defined bounds.

Importantly, vulnerability affects both of the other dimensions, but for different reasons: it increases the harm a given hazard causes, which is endangerment, and it reduces the navigability of a given route, which is feasibility. The tool treats these as two distinct effects rather than double-counting the same factor.

Beyond the scores, CERAI generates a written briefing aimed at humanitarian and political decision-makers. It highlights which IHL obligations are triggered, what the scores mean in practice, and where the greatest uncertainties or blocking constraints lie.

05

Reading the Results

Risk gauges. The endangerment and feasibility scores are shown on colour-coded dials running from low risk in green to extreme risk in deep red and purple, with intermediate bands for moderate, elevated, high, and critical.

The 75 per cent threshold. The endangerment gauge carries a marker at 75 per cent. This represents the point at which conditions are considered to demand evacuation under Article 49 of the Fourth Geneva Convention, which requires evacuation where the security of the population or imperative military reasons so demand. The threshold is indicative only. It flags that conditions resemble those which have historically triggered legal obligations; it does not itself create a legal obligation, and human legal judgment is always required.

Confidence bar. The tool shows how much of the assessment is based on real, entered data versus cautious worst-case default assumptions. When information is missing, CERAI deliberately assumes a higher-risk default, in keeping with the precautionary logic of IHL, and lowers the displayed confidence accordingly. A higher confidence figure means the conclusion rests on specific evidence rather than assumptions.

Trajectory and time-to-threshold. The tool tracks how the danger score is changing over the course of a session, whether deteriorating, stable, or improving, and offers a rough projection of how many days it might take to reach the 75 per cent obligation threshold if current trends continue. This is described explicitly as a planning aid only, since real situations rarely change in a straight line.

Historical comparisons. CERAI compares the current situation against a database of documented evacuation operations and surfaces the closest historical analogues. This mirrors how experienced coordinators reason, saying that a situation looks like Mariupol or resembles Mosul, but it makes that comparison explicit and open to challenge.

06

Testing Assumptions

Monte Carlo sensitivity analysis. Because many inputs are uncertain, CERAI can run a simulation that repeatedly varies the uncertain inputs within plausible ranges and observes how often the conclusion changes. This shows the decision-maker whether the assessment is robust or whether it hinges on a few uncertain assumptions.

What-if pressure test. The tool identifies the handful of variables that most strongly influence the outcome and lets the user adjust them to see the effect immediately. This supports scenario planning without altering the main assessment.

07

Connections to Live Data

To move beyond manual entry, the prototype connects to several external, publicly available data sources.

  • Location look-up, using Nominatim and OpenStreetMap, converts a place name into coordinates.
  • Weather, using Open-Meteo, retrieves current weather conditions for the assessed location.
  • Conflict events, using ACLED, retrieves recorded conflict-event data, which can drive a data-based trajectory calculation.
  • News signals, using GDELT, surface recent news articles relevant to the location.

The documentation also describes planned future connections to established humanitarian data systems, including UN OCHA and ReliefWeb access monitoring, the EU Global Conflict Risk Index, FEWS NET and IPC food-security data, and the INFORM Severity Index, as pathways to make the tool more automatically data-driven over time.

Each manually entered figure can also be tagged with a source credibility level, ranging from unverified, through media and NGO reporting, to UN and ICRC verified, and with a data-freshness indicator, so that the reliability and age of the underlying evidence are visible in the assessment itself.

08

Grounding in International Humanitarian Law

A defining feature of CERAI is that its structure is tied directly to specific IHL obligations. The weighting of each risk factor is justified by reference to the legal provision it relates to.

Examples cited in the tool include the following.

  • Fourth Geneva Convention, Article 49: evacuation where security demands.
  • Additional Protocol I, Article 51: protection of civilians from attack.
  • Additional Protocol I, Article 57: precautions in attack, including effective advance warning.
  • Additional Protocol I, Article 58: passive precautions, keeping civilians away from military objectives.
  • Fourth Geneva Convention, Article 54, and Additional Protocol I, Article 70: prohibition of starvation and provision of relief.
  • Special protections for children and the elderly: Fourth Geneva Convention, Articles 16 and 24, and Additional Protocol I, Articles 77 and 78.
  • Principles of non-refoulement and the safety of the receiving location.

The tool references established humanitarian doctrine and coordination frameworks from bodies including the ICRC, UN OCHA, UNHCR, WHO, IOM, and UN mine-action services, as well as the Ottawa Treaty on landmines and the Chemical Weapons Convention.

The legal design principle is stated plainly: low feasibility must never be allowed to appear to extinguish the obligation to protect civilians. A high-danger, low-feasibility situation is precisely the case that should escalate political engagement rather than be quietly resolved by an algorithm.

09

Methodological Choices

Why a geometric mean rather than a simple average. CERAI deliberately combines sub-scores in a way that penalises a single very bad factor heavily. In plain terms: if the escape corridor is essentially closed, the overall feasibility should be critically low no matter how good the weather is. A simple average would let good factors cancel out a fatal blocking constraint; CERAI's approach does not allow that.

Why vulnerability is a multiplier. Vulnerability changes how the objective conditions actually affect people, so it is applied as an adjustment within bounded limits to the objective scores rather than being averaged in alongside them as just another added factor.

Where the weights come from. The current weights were assigned by the researcher based on a review of IHL doctrine and checked for face validity against documented historical cases. The tool is candid that this is the same evidentiary standard used by established operational indices such as INFORM Severity in their early stages, and that a more rigorous, expert-panel-based weighting exercise using the Analytic Hierarchy Process is a planned next step.

What the historical cases are for. The database of past operations is used as a reference check on whether the tool's outputs align with the retrospective consensus about how dangerous those situations were, not as a statistical training set. The tool is explicit that no true statistical ground truth exists for civilian evacuation decisions.

10

Limitations and Caveats

The prototype is unusually transparent about what it cannot do. Its stated limitations are reproduced here because they are the most important part of the document for any reader considering what weight to give the tool.

It cannot read political intent. Sudden shifts in the behaviour of warring parties, or the collapse of an agreement, are outside the model.

It is only as current as its inputs. Static entries cannot capture rapidly changing battlefield conditions in real time.

It does not model armed-group behaviour. Command fragmentation among armed groups is likewise outside its scope.

The historical dataset is small. It holds a few dozen documented cases. This is enough to sanity-check the framework but not enough for robust statistical generalisation. CERAI is a structured analytic framework, not a statistical prediction model.

Correlation is not causation. Alignment between inputs and past outcomes does not prove a causal relationship.

It cannot always distinguish voluntary from coerced evacuation. The difference between a population choosing to leave and a population being displaced under duress is not something the model reliably detects.

The weights are provisional. They remain pending expert-panel validation.

Above all, the tool repeatedly states that its outputs are indicative and are not a substitute for human judgment, legal expertise, or operational assessment by qualified humanitarian and IHL professionals.

11

Practical Nature of the Tool

CERAI is delivered as a single self-contained web page that runs in an ordinary web browser, with no installation required. It can be opened directly or served locally, and it is published as a live public demonstration via GitHub Pages.

Users can save snapshots of an assessment, export results, and load pre-built example scenarios spanning moderate to extreme risk bands, making the tool suitable for teaching, demonstration, and structured discussion.

12

Intended Audience and Use

CERAI is aimed at humanitarian coordinators, political decision-makers, and legal reviewers, the people who must weigh whether and how to move civilians, and who must be able to justify those decisions afterward.

Its value is framed not as predictive accuracy but as decision support: making reasoning visible, structured, reproducible, and open to challenge. It is designed to complement, not replace, the expertise of ICRC delegates, UN coordinators, and IHL practitioners.

13

Conclusion

CERAI represents a thoughtful attempt to bring structure and transparency to one of the hardest categories of decision in armed conflict: whether, when, and how to move civilians out of danger. Its most important contribution is conceptual rather than technical. It insists on separating the danger of remaining from the possibility of leaving, so that the legal duty to protect civilians is never quietly overridden by operational difficulty.

The tool is deliberately modest about its status. It is a research prototype, its weights are provisional, its dataset is small, and it is not authorised for operational use without expert validation. But by making every score traceable, every assumption visible, and every historical comparison explicit, it offers a model for how automated tools can support humanitarian and legal judgment responsibly, assisting human experts while keeping them firmly in control of the decision.

References

Sources

  1. 01International Committee of the Red Cross. Article 49, Geneva Convention (IV) relative to the Protection of Civilian Persons in Time of War, 1949. Evacuation where the security of the population or imperative military reasons so demand.
  2. 02International Committee of the Red Cross. Article 54, Geneva Convention (IV), 1949.
  3. 03International Committee of the Red Cross. Articles 16 and 24, Geneva Convention (IV), 1949. Special protections for the wounded, the sick, children, and the elderly.
  4. 04International Committee of the Red Cross. Article 51, Additional Protocol I, 1977. Protection of the civilian population from attack.
  5. 05International Committee of the Red Cross. Article 57, Additional Protocol I, 1977. Precautions in attack, including effective advance warning.
  6. 06International Committee of the Red Cross. Article 58, Additional Protocol I, 1977. Precautions against the effects of attacks.
  7. 07International Committee of the Red Cross. Article 70, Additional Protocol I, 1977. Relief actions.
  8. 08International Committee of the Red Cross. Articles 77 and 78, Additional Protocol I, 1977. Protection of children and evacuation of children.
  9. 09Convention on the Prohibition of the Use, Stockpiling, Production and Transfer of Anti-Personnel Mines and on their Destruction (Ottawa Treaty), 1997.
  10. 10Organisation for the Prohibition of Chemical Weapons. Chemical Weapons Convention, 1993.
  11. 11United Nations Mine Action Service. Mine-action coordination and survey standards.
  12. 12ACAPS and the Joint Research Centre of the European Commission. INFORM Severity Index.
  13. 13ACLED. Armed Conflict Location and Event Data, conflict events and fatalities.
  14. 14Open-Meteo. Current weather conditions for the assessed location.
  15. 15Nominatim and OpenStreetMap. Place name to coordinate look-up.
  16. 16The GDELT Project. Global news signals relevant to the assessed location.
  17. 17UN OCHA. ReliefWeb, humanitarian access monitoring and situation reporting.
  18. 18FEWS NET. Famine Early Warning Systems Network, food-security outlooks.
  19. 19Integrated Food Security Phase Classification. IPC acute food insecurity analysis.
  20. 20European Commission Joint Research Centre. Global Conflict Risk Index (GCRI).

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file diff --git a/static-site/print/cerai/index.txt b/static-site/print/cerai/index.txt index d8b6195c3..579be757c 100644 --- a/static-site/print/cerai/index.txt +++ b/static-site/print/cerai/index.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -12:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","print","cerai",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","cerai","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -14:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -16:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","print","cerai",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","cerai","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:"$Sreact.suspense" -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1c:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -29,7 +29,7 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 19:[] 10:"$W19" 11:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1e:Ta14, @page { size: 210mm 297mm; margin: 20mm 18mm; } @@ -136,6 +136,6 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 34:["$","li","18",{"children":[["$","span",null,{"className":"print-ref-num","children":"19"}],["$","span",null,{"children":"Integrated Food Security Phase Classification. IPC acute food insecurity analysis."}]]}] 35:["$","li","19",{"children":[["$","span",null,{"className":"print-ref-num","children":"20"}],["$","span",null,{"children":"European Commission Joint Research Centre. Global Conflict Risk Index (GCRI)."}]]}] 1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -36:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +36:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 18:null 1d:[["$","title","0",{"children":"The Civilian Evacuation Risk Anticipation Index (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L36","6",{}]] diff --git a/static-site/print/digital-provenance-passport/__next._full.txt b/static-site/print/digital-provenance-passport/__next._full.txt index 9d510b626..69d6abade 100644 --- a/static-site/print/digital-provenance-passport/__next._full.txt +++ b/static-site/print/digital-provenance-passport/__next._full.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -12:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","print","digital-provenance-passport",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","digital-provenance-passport","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -14:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -16:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","print","digital-provenance-passport",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","digital-provenance-passport","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:"$Sreact.suspense" -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1c:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -29,7 +29,7 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 19:[] 10:"$W19" 11:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1e:Ta14, @page { size: 210mm 297mm; margin: 20mm 18mm; } @@ -163,6 +163,6 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 52:["$","tr","5",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Registry match"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"40 points"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"High"}],["$","td","3",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"The paid commercial check returns a match against a stolen or repatriated record"}]]}] 53:["$","figcaption",null,{"className":"mt-3 text-sm text-muted","children":"The deductions used by the web pipeline. Each condition that matches produces a red flag carrying its category, its severity, and a sentence of evidence."}] 1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -57:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +57:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 18:null 1d:[["$","title","0",{"children":"The Digital Provenance Passport (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L57","6",{}]] diff --git a/static-site/print/digital-provenance-passport/__next._head.txt b/static-site/print/digital-provenance-passport/__next._head.txt index 309fa0f86..d30eefeea 100644 --- a/static-site/print/digital-provenance-passport/__next._head.txt +++ b/static-site/print/digital-provenance-passport/__next._head.txt @@ -1,6 +1,6 @@ 1:"$Sreact.fragment" -2:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -3:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +2:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +3:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 4:"$Sreact.suspense" -5:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +5:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 0:{"rsc":["$","$1","h",{"children":[null,["$","$L2",null,{"children":[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]]}],["$","div",null,{"hidden":true,"children":["$","$L3",null,{"children":["$","$4",null,{"name":"Next.Metadata","children":[["$","title","0",{"children":"The Digital Provenance Passport (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L5","6",{}]]}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/print/digital-provenance-passport/__next._index.txt b/static-site/print/digital-provenance-passport/__next._index.txt index f98a4fd90..18b9d82b6 100644 --- a/static-site/print/digital-provenance-passport/__next._index.txt +++ b/static-site/print/digital-provenance-passport/__next._index.txt @@ -1,13 +1,13 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -d:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +d:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} 8:["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$Ld",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$Ld",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$Ld",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$Ld",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] diff --git a/static-site/print/digital-provenance-passport/__next._tree.txt b/static-site/print/digital-provenance-passport/__next._tree.txt index 85957c93c..fc7a4fb06 100644 --- a/static-site/print/digital-provenance-passport/__next._tree.txt +++ b/static-site/print/digital-provenance-passport/__next._tree.txt @@ -1,7 +1,7 @@ -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] 0:{"tree":{"name":"","param":null,"prefetchHints":16,"slots":{"children":{"name":"print","param":null,"prefetchHints":0,"slots":{"children":{"name":"slug","param":{"type":"d","key":"digital-provenance-passport","siblings":null},"prefetchHints":0,"slots":{"children":{"name":"__PAGE__","param":null,"prefetchHints":0,"slots":null}}}}}}},"staleTime":300,"buildId":"static-snapshot"} diff --git a/static-site/print/digital-provenance-passport/__next.print.txt b/static-site/print/digital-provenance-passport/__next.print.txt index 460afe242..fa617e997 100644 --- a/static-site/print/digital-provenance-passport/__next.print.txt +++ b/static-site/print/digital-provenance-passport/__next.print.txt @@ -1,4 +1,4 @@ 1:"$Sreact.fragment" -2:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -3:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +2:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +3:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] 0:{"rsc":["$","$1","c",{"children":[null,[["$","style",null,{"children":"\n /* The site is dark-first, and a dark color-scheme makes Chrome paint\n the printed sheet's margin area dark too. Paper is light. */\n html { color-scheme: light !important; }\n html, body { background: #ffffff !important; }\n body::after { content: none !important; }\n header, footer, .site-bg, .aura { display: none !important; }\n "}],["$","$L2",null,{"parallelRouterKey":"children","template":["$","$L3",null,{}]}]]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/print/digital-provenance-passport/index.html b/static-site/print/digital-provenance-passport/index.html index c6b6221a0..6138c50a0 100644 --- a/static-site/print/digital-provenance-passport/index.html +++ b/static-site/print/digital-provenance-passport/index.html @@ -1,4 +1,4 @@ -The Digital Provenance Passport (print edition) · NYU Ethical Tech CoLab

Publications · Academic report

The Digital Provenance Passport

An Automated Assistant for Tracing the Ownership History of Artworks and Cultural Objects

Ethical Tech CoLab

NYU Center for Global Affairs

July 2026

Carolina Morón, with a project collaborator whose contributions are noted in the repository's source comments. Prepared as masters research at the NYU Center for Global Affairs.

Provenance research is slow, archival, and applied to only a small fraction of the objects that need it. The Digital Provenance Passport asks what part of the first pass could be done automatically without becoming untrustworthy. It refuses to record any claim that does not carry the address of its source, computes its judgments with arithmetic a non-programmer can read and dispute, and seals its output so that an assessment cannot be quietly improved after the fact.

Key figures

40

points, the largest single deduction, applied only when a paid registry check returns a match

1

of the nine stolen-art registers consulted that can actually be queried by machine, none of them a police register

14

countries on the hand-written source-country list, a shorthand for elevated restitution exposure

2

different confidence scoring systems in the repository, which will not agree on the same object

01

Executive Summary

The Digital Provenance Passport is an experimental piece of software that takes the name of an artwork or cultural object, searches a controlled list of authoritative sources for its ownership history, flags signals associated with looting, restitution claims, documentation gaps, and suspicious pricing, and then issues a sealed digital record of everything it found and everything it did.

The system is deliberately built so that it cannot state a fact about an object unless it can point to the source that asserted it. Any claim that arrives without a source address attached is discarded before it reaches the record. This is the prototype's central safeguard against the well-documented tendency of language models to produce fluent, confident, and entirely fictional history.

The prototype scores each object on a provenance confidence scale running from 0 to 100, where a high number means the ownership history appears well-documented and a low number means it does not. The score begins at a starting value and is reduced by named, itemised deductions, each of which is displayed alongside the evidence that triggered it. There is no hidden statistical model. The rules are a short, readable list.

The system is built as an agent, meaning that it carries out a sequence of steps on its own initiative rather than waiting for a person to click through each one. Among those steps is an unusual one: it can decide for itself whether a paid commercial due-diligence search is worth buying, and if so, pay for it. The prototype uses a small automated payment standard called x402 and a test currency on a test network, so that no real money is ever at stake.

The final output is called a Passport. It is a structured record containing the object's identity, the ownership events that were found, the source behind each one, the confidence score, the red flags, any paid checks that were run, and a cryptographic seal. If a single character of that record is later altered, the seal fails and the alteration becomes visible to anyone who checks. The prototype ships with a deliberate tamper test that demonstrates this.

The prototype is not a determination of title, a legal opinion, or a substitute for a qualified provenance researcher. It runs by default in a fully offline demonstration mode using stored example data, and several of its most eye-catching features, including the commercial stolen-art database search, are simulated stand-ins rather than connections to the real service.

The work was produced as part of masters research at the New York University Center for Global Affairs, under the Ethical Tech CoLab, and was built for a hackathon on autonomous software that transacts on its own behalf. Its ambitions should be read accordingly: it is a demonstration of a method, not a deployed service.

02

Background and Rationale

The problem. Provenance research is done by hand, and it is slow. A researcher works through auction and sale catalogues, dealer stock books, household inventories, shipping and customs records, exhibition histories, correspondence, and the physical object itself, reading the labels, stamps, inscriptions, and collectors' marks on its reverse. The Getty Research Institute's Provenance Index, one of the principal reference tools in the field, holds more than two million records drawn from precisely this kind of archival material. Assembling a single credible chain of ownership can take weeks or months, and for many objects the chain can never be completed at all.

The volume problem. The number of objects that need this treatment vastly exceeds the number of specialists able to provide it. Museums hold collections running to hundreds of thousands of items each. Auction houses process consignments continuously. Buyers frequently need an answer in days. In practice, most objects that move through the market receive nothing approaching a full provenance investigation, and the first line of defence is a quick check against a stolen-art database rather than a reconstruction of where the object has actually been.

The consequences of getting it wrong are severe and asymmetric. An object with a gap in its record may be perfectly legitimate, or it may be the product of a crime that the gap exists precisely to conceal. Museums have returned major works decades after acquiring them, at considerable cost to their finances and their standing, because the questions were not asked at the point of purchase.

The gap this prototype addresses. Existing tools tend to sit at one of two extremes. Commercial databases, such as the Art Loss Register, give a fast yes-or-no answer to a narrow question: does this object appear in a register of reported thefts? They are closed, they charge per search, and by design they cannot tell a buyer anything about an object that was never reported stolen, which includes almost everything looted from an archaeological site or taken under colonial rule. Full archival provenance research, at the other extreme, is thorough but too slow and too scarce to apply at the scale of the market.

The Digital Provenance Passport proposes an intermediate layer: an automated first pass that gathers what authoritative institutions have already published about an object, organises it into a dated chain, applies a fixed and inspectable set of warning rules, and hands the human researcher a structured starting point with every source attached. It is designed to tell a researcher where to look, not to tell a court what to conclude.

A guiding design commitment runs through the code: the assessment is produced by ordinary arithmetic on evidence that has already been gathered and cited, not by asking a language model for its opinion. The repository states this plainly, describing the scoring as a transparent, editable rubric rather than a black box. That choice is what makes the rest of the tool auditable.

03

What Agent Means in This Context

The word agent has a specific meaning in software, and it is not the meaning it carries in law or in the art trade. Here it does not mean a person acting for a principal, and it does not mean a dealer's representative.

An agent, in this sense, is a program that is given a goal rather than a list of instructions, and that then chooses its own sequence of actions to reach that goal. A conventional program is a recipe: do this, then this, then stop. An agent is closer to an instruction to a research assistant: find out what you can about this object, use whatever reference tools you judge appropriate, and report back with your sources.

The practical difference is that an agent can decide, mid-task, to take an action nobody explicitly told it to take on this occasion. In this prototype the clearest example is the paid check. Having assessed the object once, the system looks at its own confidence figure, looks at the price of a commercial database search, looks at the spending limit it has been given, and decides whether buying that search is worth doing. If it decides yes, it pays and incorporates the result. If it decides no, it records why not and continues.

Two things follow from this that matter to a non-technical reader. First, an agent's autonomy is only ever as safe as the limits placed around it, which is why the spending caps are a substantive part of the design rather than a technical footnote. Second, an agent that can write is an agent that can write something untrue, which is why the sourcing rule is placed before every other consideration in the pipeline.

04

Objectives

The prototype is designed to:

  1. Reduce the time required to produce a first-pass provenance summary for a named object, from days of archival work to a single automated run.
  2. Make unsourced history structurally impossible to record, by refusing to enter any claim that does not carry the address of the source that made it, so that every claim can be traced to and checked against its origin. The limit belongs with the claim: the rule blocks an unsourced claim, not a false one, and section 13 sets out what it does not guarantee.
  3. Express the reliability of a provenance record as a single explicit number, accompanied by the itemised reasons that number is not higher.
  4. Surface the specific categories of concern that matter in cultural-property law and in anti-money-laundering practice: undocumented periods, looting and restitution signals, origin in a country with active repatriation claims, and prices that bear no relation to an object's market value.
  5. Demonstrate that an automated researcher can obtain paid due-diligence data on its own initiative, under a spending limit, and disclose in the final record exactly what it bought and what it paid.
  6. Produce a record that is portable and tamper-evident, so that an assessment can be passed between a museum, a buyer, an insurer, and a regulator without any of them having to trust the others not to have edited it.
  7. Show its working. Every score, flag, source, and payment decision appears in the output, and the underlying rules are short enough to be read and disputed by a non-programmer with the code in front of them.

05

How the System Works

The prototype runs a fixed sequence of five stages. Each stage passes its findings to the next, and the record that is eventually sealed is assembled from what the stages actually collected, not from a summary written afterwards.

Stage one, intent. The user supplies what they know about the object. Only the title is required. The optional fields are the artist, the country or culture of origin, any ownership history the user is already aware of, an asking price if the object is being bought or sold, and an estimated market value for comparison. The repository calls this stage intent is the interface, meaning that the user states what they want to know in ordinary terms and the system works out the research steps from there. The inputs are trimmed and normalised into a structured request, and a search phrase is built from them by appending the terms provenance ownership history repatriation to whatever the user typed.

Stage two, grounding. Grounding is the term used in the repository for the step that anchors the assessment in published evidence. The system sends its search phrase to a commercial web-search service called Tavily, which is a search tool built for software rather than for people: it returns clean extracted text with the address of each page it came from. Critically, the search is confined to a fixed list of permitted sources. Each result is converted into a single dated event in the object's ownership timeline, carrying the claim, the date if one is available, the place if one is available, the address of the source, the name of the institution vouching for it, and a reliability tier. Any result lacking a source address is counted and discarded, and the repository records the number of discarded claims explicitly, so that the user can see how much was thrown away.

Stage three, risk assessment. The assembled timeline, together with the user's own inputs, is examined against a short list of warning conditions. Each condition that matches produces a red flag, which is a small record containing a category, a severity of low, medium, or high, and a sentence of evidence explaining what triggered it. The same pass produces the provenance confidence score. The design principle stated in the code is that the system flags signals and cites evidence, and never asserts a legal conclusion. It does not say an object was looted. It says that authoritative sources discussing this object contain looting or restitution language, and shows the sentence.

Stage four, the paid check. The prototype then considers whether to buy a premium search of a commercial stolen-art registry. It weighs three things: the price of the check, its remaining budget, and whether its current confidence figure is already conclusive enough that a further check could not change the picture. It records its reasoning in plain sentences that appear in the interface, for example that confidence is already high with no red flags and a paid check is therefore not worth the money. If it decides to buy, it settles the payment automatically using x402. The result of the check is folded back into the assessment, and if it reports a match against a stolen or repatriated record the confidence score is cut sharply and a further red flag is added.

Stage five, the Passport. Everything accumulated across the four preceding stages is assembled into a single structured record and sealed. Throughout the run, each stage announces itself to the web interface as it happens, so that a user watches the reasoning unfold in sequence rather than receiving a finished answer with no visible derivation.

06

The Variables Explained

Everything the prototype measures can be described in ordinary terms: what it represents, why it was chosen, and how it changes the result.

The inputs. Title is the only required field, and the anchor for every subsequent search. Artist is optional and used to disambiguate, and it matters far more for paintings than for antiquities, which are usually anonymous and identified by culture and period instead. Origin is the country or culture the object came from, and it is checked directly against a list of countries with active restitution claims. Known history is whatever the user has already been told, treated as text to be scanned for warning language rather than as established fact, which is the correct treatment: the story a seller tells is evidence about the seller as much as about the object. Asking price and estimated market value together enable the pricing check; supplying neither disables it.

TierWhat it meansWeight
Verified by authorityA museum record, a UNESCO document, a registry entry, or a government record0.92
Reported in pressA news outlet or secondary coverage0.60
InferredDrawn by the software from cited context, and the least trustworthy of the three0.30
The three reliability tiers. Every fact recorded carries one of them, describing how strongly it is backed rather than how important it is. Tiers are assigned mechanically from the address of the source: a page on the Metropolitan Museum's domain is treated as verified by authority, and anything outside the recognised institutions falls to reported in press.

The gap between the first two weights is deliberate and large. It encodes the judgment that an institution's own record of an object it holds is a materially different kind of evidence from a newspaper account of the same object, and the drop to 0.30 for inference marks the point at which a reader should stop treating a line as a finding.

The confidence score. The headline number runs from 0 to 100, and it measures how well-documented the provenance appears, not how valuable or how important the object is. A high score means a clean, traceable record. A low score means gaps, warning signals, or both. The score is a measure of the evidence, not a verdict on the object.

The repository contains two implementations of this score, reflecting the fact that the project was built by more than one hand under time pressure. The web interface uses a deduction model. The score starts at 100, on the presumption that an object is well documented until something suggests otherwise, and named deductions are subtracted for each warning condition found, with the final figure held within the range 0 to 100. The command-line agent uses an accumulation model, which runs in the opposite direction. It starts at 30, a deliberately low base representing an object about which nothing has yet been established, and adds 18 points for every event confirmed by an authoritative institution and 8 points for every event reported in the press, capped at 100. It then subtracts 12 if the early history is undated or incomplete.

The two models express the same intuition from opposite ends. The deduction model asks what is wrong with this record. The accumulation model asks how much of this record has actually been established. The second is the more conservative of the two, since an object with no published history at all scores 30 rather than 100.

Following peer review, the accumulation model is now designated the canonical scorer, and the deduction model in the web pipeline is marked non-canonical in the source. Any score reported as a result comes from the canonical model. The reason accumulation wins is that an object with no published history should not score 100: a model that starts every object at a perfect score treats absence of evidence as evidence of clean provenance, which inverts the purpose of the tool. Scoring behaviour was deliberately left unchanged in that revision, so no previously reported number has silently moved, and the two models still return different figures for the same object. Reconciling the web pipeline onto the canonical model is recorded as committed future work rather than presented as an interesting property of the system. Neither model has been validated against expert judgment.

Red flagDeductionSeverityWhat triggers it
Provenance gap25 pointsHighLanguage indicating an undocumented period: undocumented, unknown owner, gap, missing, no record, and the years 1933, 1939, and 1945
Source-country origin15 pointsMediumThe stated origin, or the gathered text, mentions one of fourteen countries with active repatriation claims
Looting signal20 pointsHighLooting or illicit-trade vocabulary: looted, stolen, tomb, excavated, smuggled, repatriated, illicit, or tombaroli
Valuation outlier30 pointsHighAn asking price more than three times the stated market estimate
High value with no comparison10 pointsLowAn asking price above ten million United States dollars supplied with no market estimate to check it against
Registry match40 pointsHighThe paid commercial check returns a match against a stolen or repatriated record
The deductions used by the web pipeline. Each condition that matches produces a red flag carrying its category, its severity, and a sentence of evidence.

The three years named in the gap rule are not arbitrary: they mark the Nazi accession to power, the outbreak of the Second World War in Europe, and its end, which is the period the Washington Principles are concerned with. The gap rule carries the second-largest deduction because a gap is the single most common signature of an illegitimate acquisition. Objects with clean histories tend to have documented ones.

The fourteen source countries are Italy, Greece, Egypt, Turkey, Cambodia, China, Iraq, Peru, Mexico, Nigeria, India, Syria, Cyprus, and Thailand. The list is a shorthand for elevated restitution exposure, and its deduction is modest by design. Originating in Egypt is not a defect in an object; it is a reason to check the paperwork more carefully.

The valuation rule carries the largest deduction of the language-based rules, and it is not really a provenance rule at all. It is a money-laundering rule. Art is a recognised vehicle for moving illicit value precisely because it has no fixed price, and a wildly inflated sale is a recognised laundering technique. The threshold of three times is a judgment call rather than a figure derived from evidence.

The registry match is the largest single movement in the system, and rightly so: it is the only deduction triggered by a direct hit in a dedicated register rather than by language found in general sources. The command-line implementation adds two further categories: a repatriation precedent flag when the timeline itself cites a return or restitution event, and a transaction flag recording the result of screening the system's own payment for exposure to sanctioned addresses.

The spending variables. Because the prototype can spend money on its own initiative, the numbers that constrain that behaviour are as important as the ones that score the object. The vendor price is set by default to five United States cents, the cost of one premium search from the simulated vendor. The maximum spend per run is set by default to twenty-five cents, a hard ceiling: if the price of a check plus what has already been spent would exceed it, the check is refused and the refusal is recorded in the output. The conclusive threshold is set at 90 in the command-line agent, above which the agent will not buy a check at all, on the reasoning that the result could not change the conclusion. The web pipeline applies a comparable rule at a confidence of 85 combined with a complete absence of red flags. The design intention behind these three numbers is that an autonomous purchaser should have a price it knows, a ceiling it cannot cross, and a standard of sufficiency at which it stops buying. It is worth noting the distance between the demonstration and reality here. A real single search of the Art Loss Register has been priced in the region of sixty pounds sterling plus tax, not five cents, and at that price the economic reasoning the agent performs would carry considerably more weight.

The coverage class, and why the score is never shown alone. The confidence score is, mechanically, a measure of how much published evidence the system managed to find. That is a defensible proxy where the documentary record is dense and close to meaningless where it is not, and the density is least favourable in exactly the places this project exists to worry about. Two opposite situations produce the same low number. The first is absence within coverage: a Dutch painting sits inside auction catalogues, dealer stock books and decades of Nazi-era provenance research, so a hole in that record is itself evidence, because records would be expected to exist. The second is absence of coverage: a Cambodian temple sculpture removed during a civil war was never accessioned, catalogued or reported stolen, because no institution was in a position to report it, and it cannot appear in a stolen-property register at all. Finding nothing about it establishes nothing. The prototype therefore computes, separately from the score, which registers could in principle have named the object, and classifies it as well covered, partly covered, or structurally uncovered. The score is displayed with that classification attached and declared comparable only within it. Two decisions matter. The region used is the jurisdiction of the loss, not of manufacture and not the current location, because only the first determines which national register could hold the object; the Euphronios Krater was made in Athens, looted in Italy and held in New York, and only the Italian answer reaches the archive that recovered it. And coverage is never folded into the score, because adjusting the number by how much could be reached would produce a single figure meaning two things again, which is the defect being corrected. The effect can be read off the catalogue: the Getty Bronze scores 26 and is well covered, with an enforceable Italian confiscation order standing against it, while the Rosetta Stone scores 34 and is structurally uncovered, because no register in the set can hold a colonial-era seizure from Egypt. The two figures sit close together and mean almost opposite things. A structurally uncovered classification is an assertion about the register landscape rather than a finding about the object, and is neither exoneration nor accusation.

The demonstration mode switch. One further variable governs everything else. The system has a master setting with two positions. In mock mode, which is the default, every external connection is replaced by stored example data: no search is performed, no payment is made, and no network is required. In live mode, the system attempts real connections and quietly falls back to the stored data if any of them fail. This is candid engineering for a stage demonstration, and the reason the default is the safe one is stated in the code: so that the prototype never moves funds unless someone has explicitly asked it to. But it has an important consequence for any reader evaluating the tool. Unless the setting has been changed and real credentials supplied, an impressive-looking run is a replay of a stored example rather than live research. Both the fallback behaviour and the mode in use are disclosed on screen and in the output.

07

Reading the Results

The prototype opens on a grid of objects it already tracks, each shown as a card with a confidence figure, a repatriation status of repatriated, contested, or clear, its current holding institution, and the number of stops on its journey. Clicking a card opens that object's full dashboard.

A caution about the label on those cards. The interface prints the confidence figure with the word risk in front of it, so the Euphronios Krater appears as risk 12 out of 100. The underlying number is a cleanliness score, and 12 out of 100 means a badly compromised record, not a low level of risk. A reader who has not been told this will read the card exactly backwards. The colour coding and the accompanying risk level of high or low are correct; only the wording is misleading. It is a small defect with real potential to confuse, and it is recorded here because a plain-language review should say so.

The central display for each object is the sequence of places it has physically been, each stop carrying a year, a place, a country, a description, and a type drawn from a fixed set: origin, excavation, looting, sale, museum, repatriation, or contested. This turns a paragraph of prose into a route that can be read at a glance, and it makes the shape of a problematic history visible. An object that goes from a tomb, to a dealer in a neutral-jurisdiction city, to a major museum, within eighteen months and with no documented owner before that, has a recognisable silhouette.

Each red flag is displayed with its severity and the sentence of evidence that produced it. No flag appears without its reason. Every institution consulted is listed with the address of the page relied on, so that a researcher can go and read it. This is what distinguishes a first-pass research aid from an answer machine: the output is designed to be checked, and the checking is made easy. A search box runs the full agent on any object the user names, with the stages appearing one by one as they complete, including the payment reasoning.

The prototype ships with five real and well-documented cases, chosen to span the range of outcomes: the Euphronios Krater, looted from an Etruscan tomb near Cerveteri in 1971, bought by the Metropolitan Museum of Art in 1972 and returned to Italy in 2008; a Benin Bronze plaque taken during the British punitive expedition of 1897 and still held in London against an active Nigerian claim; the Rosetta Stone, ceded to Britain under the Capitulation of Alexandria in 1801 and the subject of repeated Egyptian requests; the Lydian Hoard, looted from burial mounds in western Turkey in the 1960s and returned by the Metropolitan Museum in 1993; and John Singer Sargent's Madame X, purchased from the artist in 1916 with an unbroken documented history. The last of these is the control case. It is the example of what a clean record looks like, and it scores 93 out of 100.

08

Data Sources and the Permitted-Source List

The single most consequential design decision in the prototype is that its research is confined to a fixed list of permitted sources. The search is restricted to the domains of:

  • The Metropolitan Museum of Art
  • UNESCO, including its World Heritage Centre
  • The Art Loss Register
  • The International Council of Museums
  • United States government cultural-heritage sites

The purpose of the restriction is to make the tier system meaningful. If the system may search anywhere, then a claim in a collector's forum post and a claim in a museum's own accession record arrive in the same shape and the distinction between them has to be reconstructed after the fact. By searching only recognised institutions, the prototype makes the authority of a claim a property of where it was found.

The named institutions are worth introducing for a reader unfamiliar with the field. UNESCO is the United Nations Educational, Scientific and Cultural Organization, and is the custodian of the principal international instrument on trafficking in cultural property. The International Council of Museums is the global professional body for museums, and publishes the Red Lists that identify categories of object at particular risk of illicit trafficking from specific regions. The Art Loss Register is a private London company that operates the largest commercial database of stolen art and conducts due-diligence checks for auction houses, dealers, insurers, and law enforcement, reportedly running several hundred thousand checks a year. The Metropolitan Museum of Art is included both as a collection record and, unavoidably, as a party to several of the best-documented restitution cases in the field.

The cost of the restriction is that the list is short, and its omissions are substantial. It does not include the Getty Provenance Index, the German Lost Art Foundation's database, Interpol's stolen works of art database, the Art Institute of Chicago, Europeana, or the national heritage authorities of any of the fourteen countries the system itself identifies as source countries. An object well documented in a French or Turkish archive and absent from these six domains will return little, and the tool will report a thin record rather than an absent one. The two are not the same thing, and the prototype does not currently distinguish them.

That omission is not evenly distributed, and the direction of the bias runs against the tool's own motivation. The case for building this system, set out in sections 2 and 11, rests on the objects that stolen-art registers cannot catch: material taken from an archaeological site or under colonial rule, never inventoried and never reported stolen. But the permitted list is five Western institutions and one commercial theft register. The tool therefore searches best where objects are already well documented, which is to say major Western museum holdings, and searches worst exactly where the motivating harm lives, which is source-country archives and colonial-era and archaeological material. The observation in section 13 that the system does not distinguish thin evidence from absent evidence is the symptom; this is the cause, and the two belong together. The consequence is that a low confidence score for a Cambodian sculpture and a low score for a Dutch painting do not mean the same thing, and the tool does not currently say so.

Adding source-country heritage authorities, the Getty Provenance Index, and the German Lost Art Foundation is therefore the first substantive extension of this work rather than an optional one. Every other improvement to the scoring, the interface, or the passport format operates on evidence the tool was able to find, and the coverage bias determines what it can find at all.

The register layer, and why it is not a set of lookups. That extension has since been partly built. The permitted list now also covers Interpol, the Federal Bureau of Investigation, the Carabinieri Command for the Protection of Cultural Heritage in Italy, the German Lost Art Foundation, and the Getty Research Institute, and the prototype puts each object to nine named registers rather than to the allowlist as an undifferentiated whole. The extension is smaller than it sounds, for a reason that governs everything the layer can honestly claim: the registers that actually certify stolen cultural property have no public programmatic interface. This was tested rather than assumed.

What each register actually permits, established by trying rather than by reading the marketing:

  • Interpol's Stolen Works of Art database, holding on the order of fifty-two thousand records of certified police information, is reachable through the ID-Art mobile application or through an account applied for and vetted by the applicant's national Interpol bureau. Neither route can be automated.
  • The FBI's National Stolen Art File is published at a web address that refuses any client which is not a browser. The Bureau's one open programmatic interface covers wanted persons and contains no art records at all; a filter requesting art crime returns the fugitive list without indicating that it has ignored the request.
  • The Carabinieri archive, at over a million objects the largest in the world, and the German Lost Art database are public search forms in HTML with no documented data interface.
  • The Art Loss Register is a commercial service, and in this prototype is the paid check described in section 9 rather than a free lookup.
  • Wikidata's structured-data endpoint is the single exception, and is genuinely queried. Its event statements carry dated values such as archaeological looting, art theft, restitution, and claim for restitution.

The layer therefore does not pretend to search those registers. Where a source is genuinely open to machine query it queries it. Where a register publishes material openly but cannot be queried, the tool searches that register's website, which is a different thing from searching its register and is labelled as such. Where neither is possible, it emits a link to the official search a person must run by hand, together with the address at which credentialed access may be applied for. Every result is presented with the access route that produced it, because the route is what makes the result interpretable: the words no evidence found mean one thing after a genuine query of a structured source and something much weaker after a keyword search of an institution's public pages.

The strongest negative the system can express is no evidence found. There is deliberately no verdict meaning clear or not stolen, and the type that represents a register result does not contain one, so no future change can produce that conclusion by accident. A stolen-property register can only contain objects that somebody was in a position to report missing. Material removed from an archaeological site, or taken under colonial administration, was never inventoried and never reported, and therefore cannot appear in such a register at all. For precisely the objects this project was built to worry about, absence from the registers is the expected condition rather than a reassuring one, and a tool that rendered it as a clean bill of health would be worse than no tool.

Scoring follows the same asymmetry. A register hit reduces confidence, a register that returns nothing adds none, and registers that could not be searched at all are counted and reported as a coverage gap, so that a check which reached two sources cannot be mistaken for one that reached nine. Each check, including those that failed and those that never ran, is written into the signed Passport together with the sentence describing what its verdict does and does not license, so that the qualification cannot be separated from the finding by anyone who handles the credential later.

The watchlist. The prototype also publishes a list of several hundred works recorded as stolen or plundered, drawn from Wikidata's theft statements and regenerated by script, so that the tool is not limited to the fifteen objects researched by hand. It is not an extract from the Interpol or FBI databases, and the interface says so at the point of display rather than in a footnote. It is community-maintained data, unreviewed and uneven in coverage. An entry is a lead to be checked against the official registers rather than a register hit, and the dates require particular care, because Wikidata sometimes attaches the restitution date to the theft statement, so a recent year on a Nazi-plunder record frequently marks a return rather than a taking. Absence from the list carries no information at all. Two decisions in its construction are worth recording, because both were errors first. The query is constrained to works of art, because without that constraint it also returns people and companies, since spoliation records attach theft events to the dispossessed as well as to their property, and a list presented as artworks that silently contained the names of victims would be both wrong and offensive. And the query had to be widened to reach events that are instances of art theft rather than subclasses of it, because named incidents such as the Isabella Stewart Gardner Museum theft are recorded as instances, and the original formulation silently omitted every object taken in a named robbery, including the most famous stolen painting in the world.

What remains missing is what mattered most at the start of this section and still does. The national heritage authorities of the fourteen source countries the system itself recognises are still absent, and no amount of register plumbing substitutes for them. The layer is built so that this shows in its output rather than being smoothed over, which is the most that can be claimed for it.

09

The Payment Layer, and Why It Is There

A reader may reasonably ask what an automated payment system is doing in a cultural-heritage research tool. The answer is partly historical: the prototype was built for a hackathon on autonomous software commerce, and the payment capability was a condition of entry. But the underlying argument is a real one and worth stating on its merits.

The most authoritative sources on stolen art are commercial and closed. The Art Loss Register does not publish its database; access is sold per search. That model works when the requester is an auction house with a subscription and an accounts department. It works badly when the requester is a researcher checking one object, a small museum, or a piece of software that has just discovered mid-task that it needs one specific answer. The friction is administrative rather than financial: the sum involved may be trivial, but obtaining an account, a key, and an invoice is not.

The x402 standard is an attempt to remove that friction. It uses the HTTP status code 402, Payment Required, reserved decades ago for exactly this purpose and left unused until recently. A server responds to a request by stating a price; the requester attaches a payment and asks again; the data is returned. There is no account and no subscription, and the exchange completes in seconds. The standard was published and open-sourced by Coinbase in 2025 and has since passed to neutral governance.

What that enables is described in the repository as paying for tools you discover: software that encounters a paywalled resource it did not know about in advance can evaluate the price and buy the answer, rather than failing and waiting for a human to arrange access. Applied to provenance, this is the difference between a research assistant that stops at the edge of the free sources and one that can obtain the commercial check when the free sources prove inconclusive.

The prototype implements this end to end, but against its own simulated vendor. The repository is explicit on the point: the paywall behaviour is genuine, in that the vendor really does refuse the request until payment settles, but the data behind it is stored example content, not the Art Loss Register's actual database. Payments are made in a test version of the USDC digital dollar on Base Sepolia, a practice network using valueless test tokens, and the repository warns in terms that the wallet must never be funded with real money. No commercial relationship with the Art Loss Register is claimed or implied.

The system also screens its own payment. Having paid, it examines the resulting transaction for exposure to sanctioned addresses or laundering patterns and records the verdict in the final document. In the demonstration this screening is illustrative rather than substantive, and the code says so. The reason it is present at all is that a tool used to detect value laundering through art should be able to demonstrate that its own transactions are clean.

10

The Passport and What the Seal Proves

The output document is issued in a standard format known as a Verifiable Credential, a specification maintained by the World Wide Web Consortium for digital records that can be checked by anyone without contacting the issuer. The format matters because a provenance assessment that only one institution's software can read is of limited use to the buyer, insurer, or regulator who receives it.

The record contains the object's identity, the full ownership timeline with a source address against every event, the confidence score, every red flag with its evidence, any paid checks with the amount paid and the transaction they produced, the list of checks performed, and the time of assessment. It is assembled from what the run actually collected. The code is explicit that the record is built from accumulated state rather than re-derived from a language model, so that the sealed document is always a faithful account of the run.

The sealing works as follows. The record is first written out in a strictly canonical form, with every field in a fixed order, so that the same content always produces an identical text. A cryptographic signature is then computed over that text. Anyone holding the document can repeat the calculation and recover the identity of the signer. If any character of the record has been changed since signing, the recovered identity will not match and the document is exposed as altered.

The prototype includes a demonstration of exactly this. Its example script signs a Passport, then alters a single field, raising the confidence score to 100, and re-checks it. The check fails and reports that the content no longer matches its seal. This is a small thing to build and a significant thing to show, because the value at stake in a provenance document is high enough to make quiet editing a genuine risk.

One conceptual point in the design deserves comment. The identity that signs the Passport is the same digital wallet identity that pays for the premium check. The repository summarises this as the observation that a wallet is already a form of public-key infrastructure, applied here to an object's identity rather than to a payment. The practical effect is that the entity that spent the money and the entity that vouched for the record are provably the same, and no separate system of certificates or key distribution is required.

What the seal proves, and does not prove, should be stated exactly. It proves that this document has not been altered since it was signed, and that it was signed by the holder of a particular key. It proves nothing whatever about whether the contents are true. A sealed record of a bad assessment is a bad assessment that cannot be quietly improved later. That is a genuine benefit, and it is a narrower one than the word verified tends to suggest to a general reader.

11

The Legal and Normative Background

The prototype's warning rules are not invented from nothing. They correspond, loosely but recognisably, to the concerns of the principal instruments in this field.

The 1970 UNESCO Convention. The Convention on the Means of Prohibiting and Preventing the Illicit Import, Export and Transfer of Ownership of Cultural Property was adopted by UNESCO in 1970 and entered into force in 1972. It has been ratified by 149 states. It requires states parties to establish national services for heritage protection and to legislate accordingly, to operate a system of export certificates, to prohibit the import of objects stolen from museums, religious institutions, and public monuments, and, under Article 7, to seize and return such objects when another state party requests it through diplomatic channels. Article 9 commits states to cooperate in concerted international action against trafficking. Its practical significance for provenance work is the date itself. The year 1970 functions as a dividing line in museum acquisition policy across much of the world: an antiquity that can be shown to have left its country of origin before 1970 sits in a different position from one that cannot. This is why the system's list of source countries, and its attention to undocumented periods, map onto a real and consequential legal boundary.

The Washington Principles on Nazi-Confiscated Art. Agreed on 3 December 1998 by more than forty states at a conference convened in Washington, these eleven non-binding principles address art confiscated by the Nazi regime and never returned. They call for such art to be identified, for archives and records to be opened to researchers, for resources to be committed to the work, for the results to be publicised so that pre-war owners and their heirs can be found, for a central registry of the information, and for claims to be resolved through a just and fair solution rather than left to the technicalities of ordinary property litigation. The fourth principle is the one most directly reflected in this software. It provides that in establishing whether a work was confiscated by the Nazis and not restituted, consideration should be given to unavoidable gaps or ambiguities in the provenance, in light of the passage of time and the circumstances of the Holocaust era. In ordinary terms: an incomplete record is not to be held against a claimant, because the destruction of records was part of what happened. This is precisely why the prototype treats a documentation gap as a signal deserving investigation rather than as an absence of evidence, and why the years 1933, 1939, and 1945 appear by name in its rules. The principles are declaratory and create no legal obligations. Their influence comes from adoption into national practice and institutional policy, and assessments of their record over the past quarter-century have been decidedly mixed.

The commercial registers. The Art Loss Register is the market's standard first-line check. Its limits are structural and well understood by practitioners, and they explain why a tool of this kind is worth building. The register records objects that were reported stolen. An object looted from an undocumented archaeological site was never inventoried and therefore was never reported, so it cannot appear. The same is true of most objects taken under colonial administration. A clean result from a stolen-art database is a meaningful answer to a narrow question and is routinely mistaken for a broader assurance than it can provide.

What the prototype does not do. It does not determine title, and it does not apply the law of any jurisdiction. Limitation periods, good-faith purchaser rules, and the availability of a remedy vary enormously between legal systems, and none of that is modelled. The code states the principle directly: the system flags signals and cites evidence, and never asserts a legal conclusion. That restraint is the correct posture for a research aid, and it should be preserved in any future development.

12

Methodological Choices

Why the scoring is arithmetic rather than learned. The prototype could have asked a language model to assign a provenance score directly. It does not. Every deduction is a written rule with a fixed value, applied to evidence that has already been gathered and cited. The reason is auditability. A number produced by a rule can be disputed by disputing the rule. A number produced by a model cannot be disputed at all, only accepted or rejected, and in a field where the output may inform a restitution claim that is not an acceptable property.

Why sourcing is enforced structurally rather than by instruction. It is possible to instruct a language model to cite its sources, and it is well-established that the instruction is sometimes ignored, including by inventing sources that do not exist. The prototype instead makes the citation a structural requirement of the data itself: a timeline event without a source address is discarded before it can be recorded, and the system counts the discards. This is a stronger guarantee than any instruction, and it is the single most defensible piece of engineering in the repository.

Why the search is restricted. The restriction is what allows reliability to be inferred from the provenance of the claim rather than assessed after the fact.

Why price is treated as provenance evidence. Including a valuation check in a provenance tool is an unusual choice, and a good one. Art is a recognised channel for moving illicit value, and a sale far above market is a recognised technique. The prototype treats an inexplicable price as a fact about the transaction in the same way it treats a gap as a fact about the record. The threshold of three times market value rests on judgment rather than on any published study.

Why failures fall back rather than stop. Every external connection in the system degrades to stored example data on error rather than halting. This was chosen so that a live demonstration cannot fail in front of an audience. It is a reasonable choice for its purpose and an unacceptable one for research use, because a tool that silently substitutes example data for real research will eventually be trusted when it should not be. The substitution is disclosed on screen, but disclosure is a weaker safeguard than refusal.

Why register access is tiered rather than flattened. The register layer could have reported a single verdict per register and kept the retrieval method internal. That would have produced a cleaner interface and a dishonest one. The nine registers are reached by four quite different routes of very unequal strength, and a verdict is only interpretable alongside the route that produced it. Presenting a keyword search of an institution's public pages in the same shape as a genuine query of a structured database would invite exactly the inference the system exists to prevent, which is that a quiet result is a reassuring one. The interface is therefore more cluttered than it would otherwise be, and the clutter is carrying the argument.

Why there is no verdict meaning clear. This is the same commitment as the structural enforcement of sourcing, applied to conclusions rather than to citations, and it is enforced the same way, in the type system rather than in a guideline. A register result can say that a match surfaced, that nothing surfaced through the access available, or that the register could not be reached. There is no fourth value, so no later change can produce a clean bill of health by inadvertence. The design principle behind both is that a rule which matters should be made impossible to violate rather than merely documented, since documentation is not read at the moment the mistake is made.

Why silence earns nothing. The scoring treats a register that returned nothing as contributing no confidence, and reports separately how many registers could not be searched at all. This is the same reasoning that made the accumulation model canonical over the deduction model: absence of evidence is not evidence of clean provenance. A system that rewarded quiet registers would score an object highest precisely when it had learned least about it, and the objects it learns least about are disproportionately the ones taken from places that never had an inventory to be missing from.

13

Limitations and Caveats

The repository and the paper both state these limits plainly. They are reproduced here because they are the most important part of the document for any reader considering what weight to give the tool.

It cannot read an archive. The overwhelming majority of provenance evidence is in physical archives, in dealer records, in correspondence, in auction catalogues that were never digitised, and on the back of the object itself. None of this is reachable by a web search. The prototype gathers what institutions have chosen to publish online, which for most objects is a small and unrepresentative fraction of what exists.

The keyword rules are crude. The system decides that a provenance gap exists by looking for particular words in the gathered text. A museum page stating that an object has no gaps in its record contains the word gap and will trigger the flag. A page discussing the successful repatriation of a different object will trigger the looting flag. These rules find the right signal often enough to be useful as a prompt for human attention, and they will produce false positives regularly. They should be read as a list of things to check, never as findings.

The source-country list is hand-written and incomplete. Fourteen countries are named. Many states with active restitution claims are not among them, and membership of the list is not a legal category. It is a useful shorthand and nothing more.

The thresholds are not empirically derived. Twenty-five points for a gap, fifteen for source-country origin, three times market value for a pricing anomaly, ninety as the point at which further checking is pointless: none of these figures comes from a study or an expert panel. They are the developer's judgment, plausibly calibrated and entirely unvalidated. This is a normal starting position for a prototype and it should not be mistaken for anything more.

There are two different scoring systems in the repository. The web interface and the command-line agent compute confidence in materially different ways and will not agree on the same object. Anyone building on this work should reconcile them before doing anything else.

The interface mislabels its own headline number. A card reading risk 12 out of 100 describes a severely compromised object, which is the opposite of what the phrasing suggests.

Most of what is impressive is simulated. In the default configuration there is no search, no payment, and no database query. The commercial stolen-art check is a stand-in with stored answers. The transaction screening returns a fixed clean verdict. The catalogue of five objects is written by hand from published cases rather than produced by the system. A reader evaluating a demonstration should establish which mode it is running in before drawing conclusions about capability.

The risk of plausible falsehood remains. This is the most serious caveat in the report. The prototype's sourcing rule is a genuine and unusually well-implemented safeguard, and it does not eliminate the problem. It guarantees that every recorded claim has a source address attached. It does not guarantee that the source says what the record claims it says, that the source is about the same object, that the extracted sentence has not lost the qualification that made it accurate, or that a page found on an authoritative domain is itself authoritative. A false provenance claim, expressed in institutional language, carrying a real link to a real museum, and sealed inside a document described as verifiable, is more dangerous than an obvious error, because every visible signal invites trust. This is a foreseeable failure mode of the design, not a hypothetical one. Every claim in a Passport must be read at its source before it is relied on, and the fact that the document is sealed and formally verifiable does nothing to change that.

It does not distinguish thin evidence from absent evidence. An object with no online record and an object with a clean record but no online presence look identical to this system. For the majority of the world's cultural objects, which are not held by major Western museums with published collection databases, this limitation is severe.

It is a prototype. It has no test suite, no validation against expert assessments, and a commit history spanning a few days of work. It should be read as a demonstration of a method and an argument about how such a tool ought to be built, not as a service.

14

Relationship to the Sibling Project

The Ethical Tech CoLab maintains a related repository, provenance-search, which addresses the same problem from a different direction. That project is a web application that identifies an artwork, including from a photograph taken on a phone in a museum, and searches a considerably wider set of free public sources: the Metropolitan Museum, the Art Institute of Chicago, a bundled extract of the Museum of Modern Art's open collection data, Wikipedia, Wikidata, and Europeana, alongside the same commercial search service used here.

The two projects share a central conviction, stated in both repositories, that the confidence score must be computed by a fixed algorithm rather than by the artificial intelligence, and both restrict their research to a list of recognised sources.

They differ in emphasis. The sibling project prioritises breadth of free sources and accessibility, including identification by camera. The project described in this report prioritises the integrity of the resulting record: the enforced sourcing rule, the evidence tiers, the tamper-evident seal, and the ability to obtain paid data autonomously. Read together they suggest a natural combination, in which the wider source coverage of the one feeds the evidentiary discipline and sealed output of the other. This report treats the present repository on its own terms; the comparison is offered only because a reader encountering both should understand that they are complementary experiments rather than duplicates.

15

Intended Audience and Use

The prototype is aimed at those who must form a view about an object's history before deciding what to do with it: museum registrars and provenance researchers, restitution and cultural-property lawyers, insurers and compliance staff at auction houses, customs and heritage-crime investigators, and the source-country authorities pursuing claims.

Its appropriate use is as a first pass. It is well suited to triage, meaning the task of deciding which objects in a large collection or a large consignment merit a researcher's time. It is well suited to preparing a starting dossier with sources already assembled and linked. It is well suited to teaching, since it makes the structure of provenance reasoning visible.

It is not suited to supporting a claim, defending an acquisition, satisfying a due-diligence obligation, or reaching any conclusion that will be acted upon without a qualified human having read the underlying sources.

16

Conclusion

The Digital Provenance Passport takes a body of work that is slow, specialised, and applied to only a small fraction of the objects that need it, and asks what part of the first pass could be done automatically without becoming untrustworthy. Its answer is more careful than most work in this area. It refuses to record any claim that does not carry its source. It searches only recognised institutions. It computes its judgments with arithmetic that a non-programmer can read and dispute. It seals its output so that the assessment cannot be quietly improved after the fact. It records what it bought, what it paid, and why it thought the purchase worthwhile.

The prototype is also, in its present state, mostly a demonstration. Its commercial database check is a stand-in, its default mode replays stored examples, its permitted-source list omits most of the field's principal resources, its keyword rules will misfire, its two scoring systems disagree with each other, and its interface labels its central number in a way that invites the opposite reading. None of this is concealed in the repository, and none of it should be concealed in a report about it.

What is durable here is the argument rather than the software. It is the proposition that when automated tools are pointed at questions where the stakes include the return of stolen property and the resolution of historical injustice, the design should make fabrication structurally impossible rather than merely discouraged, should make every judgment traceable to a rule and every rule open to challenge, and should produce a record that cannot be edited without the edit becoming visible. Those are the right commitments. This prototype implements them incompletely, states its own limits without flattering itself, and leaves a workable foundation for someone to build on.

References

Sources

  1. 01UNESCO. Convention on the Means of Prohibiting and Preventing the Illicit Import, Export and Transfer of Ownership of Cultural Property, 1970.
  2. 02United States Department of State. Washington Conference Principles on Nazi-Confiscated Art, 3 December 1998.
  3. 03The Art Loss Register. Commercial database of stolen art and due-diligence search service.
  4. 04International Council of Museums. Red Lists of cultural objects at risk.
  5. 05Getty Research Institute. Getty Provenance Index, holding more than two million records.
  6. 06The Metropolitan Museum of Art. Collection records, cited as a permitted source and as a party to several documented restitution cases.
  7. 07World Wide Web Consortium. Verifiable Credentials Data Model.
  8. 08Coinbase. x402, an open payment standard using HTTP status code 402, published 2025.
  9. 09Tavily. Web search service returning extracted text with source addresses, used for the grounding stage.
  10. 10Ethical Tech CoLab. provenance-search, the sibling repository described in Section 14.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file +

Publications · Academic report

The Digital Provenance Passport

An Automated Assistant for Tracing the Ownership History of Artworks and Cultural Objects

Ethical Tech CoLab

NYU Center for Global Affairs

July 2026

Carolina Morón, with a project collaborator whose contributions are noted in the repository's source comments. Prepared as masters research at the NYU Center for Global Affairs.

Provenance research is slow, archival, and applied to only a small fraction of the objects that need it. The Digital Provenance Passport asks what part of the first pass could be done automatically without becoming untrustworthy. It refuses to record any claim that does not carry the address of its source, computes its judgments with arithmetic a non-programmer can read and dispute, and seals its output so that an assessment cannot be quietly improved after the fact.

Key figures

40

points, the largest single deduction, applied only when a paid registry check returns a match

1

of the nine stolen-art registers consulted that can actually be queried by machine, none of them a police register

14

countries on the hand-written source-country list, a shorthand for elevated restitution exposure

2

different confidence scoring systems in the repository, which will not agree on the same object

01

Executive Summary

The Digital Provenance Passport is an experimental piece of software that takes the name of an artwork or cultural object, searches a controlled list of authoritative sources for its ownership history, flags signals associated with looting, restitution claims, documentation gaps, and suspicious pricing, and then issues a sealed digital record of everything it found and everything it did.

The system is deliberately built so that it cannot state a fact about an object unless it can point to the source that asserted it. Any claim that arrives without a source address attached is discarded before it reaches the record. This is the prototype's central safeguard against the well-documented tendency of language models to produce fluent, confident, and entirely fictional history.

The prototype scores each object on a provenance confidence scale running from 0 to 100, where a high number means the ownership history appears well-documented and a low number means it does not. The score begins at a starting value and is reduced by named, itemised deductions, each of which is displayed alongside the evidence that triggered it. There is no hidden statistical model. The rules are a short, readable list.

The system is built as an agent, meaning that it carries out a sequence of steps on its own initiative rather than waiting for a person to click through each one. Among those steps is an unusual one: it can decide for itself whether a paid commercial due-diligence search is worth buying, and if so, pay for it. The prototype uses a small automated payment standard called x402 and a test currency on a test network, so that no real money is ever at stake.

The final output is called a Passport. It is a structured record containing the object's identity, the ownership events that were found, the source behind each one, the confidence score, the red flags, any paid checks that were run, and a cryptographic seal. If a single character of that record is later altered, the seal fails and the alteration becomes visible to anyone who checks. The prototype ships with a deliberate tamper test that demonstrates this.

The prototype is not a determination of title, a legal opinion, or a substitute for a qualified provenance researcher. It runs by default in a fully offline demonstration mode using stored example data, and several of its most eye-catching features, including the commercial stolen-art database search, are simulated stand-ins rather than connections to the real service.

The work was produced as part of masters research at the New York University Center for Global Affairs, under the Ethical Tech CoLab, and was built for a hackathon on autonomous software that transacts on its own behalf. Its ambitions should be read accordingly: it is a demonstration of a method, not a deployed service.

02

Background and Rationale

The problem. Provenance research is done by hand, and it is slow. A researcher works through auction and sale catalogues, dealer stock books, household inventories, shipping and customs records, exhibition histories, correspondence, and the physical object itself, reading the labels, stamps, inscriptions, and collectors' marks on its reverse. The Getty Research Institute's Provenance Index, one of the principal reference tools in the field, holds more than two million records drawn from precisely this kind of archival material. Assembling a single credible chain of ownership can take weeks or months, and for many objects the chain can never be completed at all.

The volume problem. The number of objects that need this treatment vastly exceeds the number of specialists able to provide it. Museums hold collections running to hundreds of thousands of items each. Auction houses process consignments continuously. Buyers frequently need an answer in days. In practice, most objects that move through the market receive nothing approaching a full provenance investigation, and the first line of defence is a quick check against a stolen-art database rather than a reconstruction of where the object has actually been.

The consequences of getting it wrong are severe and asymmetric. An object with a gap in its record may be perfectly legitimate, or it may be the product of a crime that the gap exists precisely to conceal. Museums have returned major works decades after acquiring them, at considerable cost to their finances and their standing, because the questions were not asked at the point of purchase.

The gap this prototype addresses. Existing tools tend to sit at one of two extremes. Commercial databases, such as the Art Loss Register, give a fast yes-or-no answer to a narrow question: does this object appear in a register of reported thefts? They are closed, they charge per search, and by design they cannot tell a buyer anything about an object that was never reported stolen, which includes almost everything looted from an archaeological site or taken under colonial rule. Full archival provenance research, at the other extreme, is thorough but too slow and too scarce to apply at the scale of the market.

The Digital Provenance Passport proposes an intermediate layer: an automated first pass that gathers what authoritative institutions have already published about an object, organises it into a dated chain, applies a fixed and inspectable set of warning rules, and hands the human researcher a structured starting point with every source attached. It is designed to tell a researcher where to look, not to tell a court what to conclude.

A guiding design commitment runs through the code: the assessment is produced by ordinary arithmetic on evidence that has already been gathered and cited, not by asking a language model for its opinion. The repository states this plainly, describing the scoring as a transparent, editable rubric rather than a black box. That choice is what makes the rest of the tool auditable.

03

What Agent Means in This Context

The word agent has a specific meaning in software, and it is not the meaning it carries in law or in the art trade. Here it does not mean a person acting for a principal, and it does not mean a dealer's representative.

An agent, in this sense, is a program that is given a goal rather than a list of instructions, and that then chooses its own sequence of actions to reach that goal. A conventional program is a recipe: do this, then this, then stop. An agent is closer to an instruction to a research assistant: find out what you can about this object, use whatever reference tools you judge appropriate, and report back with your sources.

The practical difference is that an agent can decide, mid-task, to take an action nobody explicitly told it to take on this occasion. In this prototype the clearest example is the paid check. Having assessed the object once, the system looks at its own confidence figure, looks at the price of a commercial database search, looks at the spending limit it has been given, and decides whether buying that search is worth doing. If it decides yes, it pays and incorporates the result. If it decides no, it records why not and continues.

Two things follow from this that matter to a non-technical reader. First, an agent's autonomy is only ever as safe as the limits placed around it, which is why the spending caps are a substantive part of the design rather than a technical footnote. Second, an agent that can write is an agent that can write something untrue, which is why the sourcing rule is placed before every other consideration in the pipeline.

04

Objectives

The prototype is designed to:

  1. Reduce the time required to produce a first-pass provenance summary for a named object, from days of archival work to a single automated run.
  2. Make unsourced history structurally impossible to record, by refusing to enter any claim that does not carry the address of the source that made it, so that every claim can be traced to and checked against its origin. The limit belongs with the claim: the rule blocks an unsourced claim, not a false one, and section 13 sets out what it does not guarantee.
  3. Express the reliability of a provenance record as a single explicit number, accompanied by the itemised reasons that number is not higher.
  4. Surface the specific categories of concern that matter in cultural-property law and in anti-money-laundering practice: undocumented periods, looting and restitution signals, origin in a country with active repatriation claims, and prices that bear no relation to an object's market value.
  5. Demonstrate that an automated researcher can obtain paid due-diligence data on its own initiative, under a spending limit, and disclose in the final record exactly what it bought and what it paid.
  6. Produce a record that is portable and tamper-evident, so that an assessment can be passed between a museum, a buyer, an insurer, and a regulator without any of them having to trust the others not to have edited it.
  7. Show its working. Every score, flag, source, and payment decision appears in the output, and the underlying rules are short enough to be read and disputed by a non-programmer with the code in front of them.

05

How the System Works

The prototype runs a fixed sequence of five stages. Each stage passes its findings to the next, and the record that is eventually sealed is assembled from what the stages actually collected, not from a summary written afterwards.

Stage one, intent. The user supplies what they know about the object. Only the title is required. The optional fields are the artist, the country or culture of origin, any ownership history the user is already aware of, an asking price if the object is being bought or sold, and an estimated market value for comparison. The repository calls this stage intent is the interface, meaning that the user states what they want to know in ordinary terms and the system works out the research steps from there. The inputs are trimmed and normalised into a structured request, and a search phrase is built from them by appending the terms provenance ownership history repatriation to whatever the user typed.

Stage two, grounding. Grounding is the term used in the repository for the step that anchors the assessment in published evidence. The system sends its search phrase to a commercial web-search service called Tavily, which is a search tool built for software rather than for people: it returns clean extracted text with the address of each page it came from. Critically, the search is confined to a fixed list of permitted sources. Each result is converted into a single dated event in the object's ownership timeline, carrying the claim, the date if one is available, the place if one is available, the address of the source, the name of the institution vouching for it, and a reliability tier. Any result lacking a source address is counted and discarded, and the repository records the number of discarded claims explicitly, so that the user can see how much was thrown away.

Stage three, risk assessment. The assembled timeline, together with the user's own inputs, is examined against a short list of warning conditions. Each condition that matches produces a red flag, which is a small record containing a category, a severity of low, medium, or high, and a sentence of evidence explaining what triggered it. The same pass produces the provenance confidence score. The design principle stated in the code is that the system flags signals and cites evidence, and never asserts a legal conclusion. It does not say an object was looted. It says that authoritative sources discussing this object contain looting or restitution language, and shows the sentence.

Stage four, the paid check. The prototype then considers whether to buy a premium search of a commercial stolen-art registry. It weighs three things: the price of the check, its remaining budget, and whether its current confidence figure is already conclusive enough that a further check could not change the picture. It records its reasoning in plain sentences that appear in the interface, for example that confidence is already high with no red flags and a paid check is therefore not worth the money. If it decides to buy, it settles the payment automatically using x402. The result of the check is folded back into the assessment, and if it reports a match against a stolen or repatriated record the confidence score is cut sharply and a further red flag is added.

Stage five, the Passport. Everything accumulated across the four preceding stages is assembled into a single structured record and sealed. Throughout the run, each stage announces itself to the web interface as it happens, so that a user watches the reasoning unfold in sequence rather than receiving a finished answer with no visible derivation.

06

The Variables Explained

Everything the prototype measures can be described in ordinary terms: what it represents, why it was chosen, and how it changes the result.

The inputs. Title is the only required field, and the anchor for every subsequent search. Artist is optional and used to disambiguate, and it matters far more for paintings than for antiquities, which are usually anonymous and identified by culture and period instead. Origin is the country or culture the object came from, and it is checked directly against a list of countries with active restitution claims. Known history is whatever the user has already been told, treated as text to be scanned for warning language rather than as established fact, which is the correct treatment: the story a seller tells is evidence about the seller as much as about the object. Asking price and estimated market value together enable the pricing check; supplying neither disables it.

TierWhat it meansWeight
Verified by authorityA museum record, a UNESCO document, a registry entry, or a government record0.92
Reported in pressA news outlet or secondary coverage0.60
InferredDrawn by the software from cited context, and the least trustworthy of the three0.30
The three reliability tiers. Every fact recorded carries one of them, describing how strongly it is backed rather than how important it is. Tiers are assigned mechanically from the address of the source: a page on the Metropolitan Museum's domain is treated as verified by authority, and anything outside the recognised institutions falls to reported in press.

The gap between the first two weights is deliberate and large. It encodes the judgment that an institution's own record of an object it holds is a materially different kind of evidence from a newspaper account of the same object, and the drop to 0.30 for inference marks the point at which a reader should stop treating a line as a finding.

The confidence score. The headline number runs from 0 to 100, and it measures how well-documented the provenance appears, not how valuable or how important the object is. A high score means a clean, traceable record. A low score means gaps, warning signals, or both. The score is a measure of the evidence, not a verdict on the object.

The repository contains two implementations of this score, reflecting the fact that the project was built by more than one hand under time pressure. The web interface uses a deduction model. The score starts at 100, on the presumption that an object is well documented until something suggests otherwise, and named deductions are subtracted for each warning condition found, with the final figure held within the range 0 to 100. The command-line agent uses an accumulation model, which runs in the opposite direction. It starts at 30, a deliberately low base representing an object about which nothing has yet been established, and adds 18 points for every event confirmed by an authoritative institution and 8 points for every event reported in the press, capped at 100. It then subtracts 12 if the early history is undated or incomplete.

The two models express the same intuition from opposite ends. The deduction model asks what is wrong with this record. The accumulation model asks how much of this record has actually been established. The second is the more conservative of the two, since an object with no published history at all scores 30 rather than 100.

Following peer review, the accumulation model is now designated the canonical scorer, and the deduction model in the web pipeline is marked non-canonical in the source. Any score reported as a result comes from the canonical model. The reason accumulation wins is that an object with no published history should not score 100: a model that starts every object at a perfect score treats absence of evidence as evidence of clean provenance, which inverts the purpose of the tool. Scoring behaviour was deliberately left unchanged in that revision, so no previously reported number has silently moved, and the two models still return different figures for the same object. Reconciling the web pipeline onto the canonical model is recorded as committed future work rather than presented as an interesting property of the system. Neither model has been validated against expert judgment.

Red flagDeductionSeverityWhat triggers it
Provenance gap25 pointsHighLanguage indicating an undocumented period: undocumented, unknown owner, gap, missing, no record, and the years 1933, 1939, and 1945
Source-country origin15 pointsMediumThe stated origin, or the gathered text, mentions one of fourteen countries with active repatriation claims
Looting signal20 pointsHighLooting or illicit-trade vocabulary: looted, stolen, tomb, excavated, smuggled, repatriated, illicit, or tombaroli
Valuation outlier30 pointsHighAn asking price more than three times the stated market estimate
High value with no comparison10 pointsLowAn asking price above ten million United States dollars supplied with no market estimate to check it against
Registry match40 pointsHighThe paid commercial check returns a match against a stolen or repatriated record
The deductions used by the web pipeline. Each condition that matches produces a red flag carrying its category, its severity, and a sentence of evidence.

The three years named in the gap rule are not arbitrary: they mark the Nazi accession to power, the outbreak of the Second World War in Europe, and its end, which is the period the Washington Principles are concerned with. The gap rule carries the second-largest deduction because a gap is the single most common signature of an illegitimate acquisition. Objects with clean histories tend to have documented ones.

The fourteen source countries are Italy, Greece, Egypt, Turkey, Cambodia, China, Iraq, Peru, Mexico, Nigeria, India, Syria, Cyprus, and Thailand. The list is a shorthand for elevated restitution exposure, and its deduction is modest by design. Originating in Egypt is not a defect in an object; it is a reason to check the paperwork more carefully.

The valuation rule carries the largest deduction of the language-based rules, and it is not really a provenance rule at all. It is a money-laundering rule. Art is a recognised vehicle for moving illicit value precisely because it has no fixed price, and a wildly inflated sale is a recognised laundering technique. The threshold of three times is a judgment call rather than a figure derived from evidence.

The registry match is the largest single movement in the system, and rightly so: it is the only deduction triggered by a direct hit in a dedicated register rather than by language found in general sources. The command-line implementation adds two further categories: a repatriation precedent flag when the timeline itself cites a return or restitution event, and a transaction flag recording the result of screening the system's own payment for exposure to sanctioned addresses.

The spending variables. Because the prototype can spend money on its own initiative, the numbers that constrain that behaviour are as important as the ones that score the object. The vendor price is set by default to five United States cents, the cost of one premium search from the simulated vendor. The maximum spend per run is set by default to twenty-five cents, a hard ceiling: if the price of a check plus what has already been spent would exceed it, the check is refused and the refusal is recorded in the output. The conclusive threshold is set at 90 in the command-line agent, above which the agent will not buy a check at all, on the reasoning that the result could not change the conclusion. The web pipeline applies a comparable rule at a confidence of 85 combined with a complete absence of red flags. The design intention behind these three numbers is that an autonomous purchaser should have a price it knows, a ceiling it cannot cross, and a standard of sufficiency at which it stops buying. It is worth noting the distance between the demonstration and reality here. A real single search of the Art Loss Register has been priced in the region of sixty pounds sterling plus tax, not five cents, and at that price the economic reasoning the agent performs would carry considerably more weight.

The coverage class, and why the score is never shown alone. The confidence score is, mechanically, a measure of how much published evidence the system managed to find. That is a defensible proxy where the documentary record is dense and close to meaningless where it is not, and the density is least favourable in exactly the places this project exists to worry about. Two opposite situations produce the same low number. The first is absence within coverage: a Dutch painting sits inside auction catalogues, dealer stock books and decades of Nazi-era provenance research, so a hole in that record is itself evidence, because records would be expected to exist. The second is absence of coverage: a Cambodian temple sculpture removed during a civil war was never accessioned, catalogued or reported stolen, because no institution was in a position to report it, and it cannot appear in a stolen-property register at all. Finding nothing about it establishes nothing. The prototype therefore computes, separately from the score, which registers could in principle have named the object, and classifies it as well covered, partly covered, or structurally uncovered. The score is displayed with that classification attached and declared comparable only within it. Two decisions matter. The region used is the jurisdiction of the loss, not of manufacture and not the current location, because only the first determines which national register could hold the object; the Euphronios Krater was made in Athens, looted in Italy and held in New York, and only the Italian answer reaches the archive that recovered it. And coverage is never folded into the score, because adjusting the number by how much could be reached would produce a single figure meaning two things again, which is the defect being corrected. The effect can be read off the catalogue: the Getty Bronze scores 26 and is well covered, with an enforceable Italian confiscation order standing against it, while the Rosetta Stone scores 34 and is structurally uncovered, because no register in the set can hold a colonial-era seizure from Egypt. The two figures sit close together and mean almost opposite things. A structurally uncovered classification is an assertion about the register landscape rather than a finding about the object, and is neither exoneration nor accusation.

The demonstration mode switch. One further variable governs everything else. The system has a master setting with two positions. In mock mode, which is the default, every external connection is replaced by stored example data: no search is performed, no payment is made, and no network is required. In live mode, the system attempts real connections and quietly falls back to the stored data if any of them fail. This is candid engineering for a stage demonstration, and the reason the default is the safe one is stated in the code: so that the prototype never moves funds unless someone has explicitly asked it to. But it has an important consequence for any reader evaluating the tool. Unless the setting has been changed and real credentials supplied, an impressive-looking run is a replay of a stored example rather than live research. Both the fallback behaviour and the mode in use are disclosed on screen and in the output.

07

Reading the Results

The prototype opens on a grid of objects it already tracks, each shown as a card with a confidence figure, a repatriation status of repatriated, contested, or clear, its current holding institution, and the number of stops on its journey. Clicking a card opens that object's full dashboard.

A caution about the label on those cards. The interface prints the confidence figure with the word risk in front of it, so the Euphronios Krater appears as risk 12 out of 100. The underlying number is a cleanliness score, and 12 out of 100 means a badly compromised record, not a low level of risk. A reader who has not been told this will read the card exactly backwards. The colour coding and the accompanying risk level of high or low are correct; only the wording is misleading. It is a small defect with real potential to confuse, and it is recorded here because a plain-language review should say so.

The central display for each object is the sequence of places it has physically been, each stop carrying a year, a place, a country, a description, and a type drawn from a fixed set: origin, excavation, looting, sale, museum, repatriation, or contested. This turns a paragraph of prose into a route that can be read at a glance, and it makes the shape of a problematic history visible. An object that goes from a tomb, to a dealer in a neutral-jurisdiction city, to a major museum, within eighteen months and with no documented owner before that, has a recognisable silhouette.

Each red flag is displayed with its severity and the sentence of evidence that produced it. No flag appears without its reason. Every institution consulted is listed with the address of the page relied on, so that a researcher can go and read it. This is what distinguishes a first-pass research aid from an answer machine: the output is designed to be checked, and the checking is made easy. A search box runs the full agent on any object the user names, with the stages appearing one by one as they complete, including the payment reasoning.

The prototype ships with five real and well-documented cases, chosen to span the range of outcomes: the Euphronios Krater, looted from an Etruscan tomb near Cerveteri in 1971, bought by the Metropolitan Museum of Art in 1972 and returned to Italy in 2008; a Benin Bronze plaque taken during the British punitive expedition of 1897 and still held in London against an active Nigerian claim; the Rosetta Stone, ceded to Britain under the Capitulation of Alexandria in 1801 and the subject of repeated Egyptian requests; the Lydian Hoard, looted from burial mounds in western Turkey in the 1960s and returned by the Metropolitan Museum in 1993; and John Singer Sargent's Madame X, purchased from the artist in 1916 with an unbroken documented history. The last of these is the control case. It is the example of what a clean record looks like, and it scores 93 out of 100.

08

Data Sources and the Permitted-Source List

The single most consequential design decision in the prototype is that its research is confined to a fixed list of permitted sources. The search is restricted to the domains of:

  • The Metropolitan Museum of Art
  • UNESCO, including its World Heritage Centre
  • The Art Loss Register
  • The International Council of Museums
  • United States government cultural-heritage sites

The purpose of the restriction is to make the tier system meaningful. If the system may search anywhere, then a claim in a collector's forum post and a claim in a museum's own accession record arrive in the same shape and the distinction between them has to be reconstructed after the fact. By searching only recognised institutions, the prototype makes the authority of a claim a property of where it was found.

The named institutions are worth introducing for a reader unfamiliar with the field. UNESCO is the United Nations Educational, Scientific and Cultural Organization, and is the custodian of the principal international instrument on trafficking in cultural property. The International Council of Museums is the global professional body for museums, and publishes the Red Lists that identify categories of object at particular risk of illicit trafficking from specific regions. The Art Loss Register is a private London company that operates the largest commercial database of stolen art and conducts due-diligence checks for auction houses, dealers, insurers, and law enforcement, reportedly running several hundred thousand checks a year. The Metropolitan Museum of Art is included both as a collection record and, unavoidably, as a party to several of the best-documented restitution cases in the field.

The cost of the restriction is that the list is short, and its omissions are substantial. It does not include the Getty Provenance Index, the German Lost Art Foundation's database, Interpol's stolen works of art database, the Art Institute of Chicago, Europeana, or the national heritage authorities of any of the fourteen countries the system itself identifies as source countries. An object well documented in a French or Turkish archive and absent from these six domains will return little, and the tool will report a thin record rather than an absent one. The two are not the same thing, and the prototype does not currently distinguish them.

That omission is not evenly distributed, and the direction of the bias runs against the tool's own motivation. The case for building this system, set out in sections 2 and 11, rests on the objects that stolen-art registers cannot catch: material taken from an archaeological site or under colonial rule, never inventoried and never reported stolen. But the permitted list is five Western institutions and one commercial theft register. The tool therefore searches best where objects are already well documented, which is to say major Western museum holdings, and searches worst exactly where the motivating harm lives, which is source-country archives and colonial-era and archaeological material. The observation in section 13 that the system does not distinguish thin evidence from absent evidence is the symptom; this is the cause, and the two belong together. The consequence is that a low confidence score for a Cambodian sculpture and a low score for a Dutch painting do not mean the same thing, and the tool does not currently say so.

Adding source-country heritage authorities, the Getty Provenance Index, and the German Lost Art Foundation is therefore the first substantive extension of this work rather than an optional one. Every other improvement to the scoring, the interface, or the passport format operates on evidence the tool was able to find, and the coverage bias determines what it can find at all.

The register layer, and why it is not a set of lookups. That extension has since been partly built. The permitted list now also covers Interpol, the Federal Bureau of Investigation, the Carabinieri Command for the Protection of Cultural Heritage in Italy, the German Lost Art Foundation, and the Getty Research Institute, and the prototype puts each object to nine named registers rather than to the allowlist as an undifferentiated whole. The extension is smaller than it sounds, for a reason that governs everything the layer can honestly claim: the registers that actually certify stolen cultural property have no public programmatic interface. This was tested rather than assumed.

What each register actually permits, established by trying rather than by reading the marketing:

  • Interpol's Stolen Works of Art database, holding on the order of fifty-two thousand records of certified police information, is reachable through the ID-Art mobile application or through an account applied for and vetted by the applicant's national Interpol bureau. Neither route can be automated.
  • The FBI's National Stolen Art File is published at a web address that refuses any client which is not a browser. The Bureau's one open programmatic interface covers wanted persons and contains no art records at all; a filter requesting art crime returns the fugitive list without indicating that it has ignored the request.
  • The Carabinieri archive, at over a million objects the largest in the world, and the German Lost Art database are public search forms in HTML with no documented data interface.
  • The Art Loss Register is a commercial service, and in this prototype is the paid check described in section 9 rather than a free lookup.
  • Wikidata's structured-data endpoint is the single exception, and is genuinely queried. Its event statements carry dated values such as archaeological looting, art theft, restitution, and claim for restitution.

The layer therefore does not pretend to search those registers. Where a source is genuinely open to machine query it queries it. Where a register publishes material openly but cannot be queried, the tool searches that register's website, which is a different thing from searching its register and is labelled as such. Where neither is possible, it emits a link to the official search a person must run by hand, together with the address at which credentialed access may be applied for. Every result is presented with the access route that produced it, because the route is what makes the result interpretable: the words no evidence found mean one thing after a genuine query of a structured source and something much weaker after a keyword search of an institution's public pages.

The strongest negative the system can express is no evidence found. There is deliberately no verdict meaning clear or not stolen, and the type that represents a register result does not contain one, so no future change can produce that conclusion by accident. A stolen-property register can only contain objects that somebody was in a position to report missing. Material removed from an archaeological site, or taken under colonial administration, was never inventoried and never reported, and therefore cannot appear in such a register at all. For precisely the objects this project was built to worry about, absence from the registers is the expected condition rather than a reassuring one, and a tool that rendered it as a clean bill of health would be worse than no tool.

Scoring follows the same asymmetry. A register hit reduces confidence, a register that returns nothing adds none, and registers that could not be searched at all are counted and reported as a coverage gap, so that a check which reached two sources cannot be mistaken for one that reached nine. Each check, including those that failed and those that never ran, is written into the signed Passport together with the sentence describing what its verdict does and does not license, so that the qualification cannot be separated from the finding by anyone who handles the credential later.

The watchlist. The prototype also publishes a list of several hundred works recorded as stolen or plundered, drawn from Wikidata's theft statements and regenerated by script, so that the tool is not limited to the fifteen objects researched by hand. It is not an extract from the Interpol or FBI databases, and the interface says so at the point of display rather than in a footnote. It is community-maintained data, unreviewed and uneven in coverage. An entry is a lead to be checked against the official registers rather than a register hit, and the dates require particular care, because Wikidata sometimes attaches the restitution date to the theft statement, so a recent year on a Nazi-plunder record frequently marks a return rather than a taking. Absence from the list carries no information at all. Two decisions in its construction are worth recording, because both were errors first. The query is constrained to works of art, because without that constraint it also returns people and companies, since spoliation records attach theft events to the dispossessed as well as to their property, and a list presented as artworks that silently contained the names of victims would be both wrong and offensive. And the query had to be widened to reach events that are instances of art theft rather than subclasses of it, because named incidents such as the Isabella Stewart Gardner Museum theft are recorded as instances, and the original formulation silently omitted every object taken in a named robbery, including the most famous stolen painting in the world.

What remains missing is what mattered most at the start of this section and still does. The national heritage authorities of the fourteen source countries the system itself recognises are still absent, and no amount of register plumbing substitutes for them. The layer is built so that this shows in its output rather than being smoothed over, which is the most that can be claimed for it.

09

The Payment Layer, and Why It Is There

A reader may reasonably ask what an automated payment system is doing in a cultural-heritage research tool. The answer is partly historical: the prototype was built for a hackathon on autonomous software commerce, and the payment capability was a condition of entry. But the underlying argument is a real one and worth stating on its merits.

The most authoritative sources on stolen art are commercial and closed. The Art Loss Register does not publish its database; access is sold per search. That model works when the requester is an auction house with a subscription and an accounts department. It works badly when the requester is a researcher checking one object, a small museum, or a piece of software that has just discovered mid-task that it needs one specific answer. The friction is administrative rather than financial: the sum involved may be trivial, but obtaining an account, a key, and an invoice is not.

The x402 standard is an attempt to remove that friction. It uses the HTTP status code 402, Payment Required, reserved decades ago for exactly this purpose and left unused until recently. A server responds to a request by stating a price; the requester attaches a payment and asks again; the data is returned. There is no account and no subscription, and the exchange completes in seconds. The standard was published and open-sourced by Coinbase in 2025 and has since passed to neutral governance.

What that enables is described in the repository as paying for tools you discover: software that encounters a paywalled resource it did not know about in advance can evaluate the price and buy the answer, rather than failing and waiting for a human to arrange access. Applied to provenance, this is the difference between a research assistant that stops at the edge of the free sources and one that can obtain the commercial check when the free sources prove inconclusive.

The prototype implements this end to end, but against its own simulated vendor. The repository is explicit on the point: the paywall behaviour is genuine, in that the vendor really does refuse the request until payment settles, but the data behind it is stored example content, not the Art Loss Register's actual database. Payments are made in a test version of the USDC digital dollar on Base Sepolia, a practice network using valueless test tokens, and the repository warns in terms that the wallet must never be funded with real money. No commercial relationship with the Art Loss Register is claimed or implied.

The system also screens its own payment. Having paid, it examines the resulting transaction for exposure to sanctioned addresses or laundering patterns and records the verdict in the final document. In the demonstration this screening is illustrative rather than substantive, and the code says so. The reason it is present at all is that a tool used to detect value laundering through art should be able to demonstrate that its own transactions are clean.

10

The Passport and What the Seal Proves

The output document is issued in a standard format known as a Verifiable Credential, a specification maintained by the World Wide Web Consortium for digital records that can be checked by anyone without contacting the issuer. The format matters because a provenance assessment that only one institution's software can read is of limited use to the buyer, insurer, or regulator who receives it.

The record contains the object's identity, the full ownership timeline with a source address against every event, the confidence score, every red flag with its evidence, any paid checks with the amount paid and the transaction they produced, the list of checks performed, and the time of assessment. It is assembled from what the run actually collected. The code is explicit that the record is built from accumulated state rather than re-derived from a language model, so that the sealed document is always a faithful account of the run.

The sealing works as follows. The record is first written out in a strictly canonical form, with every field in a fixed order, so that the same content always produces an identical text. A cryptographic signature is then computed over that text. Anyone holding the document can repeat the calculation and recover the identity of the signer. If any character of the record has been changed since signing, the recovered identity will not match and the document is exposed as altered.

The prototype includes a demonstration of exactly this. Its example script signs a Passport, then alters a single field, raising the confidence score to 100, and re-checks it. The check fails and reports that the content no longer matches its seal. This is a small thing to build and a significant thing to show, because the value at stake in a provenance document is high enough to make quiet editing a genuine risk.

One conceptual point in the design deserves comment. The identity that signs the Passport is the same digital wallet identity that pays for the premium check. The repository summarises this as the observation that a wallet is already a form of public-key infrastructure, applied here to an object's identity rather than to a payment. The practical effect is that the entity that spent the money and the entity that vouched for the record are provably the same, and no separate system of certificates or key distribution is required.

What the seal proves, and does not prove, should be stated exactly. It proves that this document has not been altered since it was signed, and that it was signed by the holder of a particular key. It proves nothing whatever about whether the contents are true. A sealed record of a bad assessment is a bad assessment that cannot be quietly improved later. That is a genuine benefit, and it is a narrower one than the word verified tends to suggest to a general reader.

11

The Legal and Normative Background

The prototype's warning rules are not invented from nothing. They correspond, loosely but recognisably, to the concerns of the principal instruments in this field.

The 1970 UNESCO Convention. The Convention on the Means of Prohibiting and Preventing the Illicit Import, Export and Transfer of Ownership of Cultural Property was adopted by UNESCO in 1970 and entered into force in 1972. It has been ratified by 149 states. It requires states parties to establish national services for heritage protection and to legislate accordingly, to operate a system of export certificates, to prohibit the import of objects stolen from museums, religious institutions, and public monuments, and, under Article 7, to seize and return such objects when another state party requests it through diplomatic channels. Article 9 commits states to cooperate in concerted international action against trafficking. Its practical significance for provenance work is the date itself. The year 1970 functions as a dividing line in museum acquisition policy across much of the world: an antiquity that can be shown to have left its country of origin before 1970 sits in a different position from one that cannot. This is why the system's list of source countries, and its attention to undocumented periods, map onto a real and consequential legal boundary.

The Washington Principles on Nazi-Confiscated Art. Agreed on 3 December 1998 by more than forty states at a conference convened in Washington, these eleven non-binding principles address art confiscated by the Nazi regime and never returned. They call for such art to be identified, for archives and records to be opened to researchers, for resources to be committed to the work, for the results to be publicised so that pre-war owners and their heirs can be found, for a central registry of the information, and for claims to be resolved through a just and fair solution rather than left to the technicalities of ordinary property litigation. The fourth principle is the one most directly reflected in this software. It provides that in establishing whether a work was confiscated by the Nazis and not restituted, consideration should be given to unavoidable gaps or ambiguities in the provenance, in light of the passage of time and the circumstances of the Holocaust era. In ordinary terms: an incomplete record is not to be held against a claimant, because the destruction of records was part of what happened. This is precisely why the prototype treats a documentation gap as a signal deserving investigation rather than as an absence of evidence, and why the years 1933, 1939, and 1945 appear by name in its rules. The principles are declaratory and create no legal obligations. Their influence comes from adoption into national practice and institutional policy, and assessments of their record over the past quarter-century have been decidedly mixed.

The commercial registers. The Art Loss Register is the market's standard first-line check. Its limits are structural and well understood by practitioners, and they explain why a tool of this kind is worth building. The register records objects that were reported stolen. An object looted from an undocumented archaeological site was never inventoried and therefore was never reported, so it cannot appear. The same is true of most objects taken under colonial administration. A clean result from a stolen-art database is a meaningful answer to a narrow question and is routinely mistaken for a broader assurance than it can provide.

What the prototype does not do. It does not determine title, and it does not apply the law of any jurisdiction. Limitation periods, good-faith purchaser rules, and the availability of a remedy vary enormously between legal systems, and none of that is modelled. The code states the principle directly: the system flags signals and cites evidence, and never asserts a legal conclusion. That restraint is the correct posture for a research aid, and it should be preserved in any future development.

12

Methodological Choices

Why the scoring is arithmetic rather than learned. The prototype could have asked a language model to assign a provenance score directly. It does not. Every deduction is a written rule with a fixed value, applied to evidence that has already been gathered and cited. The reason is auditability. A number produced by a rule can be disputed by disputing the rule. A number produced by a model cannot be disputed at all, only accepted or rejected, and in a field where the output may inform a restitution claim that is not an acceptable property.

Why sourcing is enforced structurally rather than by instruction. It is possible to instruct a language model to cite its sources, and it is well-established that the instruction is sometimes ignored, including by inventing sources that do not exist. The prototype instead makes the citation a structural requirement of the data itself: a timeline event without a source address is discarded before it can be recorded, and the system counts the discards. This is a stronger guarantee than any instruction, and it is the single most defensible piece of engineering in the repository.

Why the search is restricted. The restriction is what allows reliability to be inferred from the provenance of the claim rather than assessed after the fact.

Why price is treated as provenance evidence. Including a valuation check in a provenance tool is an unusual choice, and a good one. Art is a recognised channel for moving illicit value, and a sale far above market is a recognised technique. The prototype treats an inexplicable price as a fact about the transaction in the same way it treats a gap as a fact about the record. The threshold of three times market value rests on judgment rather than on any published study.

Why failures fall back rather than stop. Every external connection in the system degrades to stored example data on error rather than halting. This was chosen so that a live demonstration cannot fail in front of an audience. It is a reasonable choice for its purpose and an unacceptable one for research use, because a tool that silently substitutes example data for real research will eventually be trusted when it should not be. The substitution is disclosed on screen, but disclosure is a weaker safeguard than refusal.

Why register access is tiered rather than flattened. The register layer could have reported a single verdict per register and kept the retrieval method internal. That would have produced a cleaner interface and a dishonest one. The nine registers are reached by four quite different routes of very unequal strength, and a verdict is only interpretable alongside the route that produced it. Presenting a keyword search of an institution's public pages in the same shape as a genuine query of a structured database would invite exactly the inference the system exists to prevent, which is that a quiet result is a reassuring one. The interface is therefore more cluttered than it would otherwise be, and the clutter is carrying the argument.

Why there is no verdict meaning clear. This is the same commitment as the structural enforcement of sourcing, applied to conclusions rather than to citations, and it is enforced the same way, in the type system rather than in a guideline. A register result can say that a match surfaced, that nothing surfaced through the access available, or that the register could not be reached. There is no fourth value, so no later change can produce a clean bill of health by inadvertence. The design principle behind both is that a rule which matters should be made impossible to violate rather than merely documented, since documentation is not read at the moment the mistake is made.

Why silence earns nothing. The scoring treats a register that returned nothing as contributing no confidence, and reports separately how many registers could not be searched at all. This is the same reasoning that made the accumulation model canonical over the deduction model: absence of evidence is not evidence of clean provenance. A system that rewarded quiet registers would score an object highest precisely when it had learned least about it, and the objects it learns least about are disproportionately the ones taken from places that never had an inventory to be missing from.

13

Limitations and Caveats

The repository and the paper both state these limits plainly. They are reproduced here because they are the most important part of the document for any reader considering what weight to give the tool.

It cannot read an archive. The overwhelming majority of provenance evidence is in physical archives, in dealer records, in correspondence, in auction catalogues that were never digitised, and on the back of the object itself. None of this is reachable by a web search. The prototype gathers what institutions have chosen to publish online, which for most objects is a small and unrepresentative fraction of what exists.

The keyword rules are crude. The system decides that a provenance gap exists by looking for particular words in the gathered text. A museum page stating that an object has no gaps in its record contains the word gap and will trigger the flag. A page discussing the successful repatriation of a different object will trigger the looting flag. These rules find the right signal often enough to be useful as a prompt for human attention, and they will produce false positives regularly. They should be read as a list of things to check, never as findings.

The source-country list is hand-written and incomplete. Fourteen countries are named. Many states with active restitution claims are not among them, and membership of the list is not a legal category. It is a useful shorthand and nothing more.

The thresholds are not empirically derived. Twenty-five points for a gap, fifteen for source-country origin, three times market value for a pricing anomaly, ninety as the point at which further checking is pointless: none of these figures comes from a study or an expert panel. They are the developer's judgment, plausibly calibrated and entirely unvalidated. This is a normal starting position for a prototype and it should not be mistaken for anything more.

There are two different scoring systems in the repository. The web interface and the command-line agent compute confidence in materially different ways and will not agree on the same object. Anyone building on this work should reconcile them before doing anything else.

The interface mislabels its own headline number. A card reading risk 12 out of 100 describes a severely compromised object, which is the opposite of what the phrasing suggests.

Most of what is impressive is simulated. In the default configuration there is no search, no payment, and no database query. The commercial stolen-art check is a stand-in with stored answers. The transaction screening returns a fixed clean verdict. The catalogue of five objects is written by hand from published cases rather than produced by the system. A reader evaluating a demonstration should establish which mode it is running in before drawing conclusions about capability.

The risk of plausible falsehood remains. This is the most serious caveat in the report. The prototype's sourcing rule is a genuine and unusually well-implemented safeguard, and it does not eliminate the problem. It guarantees that every recorded claim has a source address attached. It does not guarantee that the source says what the record claims it says, that the source is about the same object, that the extracted sentence has not lost the qualification that made it accurate, or that a page found on an authoritative domain is itself authoritative. A false provenance claim, expressed in institutional language, carrying a real link to a real museum, and sealed inside a document described as verifiable, is more dangerous than an obvious error, because every visible signal invites trust. This is a foreseeable failure mode of the design, not a hypothetical one. Every claim in a Passport must be read at its source before it is relied on, and the fact that the document is sealed and formally verifiable does nothing to change that.

It does not distinguish thin evidence from absent evidence. An object with no online record and an object with a clean record but no online presence look identical to this system. For the majority of the world's cultural objects, which are not held by major Western museums with published collection databases, this limitation is severe.

It is a prototype. It has no test suite, no validation against expert assessments, and a commit history spanning a few days of work. It should be read as a demonstration of a method and an argument about how such a tool ought to be built, not as a service.

14

Relationship to the Sibling Project

The Ethical Tech CoLab maintains a related repository, provenance-search, which addresses the same problem from a different direction. That project is a web application that identifies an artwork, including from a photograph taken on a phone in a museum, and searches a considerably wider set of free public sources: the Metropolitan Museum, the Art Institute of Chicago, a bundled extract of the Museum of Modern Art's open collection data, Wikipedia, Wikidata, and Europeana, alongside the same commercial search service used here.

The two projects share a central conviction, stated in both repositories, that the confidence score must be computed by a fixed algorithm rather than by the artificial intelligence, and both restrict their research to a list of recognised sources.

They differ in emphasis. The sibling project prioritises breadth of free sources and accessibility, including identification by camera. The project described in this report prioritises the integrity of the resulting record: the enforced sourcing rule, the evidence tiers, the tamper-evident seal, and the ability to obtain paid data autonomously. Read together they suggest a natural combination, in which the wider source coverage of the one feeds the evidentiary discipline and sealed output of the other. This report treats the present repository on its own terms; the comparison is offered only because a reader encountering both should understand that they are complementary experiments rather than duplicates.

15

Intended Audience and Use

The prototype is aimed at those who must form a view about an object's history before deciding what to do with it: museum registrars and provenance researchers, restitution and cultural-property lawyers, insurers and compliance staff at auction houses, customs and heritage-crime investigators, and the source-country authorities pursuing claims.

Its appropriate use is as a first pass. It is well suited to triage, meaning the task of deciding which objects in a large collection or a large consignment merit a researcher's time. It is well suited to preparing a starting dossier with sources already assembled and linked. It is well suited to teaching, since it makes the structure of provenance reasoning visible.

It is not suited to supporting a claim, defending an acquisition, satisfying a due-diligence obligation, or reaching any conclusion that will be acted upon without a qualified human having read the underlying sources.

16

Conclusion

The Digital Provenance Passport takes a body of work that is slow, specialised, and applied to only a small fraction of the objects that need it, and asks what part of the first pass could be done automatically without becoming untrustworthy. Its answer is more careful than most work in this area. It refuses to record any claim that does not carry its source. It searches only recognised institutions. It computes its judgments with arithmetic that a non-programmer can read and dispute. It seals its output so that the assessment cannot be quietly improved after the fact. It records what it bought, what it paid, and why it thought the purchase worthwhile.

The prototype is also, in its present state, mostly a demonstration. Its commercial database check is a stand-in, its default mode replays stored examples, its permitted-source list omits most of the field's principal resources, its keyword rules will misfire, its two scoring systems disagree with each other, and its interface labels its central number in a way that invites the opposite reading. None of this is concealed in the repository, and none of it should be concealed in a report about it.

What is durable here is the argument rather than the software. It is the proposition that when automated tools are pointed at questions where the stakes include the return of stolen property and the resolution of historical injustice, the design should make fabrication structurally impossible rather than merely discouraged, should make every judgment traceable to a rule and every rule open to challenge, and should produce a record that cannot be edited without the edit becoming visible. Those are the right commitments. This prototype implements them incompletely, states its own limits without flattering itself, and leaves a workable foundation for someone to build on.

References

Sources

  1. 01UNESCO. Convention on the Means of Prohibiting and Preventing the Illicit Import, Export and Transfer of Ownership of Cultural Property, 1970.
  2. 02United States Department of State. Washington Conference Principles on Nazi-Confiscated Art, 3 December 1998.
  3. 03The Art Loss Register. Commercial database of stolen art and due-diligence search service.
  4. 04International Council of Museums. Red Lists of cultural objects at risk.
  5. 05Getty Research Institute. Getty Provenance Index, holding more than two million records.
  6. 06The Metropolitan Museum of Art. Collection records, cited as a permitted source and as a party to several documented restitution cases.
  7. 07World Wide Web Consortium. Verifiable Credentials Data Model.
  8. 08Coinbase. x402, an open payment standard using HTTP status code 402, published 2025.
  9. 09Tavily. Web search service returning extracted text with source addresses, used for the grounding stage.
  10. 10Ethical Tech CoLab. provenance-search, the sibling repository described in Section 14.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file diff --git a/static-site/print/digital-provenance-passport/index.txt b/static-site/print/digital-provenance-passport/index.txt index 9d510b626..69d6abade 100644 --- a/static-site/print/digital-provenance-passport/index.txt +++ b/static-site/print/digital-provenance-passport/index.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -12:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","print","digital-provenance-passport",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","digital-provenance-passport","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -14:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -16:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","print","digital-provenance-passport",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","digital-provenance-passport","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:"$Sreact.suspense" -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1c:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -29,7 +29,7 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 19:[] 10:"$W19" 11:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1e:Ta14, @page { size: 210mm 297mm; margin: 20mm 18mm; } @@ -163,6 +163,6 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 52:["$","tr","5",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Registry match"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"40 points"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"High"}],["$","td","3",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"The paid commercial check returns a match against a stolen or repatriated record"}]]}] 53:["$","figcaption",null,{"className":"mt-3 text-sm text-muted","children":"The deductions used by the web pipeline. Each condition that matches produces a red flag carrying its category, its severity, and a sentence of evidence."}] 1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -57:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +57:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 18:null 1d:[["$","title","0",{"children":"The Digital Provenance Passport (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L57","6",{}]] diff --git a/static-site/print/diplomatic-simulator/__next._full.txt b/static-site/print/diplomatic-simulator/__next._full.txt index 37b738f4a..4b97e599b 100644 --- a/static-site/print/diplomatic-simulator/__next._full.txt +++ b/static-site/print/diplomatic-simulator/__next._full.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -12:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","print","diplomatic-simulator",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","diplomatic-simulator","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -14:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -16:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","print","diplomatic-simulator",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","diplomatic-simulator","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:"$Sreact.suspense" -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1c:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -29,7 +29,7 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 19:[] 10:"$W19" 11:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1e:Ta14, @page { size: 210mm 297mm; margin: 20mm 18mm; } @@ -165,6 +165,6 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 53:["$","p","13",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"The named figures are a characterisation."}]," ","Attaching the real head of government and the foreign and defence ministers to each delegation sharpens its voice but also imports the model's biases about specific, living individuals. These entries are drawn from public reporting and describe how each figure is publicly discussed; they are not claims about private conduct, and the simulator does not model any named person's actual decisions."]}] 54:["$","p","14",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"It is not decision support."}]," ","The repository states directly that the tool must not be used to inform real policy, negotiation, or intelligence judgements."]}] 1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -57:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +57:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 18:null 1d:[["$","title","0",{"children":"The Diplomatic Simulator (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L57","6",{}]] diff --git a/static-site/print/diplomatic-simulator/__next._head.txt b/static-site/print/diplomatic-simulator/__next._head.txt index c25f39ef7..ec1fba23a 100644 --- a/static-site/print/diplomatic-simulator/__next._head.txt +++ b/static-site/print/diplomatic-simulator/__next._head.txt @@ -1,6 +1,6 @@ 1:"$Sreact.fragment" -2:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -3:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +2:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +3:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 4:"$Sreact.suspense" -5:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +5:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 0:{"rsc":["$","$1","h",{"children":[null,["$","$L2",null,{"children":[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]]}],["$","div",null,{"hidden":true,"children":["$","$L3",null,{"children":["$","$4",null,{"name":"Next.Metadata","children":[["$","title","0",{"children":"The Diplomatic Simulator (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L5","6",{}]]}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/print/diplomatic-simulator/__next._index.txt b/static-site/print/diplomatic-simulator/__next._index.txt index f98a4fd90..18b9d82b6 100644 --- a/static-site/print/diplomatic-simulator/__next._index.txt +++ b/static-site/print/diplomatic-simulator/__next._index.txt @@ -1,13 +1,13 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -d:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +d:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} 8:["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$Ld",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$Ld",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$Ld",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$Ld",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] diff --git a/static-site/print/diplomatic-simulator/__next._tree.txt b/static-site/print/diplomatic-simulator/__next._tree.txt index d35da11ca..55dca4396 100644 --- a/static-site/print/diplomatic-simulator/__next._tree.txt +++ b/static-site/print/diplomatic-simulator/__next._tree.txt @@ -1,7 +1,7 @@ -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] 0:{"tree":{"name":"","param":null,"prefetchHints":16,"slots":{"children":{"name":"print","param":null,"prefetchHints":0,"slots":{"children":{"name":"slug","param":{"type":"d","key":"diplomatic-simulator","siblings":null},"prefetchHints":0,"slots":{"children":{"name":"__PAGE__","param":null,"prefetchHints":0,"slots":null}}}}}}},"staleTime":300,"buildId":"static-snapshot"} diff --git a/static-site/print/diplomatic-simulator/__next.print.txt b/static-site/print/diplomatic-simulator/__next.print.txt index 460afe242..fa617e997 100644 --- a/static-site/print/diplomatic-simulator/__next.print.txt +++ b/static-site/print/diplomatic-simulator/__next.print.txt @@ -1,4 +1,4 @@ 1:"$Sreact.fragment" -2:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -3:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +2:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +3:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] 0:{"rsc":["$","$1","c",{"children":[null,[["$","style",null,{"children":"\n /* The site is dark-first, and a dark color-scheme makes Chrome paint\n the printed sheet's margin area dark too. Paper is light. */\n html { color-scheme: light !important; }\n html, body { background: #ffffff !important; }\n body::after { content: none !important; }\n header, footer, .site-bg, .aura { display: none !important; }\n "}],["$","$L2",null,{"parallelRouterKey":"children","template":["$","$L3",null,{}]}]]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/print/diplomatic-simulator/index.html b/static-site/print/diplomatic-simulator/index.html index bc214ddc0..38d0fe473 100644 --- a/static-site/print/diplomatic-simulator/index.html +++ b/static-site/print/diplomatic-simulator/index.html @@ -1,4 +1,4 @@ -The Diplomatic Simulator (print edition) · NYU Ethical Tech CoLab

Publications · Academic report

The Diplomatic Simulator

A Multi-Party Negotiation Simulator Driven by Artificial Intelligence Agents

Ethical Tech CoLab

NYU Center for Global Affairs

July 2026

Carolina Morón. Prepared as masters research at the NYU Center for Global Affairs. The original diplomacy table demonstration and the Strait of Hormuz session report, from which the presentation layer is adapted, were built by Yorke Rhodes III.

Multi-party negotiation is difficult to teach and almost impossible to rehearse. The Diplomatic Simulator asks what happens when the delegations at a crisis table are played by artificial intelligence agents instead of by people. Each delegation is given a confidential brief and nothing else, the talks are run in rounds, and every word spoken is preserved, tagged, and scored. It is a teaching and exploration tool, not a forecasting system, and there is no ground truth anywhere in it.

Key figures

133

statements of record produced by forty-two delegation agents across six multi-party scenarios

17

negotiating tactics in the fixed vocabulary delegations use to label their own moves

40

Monte Carlo trials run for each scenario under randomly varied shocks, pressure, and mood

0

comprehensive settlements in any trial of any scenario, a ceiling the scenario packs impose rather than a finding about diplomacy

01

Executive Summary

The Diplomatic Simulator replays multi-party diplomatic negotiations in which every national delegation is played by an artificial intelligence agent, that is, a language program instructed to speak and reason in the voice of one government. It is a teaching and exploration tool. It is not a forecasting system, and its authors state plainly that it must not be used to inform real policy, negotiation, or intelligence judgements.

Six scenarios have been simulated to date: the Arctic, with seven delegations; Central Asia and the Fergana Valley, with seven; Cyprus reunification, with six; the South China Sea, with nine; the Korean Peninsula Six-Party Talks, with six; and Jammu and Kashmir, with seven. An earlier two-party session on the Strait of Hormuz, involving Iran and the United States, is preserved as the original demonstration. Across the six multi-party scenarios, forty-two delegation agents produced one hundred and thirty-three statements of record.

The organising principle of the design is information isolation. When a delegation agent writes its statement for a round, it receives only its own confidential brief, a neutral public brief that everyone shares, and the public transcript of what has already been said aloud. It never sees another government's private instructions, and it never sees another delegation's private reasoning. This mirrors the informational structure of real diplomacy, where the public record is common property and the mandate is not.

The scenario material is not invented. The confidential briefs are adapted from crisis-negotiation exercise packs associated with the United States Army War College International Strategic Crisis Negotiation Exercise programme, which are taken to universities. The file names preserved in the repository record where each pack was used, including New York University, Penn, Syracuse, and the University of Minnesota.

After the talks conclude, a separate analyst agent reads the whole transcript and produces a scoreboard, a debrief for each delegation, and a neutral summary written in the voice of a convening envoy. A further layer re-runs each scenario forty times under randomly varied external conditions, in order to show which outcomes are stable and which depend on luck.

The repository is candid that every number on the site is one artificial intelligence system's subjective judgement of another artificial intelligence system's writing. There is no ground truth anywhere in the tool. The most serious structural limitation is that a single underlying model plays every side of every table, so the adversaries are not genuinely independent minds. That limitation has now been tested once, on the Korean scenario, by re-running it with different models behind four of the six seats. The coarse outcomes replicated. The test also established something this report had not seen: on that scenario a comprehensive settlement is impossible by construction, because the concessions the parties are authorised to make do not overlap, so the zero-settlement result measures the scenario pack rather than the negotiation. The Limitations section sets out what the comparison does and does not license.

02

Background and Rationale

The problem. Crisis-negotiation exercises are among the most effective ways to teach diplomacy, because they force a participant to argue a position they may not hold, under time pressure, against people who have been told to resist them. But an exercise consumes an enormous amount of scarce human attention. A seven-party Arctic exercise needs seven teams, seven mentors, and two days. It can be run perhaps once a year at any given institution.

The second problem is that exercises leave almost no durable record. What a delegation said in round two, why it said it, and how that changed the shape of the room afterwards is rarely written down in a form anyone can study later. The learning is real but it is locked inside the experience of having been there.

The gap. Existing computational work on negotiation has largely concentrated on games with fixed rules. The best known example is CICERO, the system developed by Meta's Fundamental AI Research team and reported in Science in 2022, which reached human-level play in the board game Diplomacy by combining a language model with strategic planning. That is a considerable achievement, but a board game has a scoring function and a win condition. Real multi-party talks have neither. A separate strand of work has examined what happens when language models are placed in simulated wargames: Rivera and colleagues, publishing at the ACM Conference on Fairness, Accountability, and Transparency in 2024, found that all five models they tested displayed escalation patterns that were difficult to predict. That finding is a caution rather than an endorsement, and this prototype should be read against it.

The response. The Diplomatic Simulator occupies the space between the two. It takes the scenario packs and confidential briefs that human exercises already use, replaces the human delegations with agents, and keeps the entire record. What the tool produces is not a prediction of what governments would do. It is a legible artefact: a transcript, a set of tagged tactics, and a scoreboard, all of which can be read, disputed, and compared against what a human exercise produced from the same brief.

The design commitment that makes this worth doing is transparency about provenance. Every intermediate product of every run is committed to the repository in a readable form: the extracted party profiles, the public brief, the transcript, the analysis. A reader who doubts a score can open the transcript and check what was actually said.

03

Objectives

The prototype is designed to:

  1. Reproduce the structure of a multi-party crisis negotiation, including confidential mandates, public plenary statements, coalition formation, and a convening mediator.
  2. Preserve information isolation, so that no delegation gains an advantage the corresponding human team would not have had.
  3. Produce a complete and inspectable record of each session, including what each delegation said, which negotiating tactics it used, and how an independent analyst judged the result.
  4. Express the outcome in variables simple enough for a non-specialist to read, while keeping the reasoning behind each variable visible.
  5. Test how sensitive each outcome is to conditions no negotiator controls, by re-running each scenario under randomly varied external shocks and moods.
  6. Document its own method and its own failures, including the specific point at which an orchestration run broke and had to be repeated.

04

How the Simulator Works

The tool runs each scenario through six stages. The first five use the same reusable toolchain, held in the repository under the folder named sim; only the source documents change from scenario to scenario. Three of the six stages are ordinary computer code with no artificial intelligence involved at all.

Stage one, reading the source documents. Each scenario arrives as a set of documents: one public scenario brief describing the crisis, plus one confidential instruction file for each country. These are converted from page images into plain text. No judgement is exercised at this stage.

Stage two, building a party profile. One agent is assigned to each country. It reads that country's confidential instructions and nothing else, and distils them into a structured profile. This is the single most important step in the whole pipeline, because the profile is the entirety of what that delegation will know about itself for the rest of the negotiation.

Stage three, writing the public brief. A separate agent reads the shared scenario document and writes the neutral briefing that every delegation will see. It states the situation, lists the issues formally on the table, sets out the procedure, and fixes the vocabulary of negotiating tactics that delegations will use to label their own moves.

Stage four, the negotiation. The delegations negotiate in plenary rounds. In each round, every delegation writes exactly one statement. It is given its own profile, the public brief, and the running public transcript. It is instructed to stay in role, to ground its claims in its brief, to protect its secret bottom lines, and to label the tactics it has just used. The flagship Arctic scenario runs four rounds: an opening plenary, a positioning round, a bargaining and coalitions round, and a closing plenary. The other three scenarios run three rounds, dropping the separate positioning stage. Seven delegations across four rounds produce twenty-eight statements in the Arctic; nine delegations across three rounds produce twenty-seven in the South China Sea.

Stage five, the analysis. A single analyst agent, cast as a neutral control group, then reads the entire transcript together with every delegation's confidential profile. It is the only agent in the system that sees both sides of the information barrier. It produces the scoreboard, the per-delegation debriefs, and the convener report.

Stage six, assembly and publication. Plain computer code stitches the outputs together, standardises the tactic labels, and renders the interactive pages. No model is involved. This matters for auditing: the numbers a reader sees on the published pages are arithmetic performed on agent outputs, not a second layer of interpretation.

05

The Variables, and Why Each One Exists

Everything the simulator produces rests on a small number of variables, each of which has a plain meaning. When an agent reads a country's confidential instructions, it fills in eleven fields. Together they constitute the delegation's entire identity.

The eleven fields are as follows.

  • Delegation role. One sentence describing who this delegation formally is and how much authority it holds. The Russian profile in the Arctic scenario records a foreign ministry delegation mandated to defend Arctic sovereignty while deferring major deviations to the Foreign Minister. This field exists because a negotiator's freedom to concede is itself a variable, and a delegation that must telephone home behaves differently from one that can sign.
  • Fundamental principles. The standing commitments the government would assert regardless of this particular crisis. These give the agent something to argue from when the transcript moves somewhere its specific instructions did not anticipate.
  • Desired end state, primary. The delegation's definition of victory.
  • Desired end state, alternate. Its definition of an acceptable substitute. Having two rather than one is deliberate: a negotiator with only a maximum position cannot trade.
  • Key positions by issue. The government's stated line on each numbered issue in the public brief, stored issue by issue. This is what keeps a delegation consistent across rounds and prevents it from quietly abandoning a position it took an hour earlier.
  • Red lines. The commitments the delegation states it will not cross. In negotiation practice a red line is a declared non-negotiable limit whose value depends entirely on whether other parties believe it. The simulator makes this measurable, because the analyst later records how many of a delegation's red lines were crossed by others.
  • BATNA. A term of art from negotiation theory, introduced by Roger Fisher and William Ury in Getting to Yes in 1981. It stands for Best Alternative to a Negotiated Agreement, and it means the outcome a party falls back on if the talks collapse. It is the benchmark against which every proposal should be judged: a rational party accepts nothing worse than its BATNA. Russia's Arctic BATNA is to proceed unilaterally with resource extraction and sea-route regulation backed by its partnership with China. A delegation with an attractive fallback has little reason to concede, and recording it explicitly is what allows an agent to walk away credibly.
  • Concessions willing. What the delegation is authorised to give away, and therefore what can be traded. Without this field a negotiation of principled statements never becomes a negotiation of packages.
  • Coalition leanings. Which other delegations this one expects to work with or against. This is what allows blocs to form early rather than emerging only by accident.
  • Negotiating style. How the delegation speaks: legalistic, conciliatory, blunt. Style is not decoration. In a transcript that will later be judged, tone is part of what is being judged.
  • Private instructions. The secret material the delegation must never state verbatim. This field is the reason information isolation matters. If it were shared, the exercise would collapse into a game of open cards.

Every statement in the transcript is stored with six pieces of information: which delegation spoke, which round it was, what kind of statement it was, the full text, the list of tactics the delegation applied, and the list of other delegations it aligned itself with. The last two are what turn a wall of prose into something that can be counted.

Delegations label their own moves from a fixed list of seventeen terms, set in the public brief. The Arctic list runs:

  • anchoring
  • counter-anchoring
  • conditional-offer
  • red-line-signaled
  • verification-demand
  • deadline-pressure
  • coalition-building
  • issue-linkage
  • appeal-to-law
  • appeal-to-precedent
  • sovereignty-assertion
  • freedom-of-navigation-frame
  • side-payment
  • delay-tactic
  • principled-bargaining-frame
  • environmental-frame
  • indigenous-inclusion-frame

Several of these are established concepts rather than inventions of the project. Anchoring is the practice of stating an extreme opening figure in order to drag the eventual settlement toward it, and it draws on the anchoring effect documented by Amos Tversky and Daniel Kahneman in Science in 1974, who found that an arbitrary starting number measurably shifted people's later estimates even when they were paid to be accurate. Issue-linkage means refusing to settle one question except as part of a package with another. A side-payment is a benefit offered outside the disputed issue to buy agreement on it. Principled-bargaining-frame refers to the approach set out in Getting to Yes, which urges parties to argue from interests and objective criteria rather than from fixed positions.

Because agents do not always use the exact word from the list, a short standardising table in the assembly code folds synonyms into the canonical term. Logrolling, package-linkage, and linkage all become issue-linkage. Batna-signaling and red-line-reaffirmation both become red-line-signaled. Face-saving-formula and consensus-appeal both become principled-bargaining-frame. This is a small piece of housekeeping with a real effect on the counts, and it is worth knowing that it happens.

Each tactic label is stored as a detection record carrying a fixed confidence value of 0.8 and a source marked as self-tagging. That number is not a measurement. It is a constant, recording the fact that the label came from the speaker rather than from an independent observer. A reader should treat the tactic counts as a description of what each delegation said it was doing.

The resulting counts are informative in aggregate. In the Arctic session, coalition-building was tagged twenty-five times and red-line-signaled sixteen, against a single appeal-to-precedent. In the South China Sea session, coalition-building appears twenty-four times and conditional-offer fourteen, with delay-tactic appearing only twice. The shape of a negotiation is legible in these numbers before anyone reads a word of the transcript.

The analyst agent then produces four numbers for each delegation.

  • Satisfaction, on a scale from 0 to 1. This expresses how well the analyst judges that a delegation achieved the objectives recorded in its own profile. It is scored against that delegation's stated end states, not against any external standard of a good outcome, which is why a delegation can score well in a session that produced no agreement at all. In the Arctic session, Denmark scored 0.75 and Russia 0.38.
  • Agreements won, a simple count of the understandings the delegation secured. Canada won six in the Arctic session, the highest of the seven.
  • Red lines crossed, a count of the delegation's declared limits that other parties breached. Russia's three in the Arctic session, against zero for Denmark, is the clearest single indicator of why the two scored so differently.
  • Self-rating, a separate figure the delegation assigns to its own performance in its debrief. The United States gave itself 3.3 while the analyst assigned it a satisfaction of 0.6. Keeping the two apart is a useful discipline, since the gap between how a delegation rates itself and how a neutral observer rates it is itself diplomatically interesting.

Each debrief also lists the delegation's goals individually. Every goal carries a priority, either critical or high, a status of achieved, partial, or failed, and a short list of evidence quoting the rounds in which the outcome was settled. This is the part of the output that resists being reduced to a number, and it is the part a reader should check before trusting the number.

A later pass added a second layer to each party profile, describing not only what a delegation wants but how it characteristically bargains. A cultural-background note records the institutional register a delegation tends to negotiate in, whether a high-context consensus style that reads the room before committing or a low-context style that argues directly from the text. A named diplomatic style captures the manner a government is publicly associated with, such as China's assertive wolf-warrior posture, the blunt directness often called cowboy diplomacy in the United States, the reserved consensus-seeking common to Japan, or the flexible hedging Vietnam calls bamboo diplomacy. An alliances field lists the real blocs a country belongs to, from NATO and the Quad to the Shanghai Cooperation Organisation and the China-Russia axis, and a key-figures field names the actual head of government and the foreign and defence ministers in office in the middle of 2026, each with a short note on negotiating style.

These additions are characterisation, not prediction. The named individuals are real office-holders drawn from public reporting, and the notes describe how they are publicly discussed, not any claim about private conduct or how a named person would behave at a real table. Where an office could not be reliably confirmed the field was left empty rather than guessed. The layer sharpens a delegation's voice, and for that reason imports more of the model's biases about real people, which is why it is labelled as sourced characterisation wherever it appears.

06

Reading the Results

Alongside the scoreboard, the analyst writes a neutral summary in the voice of a Special Representative of the Secretary-General, the title given to a senior envoy appointed by the United Nations Secretary-General to convene and mediate. The report carries a headline, a summary, a list of key outcomes, and a list of unresolved issues. The Arctic report concluded that the conference produced a scaffolding of cooperation tracks but no binding settlement, and observed that what broke through was the low-politics agenda, where trust was cheapest, while everything touching sovereignty, jurisdiction, and membership stalled.

The published site replays each session as a negotiation table, with a scenario picker, the full transcript, and the session insights attached to each statement. A separate dashboard page presents each delegation's private reasoning, its proposals, the tactic labels attached to its statements, and the scoreboard, for any chosen scenario.

A third page streams a completed session statement by statement, at adjustable speed, to convey the pace of a plenary. The page states explicitly that this is a replay of a finished simulation and that nothing is being generated in the reader's browser.

Each scenario also has a standalone written report combining the overview, the convener report, the scoreboard, and the complete transcript, for readers who want the record rather than the interface.

07

The Six Scenarios

The Arctic. Seven delegations, four rounds. The issues are the overlapping continental shelf claims of Russia, Canada, and Denmark over the Lomonosov Ridge, which remain before the United Nations Commission on the Limits of the Continental Shelf; the contested legal status of the Northwest Passage and the Northern Sea Route, which Canada and Russia respectively treat as waters under their own control and which the United States regards as international straits; the equal-access provisions of the 1920 Svalbard Treaty and the fisheries zone Norway maintains around the archipelago; militarisation; resources; environment; governance; and the inclusion of Indigenous peoples. The scenario notes that the United States is not a party to the United Nations Convention on the Law of the Sea, which shapes what legal arguments it can and cannot make.

Central Asia and the Fergana Valley. Seven delegations. The issues are the enclave-riddled borders inherited from Soviet administrative divisions, the sharing of Syr Darya water between upstream and downstream states, hydropower, and the competing influence of outside powers.

Cyprus. Six delegations. The issues are the choice between a bizonal bicommunal federation and a two-state settlement, territory, property, security guarantees, and the disputed gas fields of the eastern Mediterranean. This is the only scenario whose source pack predates the others, and it is the scenario the simulator finds hardest to settle.

The South China Sea. Nine delegations, the largest table. The issues are the competing maritime claims, the application of the Law of the Sea Convention, freedom of navigation, and the long-running attempt to agree a code of conduct between China and the states of the Association of Southeast Asian Nations.

The Korean Peninsula. Six delegations, three rounds, convened in Beijing under a Chinese chair as a resumption of the Six-Party Talks. The parties are China, North Korea, South Korea, Japan, Russia, and the United States. The issues are guarantees of the non-use of force and whether a Korean War peace treaty falls inside the process; whether a North Korea returned to the Non-Proliferation Treaty in good standing may hold a civilian nuclear programme; diplomatic normalisation and whether a formal end to the war precedes or follows denuclearisation; sanctions relief and the timing of humanitarian assistance; complete verified dismantlement in one comprehensive package against a rewarded step-by-step sequence; and light-water reactors under a revived Korean Energy Development Organisation. The convening report recorded that Beijing reopened three negotiating channels and kept the Six-Party framework alive without reaching a settlement.

Jammu and Kashmir. Seven delegations, three rounds, convened in Geneva under a United Nations Special Representative pursuant to Security Council Resolution 2900, and joined for the first time by a delegation from Jammu and Kashmir itself alongside India, Pakistan, China, Russia, the United Kingdom, and the United States. The talks are organised into three chapters. The territory chapter covers sovereignty over the former princely state, whether the Line of Control becomes an international border, and the Sino-Indian claims to Aksai Chin and the Shaksam Valley. The governance chapter covers the end state for Jammu and Kashmir, the status of Article 370, and the terms of any plebiscite under the troop-withdrawal conditions of Resolution 47. The human rights chapter covers whether violations are investigated internationally or domestically, the status and repatriation of roughly 950,000 displaced people, and access to the camps and the Kashmir Valley. The convening report recorded that Geneva closed with a single converged instrument, a screened humanitarian relief mechanism for the displaced.

Convening plausibility. A separate question is how plausible it is, in the real world of 2026, that these parties would actually convene and bargain in good faith in the format each scenario stages. This is an editorial judgement about real-world posture, offered as a caveat rather than a result: it is not an output of the simulator and it is not a forecast. A scenario can be a first-rate teaching exercise while describing a table that no government would currently join, and the zero-settlement finding partly measures how far each pack sits from a real, willing negotiation. The ratings below are the author's, on a simple scale of low, moderate, and high.

Central Asia, high. The rare case where good-faith multilateral bargaining actually succeeded: Kyrgyzstan, Tajikistan, and Uzbekistan signed the Treaty of Khujand fixing their border tripoint on 31 March 2025, following the Kyrgyz-Tajik delimitation treaty of the same month. Real negotiation in roughly this format is live and productive.

Cyprus, moderate. The leaders and guarantors do convene under the United Nations, and a fresh informal round was in preparation through 2026, but since the collapse of the Crans-Montana talks in 2017 the sides remain split between a bizonal federation and a two-state settlement, so a good-faith settlement of this scope is not currently in reach.

The Arctic, low. The Arctic Council's political track between Russia and the seven Western states has been suspended since Russia's 2022 invasion of Ukraine, and United States pressure on Denmark over Greenland compounds the freeze. A good-faith seven-party table is not currently thinkable, and China is not a Council member.

The South China Sea, low. China declares the 2016 arbitral award null and void and insists on bilateral rather than multilateral resolution, and the nearest real analogue, the ASEAN-China Code of Conduct, has crawled since 2018 without conclusion.

The Korean Peninsula, low. The Six-Party Talks have been defunct since 2009; North Korea has written its nuclear status into its constitution, adopted a hostile-two-states doctrine renouncing unification, and secured through its 2024 partnership with Russia much of what it once sought at the table.

Jammu and Kashmir, low. India treats the territory as strictly internal, a stance hardened after the 2019 revocation of Article 370, and refuses third-party mediation, so a Security Council conference seating a separate Jammu and Kashmir delegation is unthinkable to New Delhi even as Pakistan seeks exactly that.

Four of the six scenarios rate low, which is a caution against reading any negotiation result here as commentary on how these disputes are likely to move. The packs were built to be teachable, which often means describing a table more balanced and more willing than the real one.

08

Testing Under Uncertainty

A Monte Carlo simulation estimates the range of possible outcomes by running a model many times with randomly drawn inputs and counting how often each result occurs. The technique was devised at Los Alamos in 1946 by Stanislaw Ulam and developed with John von Neumann and Nicholas Metropolis, who supplied the name; Metropolis later recorded that it referred to an uncle of Ulam's who kept borrowing money because he just had to go to Monte Carlo.

A single run tells one story. Whether that story was inevitable or a fluke is a different question, and it is usually the more useful one. Saying that a deal fails in two runs out of eight, and specifically when a hardline shock coincides with a confrontational mood, tells a reader far more about fragility than any single result can.

Each trial draws a fresh value for four things no delegation controls.

  • External shock, drawn from a list written for each scenario. The Arctic list includes a record ice-free summer opening new routes, a Russian naval incident involving a NATO vessel, and a major hydrocarbon discovery in contested waters. The Central Asian list includes a severe drought collapsing Syr Darya flows and a deadly border clash. The South China Sea list includes a collision at Second Thomas Shoal and the declaration of an air defence identification zone. One option on every list is that no major shock occurs, so that the absence of a crisis is itself a sampled condition.
  • Mediator pressure, set to low, medium, or high. This represents how hard the convening envoy pushes for a result.
  • Overall mood, set to cooperative, neutral, or confrontational. This represents the atmosphere in the room, which experienced negotiators treat as a real variable rather than a soft one.
  • Momentum, recording which delegation enters with the initiative.

An agent simulates the outcome under that specific draw and reports the type of deal reached, each party's satisfaction, whether each party's primary goal was achieved, partial, or failed, whether its red line held, and which coalitions formed. Deal types are recorded on a five-point ordered scale: comprehensive, framework, partial, stalemate, breakdown. Ordinary code, with no model involved, then counts the deal types, computes each delegation's satisfaction as a mean, a standard deviation, a median, and a lowest and highest value, calculates the percentage of trials in which each red line held, and ranks the coalitions by how often they recurred. A wide spread means the outcome depends on conditions; a narrow one means it is robust.

Forty trials were run for each scenario, and twenty for the earlier Strait of Hormuz session. Across every trial of every scenario, not one produced a comprehensive settlement. Central Asia was the most stable, returning a framework agreement in twenty-six trials of forty and a partial in thirteen, with a single stalemate and no breakdowns. The Arctic split evenly, sixteen frameworks and sixteen partials, with seven stalemates and a single breakdown. Cyprus was among the most fragile of the original four, returning twelve stalemates and seven breakdowns against twelve frameworks. Of the two scenarios added later, the Korean Peninsula proved the most fragile table in the whole set, reaching a framework in only four trials against eleven partials, sixteen stalemates, and nine breakdowns, while Jammu and Kashmir returned six frameworks, seventeen partials, ten stalemates, and seven breakdowns. Individual variables are equally revealing: in the Arctic trials, China's red line held in eighteen runs of forty and the United States red line in thirty-five of forty, while the five other delegations held theirs in all forty. Counts are given rather than percentages throughout, because a percentage taken over forty trials invites a reader to treat it as a rate estimated from a sample, and it is a census of forty runs of one scenario.

The absence of a comprehensive settlement across every trial reads as a finding about diplomacy and is not one, at least on the scenario where it has been examined. The mixed-model comparison described under Limitations checked the Korean profiles against each other and found that the maximum concession the DPRK is authorised to make and the minimum the United States, the Republic of Korea and Japan are authorised to accept do not overlap at any point. The DPRK's red line forbids reducing its deterrent without acceptable compensation; the other three forbid any residual programme whatsoever. On that table a comprehensive settlement is unreachable by construction, so its absence measures the scenario pack rather than the negotiation, and the figure cannot discriminate between two runs, two models, or two sets of tactics. The same check has not been run on the other five scenarios, and until it has, the zero-settlement result should be read as a property the scenarios may impose rather than an outcome the talks produced. What the trials do measure, and measure usefully, is the distribution below that ceiling: how often a table reaches a framework rather than a partial, how often it breaks down, and which red lines survive varied conditions.

Three further variables were added to each trial. The first records the blocs that formed, distinguishing a standing alliance, such as NATO or the China-Russia axis, from a situational partner of convenience, and rating each bloc's cohesion as tight, loose, or fractured, so that a coalition's strength and durability are legible and not only its membership. The second records the form of any agreement, on a scale from none through a communiqué, a memorandum, and a framework to a full treaty, separating the substance of a deal from its legal weight. The third records, for each party, whether it could realistically ratify and deliver the outcome at home, since a deal at the table is not a deal until it survives domestic ratification.

Because a breakdown can be a win for a party that prefers no agreement, the probability of success is reported not as one number but as several: the share of trials reaching any agreement, the share reaching a framework or better, each party's rate of attaining its primary goal, and the rate at which red lines and ratification hold. Across the six scenarios these figures track the fragility ordering seen in the deal types. Central Asia reaches some agreement in almost every trial and a framework or better in about two thirds of them, while the Korean Peninsula reaches any agreement in fewer than two trials in five and a framework or better in only one in ten. A high in-simulation success rate on a low-plausibility table still describes the scenario pack more than it describes the world.

The methodology page is explicit that each trial is a reduced-form outcome simulation conditioned on the random draw, not a complete re-run of the multi-round negotiation. What the exercise samples is the model's own distribution over plausible outcomes. It is not an empirical distribution of real events, and the randomised conditions are illustrative rather than calibrated probabilities. The right reading is how sensitive the model thinks this outcome is to shocks and mood, and nothing stronger.

09

Grounding in Negotiation Theory

The variables are not arbitrary. Recording a BATNA, red lines, and a set of tradeable concessions for every party reproduces the standard analytic apparatus of the Harvard Negotiation Project, and it is what allows the simulator to distinguish a delegation that conceded from a delegation that had nothing to gain by holding out.

The distinction between claiming value and creating value, developed in the study of labour negotiations by Richard Walton and Robert McKersie in 1965, appears in the structure of the tactic vocabulary. Anchoring, counter-anchoring, and deadline-pressure are moves that divide a fixed quantity. Issue-linkage, side-payments, and conditional offers are moves that enlarge what is available to divide. The counts therefore say something about what kind of negotiation took place, not merely how much of it there was.

One well-established feature of real diplomacy is deliberately absent. Robert Putnam's account of two-level games, published in International Organization in 1988, describes negotiators bargaining simultaneously at an international table and a domestic one, where a leader whose parliament will reject a deal has less room to agree but more leverage to refuse. The simulator captures a trace of this in the delegation role field, which records how much authority a delegation holds, but it does not model the domestic table. No agent faces a legislature, a coalition partner, or an election. This is a substantial simplification of what constrains real negotiators.

The convening role is modelled at a similarly light touch. The public brief provides that a Special Representative chairs the plenary and may invite proposals and summarise, and the analyst writes in that voice afterwards. But no agent plays the mediator during the talks. Mediator pressure appears only as a randomised condition in the Monte Carlo layer, never as an actor making choices in the room.

10

Provenance of the Scenario Material

The scenario packs are adapted from crisis-negotiation exercise materials associated with the International Strategic Crisis Negotiation Exercise programme run by the Center for Strategic Leadership at the United States Army War College, which takes the exercise to universities each year and has developed regional scenario sets including the Arctic, Cyprus, and the South China Sea.

The repository preserves the source documents themselves, organised by scenario, with one public scenario brief and one confidential instruction file per country. Their file names record the institution and academic year of the exercise from which each pack came, which is an unusually clean audit trail for material of this kind.

The material is notional exercise content written for teaching. It does not represent the actual negotiating position of any government, and the simulator inherits that status. Search-engine indexing of the published site is disabled, both through a site-wide instruction file and through per-page tags, so that AI-generated statements attributed to real states do not circulate as though they were reporting.

11

Limitations and Caveats

The repository documents its own limitations at length, and they are serious enough to state in full.

One mind plays every side. All delegations are driven by the same underlying model. Information is isolated, but the adversaries are not genuinely independent actors, and shared assumptions or style can leak across roles. This is the deepest limitation in the design, because it undercuts the premise that a negotiation is a meeting of different minds.

That confound has now been tested, on one scenario. An independent peer review put it plainly: the monoculture is acknowledged but never measured, and the test is cheap. It has since been run for the Korean Peninsula, as a controlled comparison against the single-model baseline holding scenario, profiles, brief, prompts, rounds and information isolation constant and varying only which model sits behind each seat. The two delegations defending the hardest positions stayed on the strongest model, so that a weaker model self-breaching a red line could not be confused with the position failing; the chair and Russia moved to a mid-tier model and the two allied delegations to the smallest.

What that test found is mostly adverse to this report. The coarse outcomes replicated: substituting four of six seats across two model tiers did not change whether the mutual freeze was accepted, whether sanctions moved, or whether a settlement was reached. Three finer outcomes did not replicate, and the bilateral channel on the abductions agreed in the baseline was never convened in the second run. But the central result is that the zero-settlements figure is over-determined by the scenario pack rather than produced by the talks, and it is set out under the Monte Carlo results above. Two further cautions belong with any citation of the comparison. One headline outcome moved between the two runs at a seat whose model never changed, which is ordinary run-to-run variance rather than an effect of substitution, and it bounds how much weight any single reported difference can carry. And the baseline is not a clean reference: it contains the same families of error as the mixed run, including two delegations asserting a five-party coalition the public record does not support and attributing it to the record.

Model and role cannot be separated in that test. Both seats on the smallest model were the two allied non-American delegations and both seats on the mid-tier model were the two step-by-step delegations, so the model and the chair it sits in are perfectly confounded. The claim that a weaker model degrades role fidelity and the claim that the Seoul and Tokyo seats are simply harder to play are indistinguishable on this data. There is one run in each arm, so nothing in the comparison is a rate, a frequency, or an effect size, and it should not be cited as one. Rotating each model across each seat, replicating enough to turn instances into rates, and re-auditing the transcripts with a coder blind to the model assignment are the work that would license more.

There is no ground truth. Nothing in the tool has been validated against a real negotiation, an expert assessment, or a historical outcome. The scoreboard numbers are one system's subjective judgement of another system's writing. Satisfaction should be read as how the analyst rated the argument on the page, not as an assessment of any national interest.

Agents can fabricate. A language model can invent a plausible-sounding treaty, figure, or precedent that appears nowhere in its brief, and can drift from a government's real position in ways that read smoothly.

Training data shapes the outcome. Which arguments sound strong, which framings feel reasonable, and how sympathetically each actor is portrayed all reflect biases in the material the model learned from. A simulator in which one delegation consistently argues more persuasively than another may be reporting a fact about the model rather than about the dispute.

The exercise is compressed. A multi-day negotiation becomes three or four short plenary rounds. Caucuses, bilateral side-channels, corridor conversations, and escalation dynamics are simplified or absent. Much of what determines the outcome of real talks happens outside the plenary, and none of that is here.

The samples are small. Forty trials per scenario is enough to show that Cyprus is fragile and Central Asia is not, and that the two scenarios added later, Korea and Jammu and Kashmir, are more fragile still. It is not enough to support a percentage figure quoted on its own.

The trial count. The Monte Carlo layer now holds forty trials for each of the six scenarios and a further twenty for the earlier Strait of Hormuz session, for two hundred and sixty in total. Anyone reconciling this figure against an earlier count should note that the run was topped up from an initial eight, then twenty, trials per scenario.

The project records a failure honestly. The first orchestration run struck a transient network outage that stopped roughly half its agents mid-negotiation, and a targeted re-run completed the missing rounds. The repository states that no gaps were papered over and that every published statement is a genuine agent output, with the intermediate artefacts committed for inspection. Because language models are not deterministic, the exact wording of any repeat run will differ even from identical source documents.

The isolation has been audited on every scenario. A cross-party contamination audit now exists for all six multi-party scenarios, tracing each statement to the actor's own brief, the shared public brief, or legitimate public inference, and checking whether any delegation's concealed cards ever surfaced through a rival's mouth. None did: in every scenario the tell-tale secrets stayed dark, and the audits rate information integrity as high. What they flag instead is stylistic homogenisation, the single-author fingerprint of one model writing every side, which is the artefact the mixed-model test is built to probe.

The named figures are a characterisation. Attaching the real head of government and the foreign and defence ministers to each delegation sharpens its voice but also imports the model's biases about specific, living individuals. These entries are drawn from public reporting and describe how each figure is publicly discussed; they are not claims about private conduct, and the simulator does not model any named person's actual decisions.

It is not decision support. The repository states directly that the tool must not be used to inform real policy, negotiation, or intelligence judgements.

12

Intended Audience and Use

The prototype is built for teaching and for structured exploration. Its most defensible use is as a companion to a human exercise rather than a replacement for one. A class that has run the Arctic scenario itself can compare its own transcript against the simulated one and argue about where the agents were wrong, which is a more demanding exercise than either activity alone.

It is also of interest to researchers examining what happens when language models are asked to hold a position under pressure. The run artefacts are complete enough to support that kind of scrutiny, which is precisely what makes the honest limitations section valuable rather than decorative.

It is of no use to anyone seeking to know what any government will actually do. The tool says so itself, and this report repeats it because the output is fluent enough to be mistaken for analysis.

13

Conclusion

The Diplomatic Simulator is a modest and carefully documented prototype. Its contribution is less the simulated negotiations themselves than the discipline imposed on them: confidential mandates that stay confidential, statements that are counted and labelled, scores that are defined against each delegation's own stated goals rather than an invented standard of success, and a sensitivity layer that reports how much of the outcome was luck.

The most valuable thing in the repository may be its methodology page, which sets out what the agents do, what they do not do, which model produced each layer, where the run broke, and why none of the numbers carry ground truth. Tools of this kind will become common, and their outputs will be fluent enough that readers extend them more credibility than they have earned. A prototype that publishes its own failure modes alongside its results, and that turns off search indexing so its invented statements are not mistaken for reporting, sets a standard worth borrowing.

References

Sources

  1. 01Meta Fundamental AI Research Diplomacy Team. (2022). Human-level play in the game of Diplomacy by combining language models with strategic reasoning. Science, volume 378, issue 6624.
  2. 02Rivera, J.P., Mukobi, G., Reuel, A., Lamparth, M., Smith, C., and Schneider, J. (2024). Escalation Risks from Language Models in Military and Diplomatic Decision-Making. ACM Conference on Fairness, Accountability, and Transparency.
  3. 03Fisher, R. and Ury, W. (1981). Getting to Yes: Negotiating Agreement Without Giving In. Houghton Mifflin. Source of the BATNA concept and of principled bargaining.
  4. 04Tversky, A. and Kahneman, D. (1974). Judgment under Uncertainty: Heuristics and Biases. Science, volume 185, issue 4157. Source of the anchoring effect.
  5. 05Walton, R.E. and McKersie, R.B. (1965). A Behavioral Theory of Labor Negotiations. McGraw-Hill. Source of the distinction between claiming value and creating value.
  6. 06Putnam, R.D. (1988). Diplomacy and Domestic Politics: The Logic of Two-Level Games. International Organization, volume 42, issue 3.
  7. 07Metropolis, N. (1987). The Beginning of the Monte Carlo Method. Los Alamos Science, special issue.
  8. 08United States Army War College, Center for Strategic Leadership. International Strategic Crisis Negotiation Exercise. Source of the adapted scenario packs and confidential briefs.
  9. 09Convening plausibility, Central Asia. Uzbekistan, Tajikistan, and Kyrgyzstan sign landmark border agreement (Treaty of Khujand), 31 March 2025.
  10. 10Convening plausibility, Cyprus. UN Secretary-General to visit Cyprus for stalled peace talks, 2026.
  11. 11Convening plausibility, South China Sea. Wang Yi reiterates China's rejection of the 2016 arbitration ruling, South China Morning Post, 2025.
  12. 12Convening plausibility, Korea. Kim Jong Un's 'two hostile states' declaration and its legal implications, Daily NK.
  13. 13Convening plausibility, Jammu and Kashmir. Modi tells Trump India will not accept third-party mediation on Kashmir, BBC, 2025.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file +

Publications · Academic report

The Diplomatic Simulator

A Multi-Party Negotiation Simulator Driven by Artificial Intelligence Agents

Ethical Tech CoLab

NYU Center for Global Affairs

July 2026

Carolina Morón. Prepared as masters research at the NYU Center for Global Affairs. The original diplomacy table demonstration and the Strait of Hormuz session report, from which the presentation layer is adapted, were built by Yorke Rhodes III.

Multi-party negotiation is difficult to teach and almost impossible to rehearse. The Diplomatic Simulator asks what happens when the delegations at a crisis table are played by artificial intelligence agents instead of by people. Each delegation is given a confidential brief and nothing else, the talks are run in rounds, and every word spoken is preserved, tagged, and scored. It is a teaching and exploration tool, not a forecasting system, and there is no ground truth anywhere in it.

Key figures

133

statements of record produced by forty-two delegation agents across six multi-party scenarios

17

negotiating tactics in the fixed vocabulary delegations use to label their own moves

40

Monte Carlo trials run for each scenario under randomly varied shocks, pressure, and mood

0

comprehensive settlements in any trial of any scenario, a ceiling the scenario packs impose rather than a finding about diplomacy

01

Executive Summary

The Diplomatic Simulator replays multi-party diplomatic negotiations in which every national delegation is played by an artificial intelligence agent, that is, a language program instructed to speak and reason in the voice of one government. It is a teaching and exploration tool. It is not a forecasting system, and its authors state plainly that it must not be used to inform real policy, negotiation, or intelligence judgements.

Six scenarios have been simulated to date: the Arctic, with seven delegations; Central Asia and the Fergana Valley, with seven; Cyprus reunification, with six; the South China Sea, with nine; the Korean Peninsula Six-Party Talks, with six; and Jammu and Kashmir, with seven. An earlier two-party session on the Strait of Hormuz, involving Iran and the United States, is preserved as the original demonstration. Across the six multi-party scenarios, forty-two delegation agents produced one hundred and thirty-three statements of record.

The organising principle of the design is information isolation. When a delegation agent writes its statement for a round, it receives only its own confidential brief, a neutral public brief that everyone shares, and the public transcript of what has already been said aloud. It never sees another government's private instructions, and it never sees another delegation's private reasoning. This mirrors the informational structure of real diplomacy, where the public record is common property and the mandate is not.

The scenario material is not invented. The confidential briefs are adapted from crisis-negotiation exercise packs associated with the United States Army War College International Strategic Crisis Negotiation Exercise programme, which are taken to universities. The file names preserved in the repository record where each pack was used, including New York University, Penn, Syracuse, and the University of Minnesota.

After the talks conclude, a separate analyst agent reads the whole transcript and produces a scoreboard, a debrief for each delegation, and a neutral summary written in the voice of a convening envoy. A further layer re-runs each scenario forty times under randomly varied external conditions, in order to show which outcomes are stable and which depend on luck.

The repository is candid that every number on the site is one artificial intelligence system's subjective judgement of another artificial intelligence system's writing. There is no ground truth anywhere in the tool. The most serious structural limitation is that a single underlying model plays every side of every table, so the adversaries are not genuinely independent minds. That limitation has now been tested once, on the Korean scenario, by re-running it with different models behind four of the six seats. The coarse outcomes replicated. The test also established something this report had not seen: on that scenario a comprehensive settlement is impossible by construction, because the concessions the parties are authorised to make do not overlap, so the zero-settlement result measures the scenario pack rather than the negotiation. The Limitations section sets out what the comparison does and does not license.

02

Background and Rationale

The problem. Crisis-negotiation exercises are among the most effective ways to teach diplomacy, because they force a participant to argue a position they may not hold, under time pressure, against people who have been told to resist them. But an exercise consumes an enormous amount of scarce human attention. A seven-party Arctic exercise needs seven teams, seven mentors, and two days. It can be run perhaps once a year at any given institution.

The second problem is that exercises leave almost no durable record. What a delegation said in round two, why it said it, and how that changed the shape of the room afterwards is rarely written down in a form anyone can study later. The learning is real but it is locked inside the experience of having been there.

The gap. Existing computational work on negotiation has largely concentrated on games with fixed rules. The best known example is CICERO, the system developed by Meta's Fundamental AI Research team and reported in Science in 2022, which reached human-level play in the board game Diplomacy by combining a language model with strategic planning. That is a considerable achievement, but a board game has a scoring function and a win condition. Real multi-party talks have neither. A separate strand of work has examined what happens when language models are placed in simulated wargames: Rivera and colleagues, publishing at the ACM Conference on Fairness, Accountability, and Transparency in 2024, found that all five models they tested displayed escalation patterns that were difficult to predict. That finding is a caution rather than an endorsement, and this prototype should be read against it.

The response. The Diplomatic Simulator occupies the space between the two. It takes the scenario packs and confidential briefs that human exercises already use, replaces the human delegations with agents, and keeps the entire record. What the tool produces is not a prediction of what governments would do. It is a legible artefact: a transcript, a set of tagged tactics, and a scoreboard, all of which can be read, disputed, and compared against what a human exercise produced from the same brief.

The design commitment that makes this worth doing is transparency about provenance. Every intermediate product of every run is committed to the repository in a readable form: the extracted party profiles, the public brief, the transcript, the analysis. A reader who doubts a score can open the transcript and check what was actually said.

03

Objectives

The prototype is designed to:

  1. Reproduce the structure of a multi-party crisis negotiation, including confidential mandates, public plenary statements, coalition formation, and a convening mediator.
  2. Preserve information isolation, so that no delegation gains an advantage the corresponding human team would not have had.
  3. Produce a complete and inspectable record of each session, including what each delegation said, which negotiating tactics it used, and how an independent analyst judged the result.
  4. Express the outcome in variables simple enough for a non-specialist to read, while keeping the reasoning behind each variable visible.
  5. Test how sensitive each outcome is to conditions no negotiator controls, by re-running each scenario under randomly varied external shocks and moods.
  6. Document its own method and its own failures, including the specific point at which an orchestration run broke and had to be repeated.

04

How the Simulator Works

The tool runs each scenario through six stages. The first five use the same reusable toolchain, held in the repository under the folder named sim; only the source documents change from scenario to scenario. Three of the six stages are ordinary computer code with no artificial intelligence involved at all.

Stage one, reading the source documents. Each scenario arrives as a set of documents: one public scenario brief describing the crisis, plus one confidential instruction file for each country. These are converted from page images into plain text. No judgement is exercised at this stage.

Stage two, building a party profile. One agent is assigned to each country. It reads that country's confidential instructions and nothing else, and distils them into a structured profile. This is the single most important step in the whole pipeline, because the profile is the entirety of what that delegation will know about itself for the rest of the negotiation.

Stage three, writing the public brief. A separate agent reads the shared scenario document and writes the neutral briefing that every delegation will see. It states the situation, lists the issues formally on the table, sets out the procedure, and fixes the vocabulary of negotiating tactics that delegations will use to label their own moves.

Stage four, the negotiation. The delegations negotiate in plenary rounds. In each round, every delegation writes exactly one statement. It is given its own profile, the public brief, and the running public transcript. It is instructed to stay in role, to ground its claims in its brief, to protect its secret bottom lines, and to label the tactics it has just used. The flagship Arctic scenario runs four rounds: an opening plenary, a positioning round, a bargaining and coalitions round, and a closing plenary. The other three scenarios run three rounds, dropping the separate positioning stage. Seven delegations across four rounds produce twenty-eight statements in the Arctic; nine delegations across three rounds produce twenty-seven in the South China Sea.

Stage five, the analysis. A single analyst agent, cast as a neutral control group, then reads the entire transcript together with every delegation's confidential profile. It is the only agent in the system that sees both sides of the information barrier. It produces the scoreboard, the per-delegation debriefs, and the convener report.

Stage six, assembly and publication. Plain computer code stitches the outputs together, standardises the tactic labels, and renders the interactive pages. No model is involved. This matters for auditing: the numbers a reader sees on the published pages are arithmetic performed on agent outputs, not a second layer of interpretation.

05

The Variables, and Why Each One Exists

Everything the simulator produces rests on a small number of variables, each of which has a plain meaning. When an agent reads a country's confidential instructions, it fills in eleven fields. Together they constitute the delegation's entire identity.

The eleven fields are as follows.

  • Delegation role. One sentence describing who this delegation formally is and how much authority it holds. The Russian profile in the Arctic scenario records a foreign ministry delegation mandated to defend Arctic sovereignty while deferring major deviations to the Foreign Minister. This field exists because a negotiator's freedom to concede is itself a variable, and a delegation that must telephone home behaves differently from one that can sign.
  • Fundamental principles. The standing commitments the government would assert regardless of this particular crisis. These give the agent something to argue from when the transcript moves somewhere its specific instructions did not anticipate.
  • Desired end state, primary. The delegation's definition of victory.
  • Desired end state, alternate. Its definition of an acceptable substitute. Having two rather than one is deliberate: a negotiator with only a maximum position cannot trade.
  • Key positions by issue. The government's stated line on each numbered issue in the public brief, stored issue by issue. This is what keeps a delegation consistent across rounds and prevents it from quietly abandoning a position it took an hour earlier.
  • Red lines. The commitments the delegation states it will not cross. In negotiation practice a red line is a declared non-negotiable limit whose value depends entirely on whether other parties believe it. The simulator makes this measurable, because the analyst later records how many of a delegation's red lines were crossed by others.
  • BATNA. A term of art from negotiation theory, introduced by Roger Fisher and William Ury in Getting to Yes in 1981. It stands for Best Alternative to a Negotiated Agreement, and it means the outcome a party falls back on if the talks collapse. It is the benchmark against which every proposal should be judged: a rational party accepts nothing worse than its BATNA. Russia's Arctic BATNA is to proceed unilaterally with resource extraction and sea-route regulation backed by its partnership with China. A delegation with an attractive fallback has little reason to concede, and recording it explicitly is what allows an agent to walk away credibly.
  • Concessions willing. What the delegation is authorised to give away, and therefore what can be traded. Without this field a negotiation of principled statements never becomes a negotiation of packages.
  • Coalition leanings. Which other delegations this one expects to work with or against. This is what allows blocs to form early rather than emerging only by accident.
  • Negotiating style. How the delegation speaks: legalistic, conciliatory, blunt. Style is not decoration. In a transcript that will later be judged, tone is part of what is being judged.
  • Private instructions. The secret material the delegation must never state verbatim. This field is the reason information isolation matters. If it were shared, the exercise would collapse into a game of open cards.

Every statement in the transcript is stored with six pieces of information: which delegation spoke, which round it was, what kind of statement it was, the full text, the list of tactics the delegation applied, and the list of other delegations it aligned itself with. The last two are what turn a wall of prose into something that can be counted.

Delegations label their own moves from a fixed list of seventeen terms, set in the public brief. The Arctic list runs:

  • anchoring
  • counter-anchoring
  • conditional-offer
  • red-line-signaled
  • verification-demand
  • deadline-pressure
  • coalition-building
  • issue-linkage
  • appeal-to-law
  • appeal-to-precedent
  • sovereignty-assertion
  • freedom-of-navigation-frame
  • side-payment
  • delay-tactic
  • principled-bargaining-frame
  • environmental-frame
  • indigenous-inclusion-frame

Several of these are established concepts rather than inventions of the project. Anchoring is the practice of stating an extreme opening figure in order to drag the eventual settlement toward it, and it draws on the anchoring effect documented by Amos Tversky and Daniel Kahneman in Science in 1974, who found that an arbitrary starting number measurably shifted people's later estimates even when they were paid to be accurate. Issue-linkage means refusing to settle one question except as part of a package with another. A side-payment is a benefit offered outside the disputed issue to buy agreement on it. Principled-bargaining-frame refers to the approach set out in Getting to Yes, which urges parties to argue from interests and objective criteria rather than from fixed positions.

Because agents do not always use the exact word from the list, a short standardising table in the assembly code folds synonyms into the canonical term. Logrolling, package-linkage, and linkage all become issue-linkage. Batna-signaling and red-line-reaffirmation both become red-line-signaled. Face-saving-formula and consensus-appeal both become principled-bargaining-frame. This is a small piece of housekeeping with a real effect on the counts, and it is worth knowing that it happens.

Each tactic label is stored as a detection record carrying a fixed confidence value of 0.8 and a source marked as self-tagging. That number is not a measurement. It is a constant, recording the fact that the label came from the speaker rather than from an independent observer. A reader should treat the tactic counts as a description of what each delegation said it was doing.

The resulting counts are informative in aggregate. In the Arctic session, coalition-building was tagged twenty-five times and red-line-signaled sixteen, against a single appeal-to-precedent. In the South China Sea session, coalition-building appears twenty-four times and conditional-offer fourteen, with delay-tactic appearing only twice. The shape of a negotiation is legible in these numbers before anyone reads a word of the transcript.

The analyst agent then produces four numbers for each delegation.

  • Satisfaction, on a scale from 0 to 1. This expresses how well the analyst judges that a delegation achieved the objectives recorded in its own profile. It is scored against that delegation's stated end states, not against any external standard of a good outcome, which is why a delegation can score well in a session that produced no agreement at all. In the Arctic session, Denmark scored 0.75 and Russia 0.38.
  • Agreements won, a simple count of the understandings the delegation secured. Canada won six in the Arctic session, the highest of the seven.
  • Red lines crossed, a count of the delegation's declared limits that other parties breached. Russia's three in the Arctic session, against zero for Denmark, is the clearest single indicator of why the two scored so differently.
  • Self-rating, a separate figure the delegation assigns to its own performance in its debrief. The United States gave itself 3.3 while the analyst assigned it a satisfaction of 0.6. Keeping the two apart is a useful discipline, since the gap between how a delegation rates itself and how a neutral observer rates it is itself diplomatically interesting.

Each debrief also lists the delegation's goals individually. Every goal carries a priority, either critical or high, a status of achieved, partial, or failed, and a short list of evidence quoting the rounds in which the outcome was settled. This is the part of the output that resists being reduced to a number, and it is the part a reader should check before trusting the number.

A later pass added a second layer to each party profile, describing not only what a delegation wants but how it characteristically bargains. A cultural-background note records the institutional register a delegation tends to negotiate in, whether a high-context consensus style that reads the room before committing or a low-context style that argues directly from the text. A named diplomatic style captures the manner a government is publicly associated with, such as China's assertive wolf-warrior posture, the blunt directness often called cowboy diplomacy in the United States, the reserved consensus-seeking common to Japan, or the flexible hedging Vietnam calls bamboo diplomacy. An alliances field lists the real blocs a country belongs to, from NATO and the Quad to the Shanghai Cooperation Organisation and the China-Russia axis, and a key-figures field names the actual head of government and the foreign and defence ministers in office in the middle of 2026, each with a short note on negotiating style.

These additions are characterisation, not prediction. The named individuals are real office-holders drawn from public reporting, and the notes describe how they are publicly discussed, not any claim about private conduct or how a named person would behave at a real table. Where an office could not be reliably confirmed the field was left empty rather than guessed. The layer sharpens a delegation's voice, and for that reason imports more of the model's biases about real people, which is why it is labelled as sourced characterisation wherever it appears.

06

Reading the Results

Alongside the scoreboard, the analyst writes a neutral summary in the voice of a Special Representative of the Secretary-General, the title given to a senior envoy appointed by the United Nations Secretary-General to convene and mediate. The report carries a headline, a summary, a list of key outcomes, and a list of unresolved issues. The Arctic report concluded that the conference produced a scaffolding of cooperation tracks but no binding settlement, and observed that what broke through was the low-politics agenda, where trust was cheapest, while everything touching sovereignty, jurisdiction, and membership stalled.

The published site replays each session as a negotiation table, with a scenario picker, the full transcript, and the session insights attached to each statement. A separate dashboard page presents each delegation's private reasoning, its proposals, the tactic labels attached to its statements, and the scoreboard, for any chosen scenario.

A third page streams a completed session statement by statement, at adjustable speed, to convey the pace of a plenary. The page states explicitly that this is a replay of a finished simulation and that nothing is being generated in the reader's browser.

Each scenario also has a standalone written report combining the overview, the convener report, the scoreboard, and the complete transcript, for readers who want the record rather than the interface.

07

The Six Scenarios

The Arctic. Seven delegations, four rounds. The issues are the overlapping continental shelf claims of Russia, Canada, and Denmark over the Lomonosov Ridge, which remain before the United Nations Commission on the Limits of the Continental Shelf; the contested legal status of the Northwest Passage and the Northern Sea Route, which Canada and Russia respectively treat as waters under their own control and which the United States regards as international straits; the equal-access provisions of the 1920 Svalbard Treaty and the fisheries zone Norway maintains around the archipelago; militarisation; resources; environment; governance; and the inclusion of Indigenous peoples. The scenario notes that the United States is not a party to the United Nations Convention on the Law of the Sea, which shapes what legal arguments it can and cannot make.

Central Asia and the Fergana Valley. Seven delegations. The issues are the enclave-riddled borders inherited from Soviet administrative divisions, the sharing of Syr Darya water between upstream and downstream states, hydropower, and the competing influence of outside powers.

Cyprus. Six delegations. The issues are the choice between a bizonal bicommunal federation and a two-state settlement, territory, property, security guarantees, and the disputed gas fields of the eastern Mediterranean. This is the only scenario whose source pack predates the others, and it is the scenario the simulator finds hardest to settle.

The South China Sea. Nine delegations, the largest table. The issues are the competing maritime claims, the application of the Law of the Sea Convention, freedom of navigation, and the long-running attempt to agree a code of conduct between China and the states of the Association of Southeast Asian Nations.

The Korean Peninsula. Six delegations, three rounds, convened in Beijing under a Chinese chair as a resumption of the Six-Party Talks. The parties are China, North Korea, South Korea, Japan, Russia, and the United States. The issues are guarantees of the non-use of force and whether a Korean War peace treaty falls inside the process; whether a North Korea returned to the Non-Proliferation Treaty in good standing may hold a civilian nuclear programme; diplomatic normalisation and whether a formal end to the war precedes or follows denuclearisation; sanctions relief and the timing of humanitarian assistance; complete verified dismantlement in one comprehensive package against a rewarded step-by-step sequence; and light-water reactors under a revived Korean Energy Development Organisation. The convening report recorded that Beijing reopened three negotiating channels and kept the Six-Party framework alive without reaching a settlement.

Jammu and Kashmir. Seven delegations, three rounds, convened in Geneva under a United Nations Special Representative pursuant to Security Council Resolution 2900, and joined for the first time by a delegation from Jammu and Kashmir itself alongside India, Pakistan, China, Russia, the United Kingdom, and the United States. The talks are organised into three chapters. The territory chapter covers sovereignty over the former princely state, whether the Line of Control becomes an international border, and the Sino-Indian claims to Aksai Chin and the Shaksam Valley. The governance chapter covers the end state for Jammu and Kashmir, the status of Article 370, and the terms of any plebiscite under the troop-withdrawal conditions of Resolution 47. The human rights chapter covers whether violations are investigated internationally or domestically, the status and repatriation of roughly 950,000 displaced people, and access to the camps and the Kashmir Valley. The convening report recorded that Geneva closed with a single converged instrument, a screened humanitarian relief mechanism for the displaced.

Convening plausibility. A separate question is how plausible it is, in the real world of 2026, that these parties would actually convene and bargain in good faith in the format each scenario stages. This is an editorial judgement about real-world posture, offered as a caveat rather than a result: it is not an output of the simulator and it is not a forecast. A scenario can be a first-rate teaching exercise while describing a table that no government would currently join, and the zero-settlement finding partly measures how far each pack sits from a real, willing negotiation. The ratings below are the author's, on a simple scale of low, moderate, and high.

Central Asia, high. The rare case where good-faith multilateral bargaining actually succeeded: Kyrgyzstan, Tajikistan, and Uzbekistan signed the Treaty of Khujand fixing their border tripoint on 31 March 2025, following the Kyrgyz-Tajik delimitation treaty of the same month. Real negotiation in roughly this format is live and productive.

Cyprus, moderate. The leaders and guarantors do convene under the United Nations, and a fresh informal round was in preparation through 2026, but since the collapse of the Crans-Montana talks in 2017 the sides remain split between a bizonal federation and a two-state settlement, so a good-faith settlement of this scope is not currently in reach.

The Arctic, low. The Arctic Council's political track between Russia and the seven Western states has been suspended since Russia's 2022 invasion of Ukraine, and United States pressure on Denmark over Greenland compounds the freeze. A good-faith seven-party table is not currently thinkable, and China is not a Council member.

The South China Sea, low. China declares the 2016 arbitral award null and void and insists on bilateral rather than multilateral resolution, and the nearest real analogue, the ASEAN-China Code of Conduct, has crawled since 2018 without conclusion.

The Korean Peninsula, low. The Six-Party Talks have been defunct since 2009; North Korea has written its nuclear status into its constitution, adopted a hostile-two-states doctrine renouncing unification, and secured through its 2024 partnership with Russia much of what it once sought at the table.

Jammu and Kashmir, low. India treats the territory as strictly internal, a stance hardened after the 2019 revocation of Article 370, and refuses third-party mediation, so a Security Council conference seating a separate Jammu and Kashmir delegation is unthinkable to New Delhi even as Pakistan seeks exactly that.

Four of the six scenarios rate low, which is a caution against reading any negotiation result here as commentary on how these disputes are likely to move. The packs were built to be teachable, which often means describing a table more balanced and more willing than the real one.

08

Testing Under Uncertainty

A Monte Carlo simulation estimates the range of possible outcomes by running a model many times with randomly drawn inputs and counting how often each result occurs. The technique was devised at Los Alamos in 1946 by Stanislaw Ulam and developed with John von Neumann and Nicholas Metropolis, who supplied the name; Metropolis later recorded that it referred to an uncle of Ulam's who kept borrowing money because he just had to go to Monte Carlo.

A single run tells one story. Whether that story was inevitable or a fluke is a different question, and it is usually the more useful one. Saying that a deal fails in two runs out of eight, and specifically when a hardline shock coincides with a confrontational mood, tells a reader far more about fragility than any single result can.

Each trial draws a fresh value for four things no delegation controls.

  • External shock, drawn from a list written for each scenario. The Arctic list includes a record ice-free summer opening new routes, a Russian naval incident involving a NATO vessel, and a major hydrocarbon discovery in contested waters. The Central Asian list includes a severe drought collapsing Syr Darya flows and a deadly border clash. The South China Sea list includes a collision at Second Thomas Shoal and the declaration of an air defence identification zone. One option on every list is that no major shock occurs, so that the absence of a crisis is itself a sampled condition.
  • Mediator pressure, set to low, medium, or high. This represents how hard the convening envoy pushes for a result.
  • Overall mood, set to cooperative, neutral, or confrontational. This represents the atmosphere in the room, which experienced negotiators treat as a real variable rather than a soft one.
  • Momentum, recording which delegation enters with the initiative.

An agent simulates the outcome under that specific draw and reports the type of deal reached, each party's satisfaction, whether each party's primary goal was achieved, partial, or failed, whether its red line held, and which coalitions formed. Deal types are recorded on a five-point ordered scale: comprehensive, framework, partial, stalemate, breakdown. Ordinary code, with no model involved, then counts the deal types, computes each delegation's satisfaction as a mean, a standard deviation, a median, and a lowest and highest value, calculates the percentage of trials in which each red line held, and ranks the coalitions by how often they recurred. A wide spread means the outcome depends on conditions; a narrow one means it is robust.

Forty trials were run for each scenario, and twenty for the earlier Strait of Hormuz session. Across every trial of every scenario, not one produced a comprehensive settlement. Central Asia was the most stable, returning a framework agreement in twenty-six trials of forty and a partial in thirteen, with a single stalemate and no breakdowns. The Arctic split evenly, sixteen frameworks and sixteen partials, with seven stalemates and a single breakdown. Cyprus was among the most fragile of the original four, returning twelve stalemates and seven breakdowns against twelve frameworks. Of the two scenarios added later, the Korean Peninsula proved the most fragile table in the whole set, reaching a framework in only four trials against eleven partials, sixteen stalemates, and nine breakdowns, while Jammu and Kashmir returned six frameworks, seventeen partials, ten stalemates, and seven breakdowns. Individual variables are equally revealing: in the Arctic trials, China's red line held in eighteen runs of forty and the United States red line in thirty-five of forty, while the five other delegations held theirs in all forty. Counts are given rather than percentages throughout, because a percentage taken over forty trials invites a reader to treat it as a rate estimated from a sample, and it is a census of forty runs of one scenario.

The absence of a comprehensive settlement across every trial reads as a finding about diplomacy and is not one, at least on the scenario where it has been examined. The mixed-model comparison described under Limitations checked the Korean profiles against each other and found that the maximum concession the DPRK is authorised to make and the minimum the United States, the Republic of Korea and Japan are authorised to accept do not overlap at any point. The DPRK's red line forbids reducing its deterrent without acceptable compensation; the other three forbid any residual programme whatsoever. On that table a comprehensive settlement is unreachable by construction, so its absence measures the scenario pack rather than the negotiation, and the figure cannot discriminate between two runs, two models, or two sets of tactics. The same check has not been run on the other five scenarios, and until it has, the zero-settlement result should be read as a property the scenarios may impose rather than an outcome the talks produced. What the trials do measure, and measure usefully, is the distribution below that ceiling: how often a table reaches a framework rather than a partial, how often it breaks down, and which red lines survive varied conditions.

Three further variables were added to each trial. The first records the blocs that formed, distinguishing a standing alliance, such as NATO or the China-Russia axis, from a situational partner of convenience, and rating each bloc's cohesion as tight, loose, or fractured, so that a coalition's strength and durability are legible and not only its membership. The second records the form of any agreement, on a scale from none through a communiqué, a memorandum, and a framework to a full treaty, separating the substance of a deal from its legal weight. The third records, for each party, whether it could realistically ratify and deliver the outcome at home, since a deal at the table is not a deal until it survives domestic ratification.

Because a breakdown can be a win for a party that prefers no agreement, the probability of success is reported not as one number but as several: the share of trials reaching any agreement, the share reaching a framework or better, each party's rate of attaining its primary goal, and the rate at which red lines and ratification hold. Across the six scenarios these figures track the fragility ordering seen in the deal types. Central Asia reaches some agreement in almost every trial and a framework or better in about two thirds of them, while the Korean Peninsula reaches any agreement in fewer than two trials in five and a framework or better in only one in ten. A high in-simulation success rate on a low-plausibility table still describes the scenario pack more than it describes the world.

The methodology page is explicit that each trial is a reduced-form outcome simulation conditioned on the random draw, not a complete re-run of the multi-round negotiation. What the exercise samples is the model's own distribution over plausible outcomes. It is not an empirical distribution of real events, and the randomised conditions are illustrative rather than calibrated probabilities. The right reading is how sensitive the model thinks this outcome is to shocks and mood, and nothing stronger.

09

Grounding in Negotiation Theory

The variables are not arbitrary. Recording a BATNA, red lines, and a set of tradeable concessions for every party reproduces the standard analytic apparatus of the Harvard Negotiation Project, and it is what allows the simulator to distinguish a delegation that conceded from a delegation that had nothing to gain by holding out.

The distinction between claiming value and creating value, developed in the study of labour negotiations by Richard Walton and Robert McKersie in 1965, appears in the structure of the tactic vocabulary. Anchoring, counter-anchoring, and deadline-pressure are moves that divide a fixed quantity. Issue-linkage, side-payments, and conditional offers are moves that enlarge what is available to divide. The counts therefore say something about what kind of negotiation took place, not merely how much of it there was.

One well-established feature of real diplomacy is deliberately absent. Robert Putnam's account of two-level games, published in International Organization in 1988, describes negotiators bargaining simultaneously at an international table and a domestic one, where a leader whose parliament will reject a deal has less room to agree but more leverage to refuse. The simulator captures a trace of this in the delegation role field, which records how much authority a delegation holds, but it does not model the domestic table. No agent faces a legislature, a coalition partner, or an election. This is a substantial simplification of what constrains real negotiators.

The convening role is modelled at a similarly light touch. The public brief provides that a Special Representative chairs the plenary and may invite proposals and summarise, and the analyst writes in that voice afterwards. But no agent plays the mediator during the talks. Mediator pressure appears only as a randomised condition in the Monte Carlo layer, never as an actor making choices in the room.

10

Provenance of the Scenario Material

The scenario packs are adapted from crisis-negotiation exercise materials associated with the International Strategic Crisis Negotiation Exercise programme run by the Center for Strategic Leadership at the United States Army War College, which takes the exercise to universities each year and has developed regional scenario sets including the Arctic, Cyprus, and the South China Sea.

The repository preserves the source documents themselves, organised by scenario, with one public scenario brief and one confidential instruction file per country. Their file names record the institution and academic year of the exercise from which each pack came, which is an unusually clean audit trail for material of this kind.

The material is notional exercise content written for teaching. It does not represent the actual negotiating position of any government, and the simulator inherits that status. Search-engine indexing of the published site is disabled, both through a site-wide instruction file and through per-page tags, so that AI-generated statements attributed to real states do not circulate as though they were reporting.

11

Limitations and Caveats

The repository documents its own limitations at length, and they are serious enough to state in full.

One mind plays every side. All delegations are driven by the same underlying model. Information is isolated, but the adversaries are not genuinely independent actors, and shared assumptions or style can leak across roles. This is the deepest limitation in the design, because it undercuts the premise that a negotiation is a meeting of different minds.

That confound has now been tested, on one scenario. An independent peer review put it plainly: the monoculture is acknowledged but never measured, and the test is cheap. It has since been run for the Korean Peninsula, as a controlled comparison against the single-model baseline holding scenario, profiles, brief, prompts, rounds and information isolation constant and varying only which model sits behind each seat. The two delegations defending the hardest positions stayed on the strongest model, so that a weaker model self-breaching a red line could not be confused with the position failing; the chair and Russia moved to a mid-tier model and the two allied delegations to the smallest.

What that test found is mostly adverse to this report. The coarse outcomes replicated: substituting four of six seats across two model tiers did not change whether the mutual freeze was accepted, whether sanctions moved, or whether a settlement was reached. Three finer outcomes did not replicate, and the bilateral channel on the abductions agreed in the baseline was never convened in the second run. But the central result is that the zero-settlements figure is over-determined by the scenario pack rather than produced by the talks, and it is set out under the Monte Carlo results above. Two further cautions belong with any citation of the comparison. One headline outcome moved between the two runs at a seat whose model never changed, which is ordinary run-to-run variance rather than an effect of substitution, and it bounds how much weight any single reported difference can carry. And the baseline is not a clean reference: it contains the same families of error as the mixed run, including two delegations asserting a five-party coalition the public record does not support and attributing it to the record.

Model and role cannot be separated in that test. Both seats on the smallest model were the two allied non-American delegations and both seats on the mid-tier model were the two step-by-step delegations, so the model and the chair it sits in are perfectly confounded. The claim that a weaker model degrades role fidelity and the claim that the Seoul and Tokyo seats are simply harder to play are indistinguishable on this data. There is one run in each arm, so nothing in the comparison is a rate, a frequency, or an effect size, and it should not be cited as one. Rotating each model across each seat, replicating enough to turn instances into rates, and re-auditing the transcripts with a coder blind to the model assignment are the work that would license more.

There is no ground truth. Nothing in the tool has been validated against a real negotiation, an expert assessment, or a historical outcome. The scoreboard numbers are one system's subjective judgement of another system's writing. Satisfaction should be read as how the analyst rated the argument on the page, not as an assessment of any national interest.

Agents can fabricate. A language model can invent a plausible-sounding treaty, figure, or precedent that appears nowhere in its brief, and can drift from a government's real position in ways that read smoothly.

Training data shapes the outcome. Which arguments sound strong, which framings feel reasonable, and how sympathetically each actor is portrayed all reflect biases in the material the model learned from. A simulator in which one delegation consistently argues more persuasively than another may be reporting a fact about the model rather than about the dispute.

The exercise is compressed. A multi-day negotiation becomes three or four short plenary rounds. Caucuses, bilateral side-channels, corridor conversations, and escalation dynamics are simplified or absent. Much of what determines the outcome of real talks happens outside the plenary, and none of that is here.

The samples are small. Forty trials per scenario is enough to show that Cyprus is fragile and Central Asia is not, and that the two scenarios added later, Korea and Jammu and Kashmir, are more fragile still. It is not enough to support a percentage figure quoted on its own.

The trial count. The Monte Carlo layer now holds forty trials for each of the six scenarios and a further twenty for the earlier Strait of Hormuz session, for two hundred and sixty in total. Anyone reconciling this figure against an earlier count should note that the run was topped up from an initial eight, then twenty, trials per scenario.

The project records a failure honestly. The first orchestration run struck a transient network outage that stopped roughly half its agents mid-negotiation, and a targeted re-run completed the missing rounds. The repository states that no gaps were papered over and that every published statement is a genuine agent output, with the intermediate artefacts committed for inspection. Because language models are not deterministic, the exact wording of any repeat run will differ even from identical source documents.

The isolation has been audited on every scenario. A cross-party contamination audit now exists for all six multi-party scenarios, tracing each statement to the actor's own brief, the shared public brief, or legitimate public inference, and checking whether any delegation's concealed cards ever surfaced through a rival's mouth. None did: in every scenario the tell-tale secrets stayed dark, and the audits rate information integrity as high. What they flag instead is stylistic homogenisation, the single-author fingerprint of one model writing every side, which is the artefact the mixed-model test is built to probe.

The named figures are a characterisation. Attaching the real head of government and the foreign and defence ministers to each delegation sharpens its voice but also imports the model's biases about specific, living individuals. These entries are drawn from public reporting and describe how each figure is publicly discussed; they are not claims about private conduct, and the simulator does not model any named person's actual decisions.

It is not decision support. The repository states directly that the tool must not be used to inform real policy, negotiation, or intelligence judgements.

12

Intended Audience and Use

The prototype is built for teaching and for structured exploration. Its most defensible use is as a companion to a human exercise rather than a replacement for one. A class that has run the Arctic scenario itself can compare its own transcript against the simulated one and argue about where the agents were wrong, which is a more demanding exercise than either activity alone.

It is also of interest to researchers examining what happens when language models are asked to hold a position under pressure. The run artefacts are complete enough to support that kind of scrutiny, which is precisely what makes the honest limitations section valuable rather than decorative.

It is of no use to anyone seeking to know what any government will actually do. The tool says so itself, and this report repeats it because the output is fluent enough to be mistaken for analysis.

13

Conclusion

The Diplomatic Simulator is a modest and carefully documented prototype. Its contribution is less the simulated negotiations themselves than the discipline imposed on them: confidential mandates that stay confidential, statements that are counted and labelled, scores that are defined against each delegation's own stated goals rather than an invented standard of success, and a sensitivity layer that reports how much of the outcome was luck.

The most valuable thing in the repository may be its methodology page, which sets out what the agents do, what they do not do, which model produced each layer, where the run broke, and why none of the numbers carry ground truth. Tools of this kind will become common, and their outputs will be fluent enough that readers extend them more credibility than they have earned. A prototype that publishes its own failure modes alongside its results, and that turns off search indexing so its invented statements are not mistaken for reporting, sets a standard worth borrowing.

References

Sources

  1. 01Meta Fundamental AI Research Diplomacy Team. (2022). Human-level play in the game of Diplomacy by combining language models with strategic reasoning. Science, volume 378, issue 6624.
  2. 02Rivera, J.P., Mukobi, G., Reuel, A., Lamparth, M., Smith, C., and Schneider, J. (2024). Escalation Risks from Language Models in Military and Diplomatic Decision-Making. ACM Conference on Fairness, Accountability, and Transparency.
  3. 03Fisher, R. and Ury, W. (1981). Getting to Yes: Negotiating Agreement Without Giving In. Houghton Mifflin. Source of the BATNA concept and of principled bargaining.
  4. 04Tversky, A. and Kahneman, D. (1974). Judgment under Uncertainty: Heuristics and Biases. Science, volume 185, issue 4157. Source of the anchoring effect.
  5. 05Walton, R.E. and McKersie, R.B. (1965). A Behavioral Theory of Labor Negotiations. McGraw-Hill. Source of the distinction between claiming value and creating value.
  6. 06Putnam, R.D. (1988). Diplomacy and Domestic Politics: The Logic of Two-Level Games. International Organization, volume 42, issue 3.
  7. 07Metropolis, N. (1987). The Beginning of the Monte Carlo Method. Los Alamos Science, special issue.
  8. 08United States Army War College, Center for Strategic Leadership. International Strategic Crisis Negotiation Exercise. Source of the adapted scenario packs and confidential briefs.
  9. 09Convening plausibility, Central Asia. Uzbekistan, Tajikistan, and Kyrgyzstan sign landmark border agreement (Treaty of Khujand), 31 March 2025.
  10. 10Convening plausibility, Cyprus. UN Secretary-General to visit Cyprus for stalled peace talks, 2026.
  11. 11Convening plausibility, South China Sea. Wang Yi reiterates China's rejection of the 2016 arbitration ruling, South China Morning Post, 2025.
  12. 12Convening plausibility, Korea. Kim Jong Un's 'two hostile states' declaration and its legal implications, Daily NK.
  13. 13Convening plausibility, Jammu and Kashmir. Modi tells Trump India will not accept third-party mediation on Kashmir, BBC, 2025.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file diff --git a/static-site/print/diplomatic-simulator/index.txt b/static-site/print/diplomatic-simulator/index.txt index 37b738f4a..4b97e599b 100644 --- a/static-site/print/diplomatic-simulator/index.txt +++ b/static-site/print/diplomatic-simulator/index.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -12:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","print","diplomatic-simulator",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","diplomatic-simulator","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -14:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -16:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","print","diplomatic-simulator",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","diplomatic-simulator","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:"$Sreact.suspense" -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1c:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -29,7 +29,7 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 19:[] 10:"$W19" 11:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1e:Ta14, @page { size: 210mm 297mm; margin: 20mm 18mm; } @@ -165,6 +165,6 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 53:["$","p","13",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"The named figures are a characterisation."}]," ","Attaching the real head of government and the foreign and defence ministers to each delegation sharpens its voice but also imports the model's biases about specific, living individuals. These entries are drawn from public reporting and describe how each figure is publicly discussed; they are not claims about private conduct, and the simulator does not model any named person's actual decisions."]}] 54:["$","p","14",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"It is not decision support."}]," ","The repository states directly that the tool must not be used to inform real policy, negotiation, or intelligence judgements."]}] 1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -57:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +57:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 18:null 1d:[["$","title","0",{"children":"The Diplomatic Simulator (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L57","6",{}]] diff --git a/static-site/print/ercf/__next._full.txt b/static-site/print/ercf/__next._full.txt index 2966c90b4..02dd61633 100644 --- a/static-site/print/ercf/__next._full.txt +++ b/static-site/print/ercf/__next._full.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -12:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","print","ercf",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","ercf","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -14:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -16:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","print","ercf",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","ercf","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:"$Sreact.suspense" -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1c:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -29,7 +29,7 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 19:[] 10:"$W19" 11:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1e:Ta14, @page { size: 210mm 297mm; margin: 20mm 18mm; } @@ -184,6 +184,6 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 67:["$","tr","8",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Trauma kit"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"200 US dollars"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Unvalidated. The ICRC does not publish per-kit pricing."}]]}] 68:["$","tr","9",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Radio"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"500 US dollars each"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Within the documented procurement range for professional handheld VHF units."}]]}] 1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -6d:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +6d:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 18:null 1d:[["$","title","0",{"children":"The Evacuation Risk and Cost Framework (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L6d","6",{}]] diff --git a/static-site/print/ercf/__next._head.txt b/static-site/print/ercf/__next._head.txt index 5f053e8c6..1b973b22f 100644 --- a/static-site/print/ercf/__next._head.txt +++ b/static-site/print/ercf/__next._head.txt @@ -1,6 +1,6 @@ 1:"$Sreact.fragment" -2:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -3:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +2:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +3:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 4:"$Sreact.suspense" -5:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +5:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 0:{"rsc":["$","$1","h",{"children":[null,["$","$L2",null,{"children":[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]]}],["$","div",null,{"hidden":true,"children":["$","$L3",null,{"children":["$","$4",null,{"name":"Next.Metadata","children":[["$","title","0",{"children":"The Evacuation Risk and Cost Framework (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L5","6",{}]]}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/print/ercf/__next._index.txt b/static-site/print/ercf/__next._index.txt index f98a4fd90..18b9d82b6 100644 --- a/static-site/print/ercf/__next._index.txt +++ b/static-site/print/ercf/__next._index.txt @@ -1,13 +1,13 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -d:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +d:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} 8:["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$Ld",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$Ld",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$Ld",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$Ld",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] diff --git a/static-site/print/ercf/__next._tree.txt b/static-site/print/ercf/__next._tree.txt index 172ef7ede..54a73df99 100644 --- a/static-site/print/ercf/__next._tree.txt +++ b/static-site/print/ercf/__next._tree.txt @@ -1,7 +1,7 @@ -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] 0:{"tree":{"name":"","param":null,"prefetchHints":16,"slots":{"children":{"name":"print","param":null,"prefetchHints":0,"slots":{"children":{"name":"slug","param":{"type":"d","key":"ercf","siblings":null},"prefetchHints":0,"slots":{"children":{"name":"__PAGE__","param":null,"prefetchHints":0,"slots":null}}}}}}},"staleTime":300,"buildId":"static-snapshot"} diff --git a/static-site/print/ercf/__next.print.txt b/static-site/print/ercf/__next.print.txt index 460afe242..fa617e997 100644 --- a/static-site/print/ercf/__next.print.txt +++ b/static-site/print/ercf/__next.print.txt @@ -1,4 +1,4 @@ 1:"$Sreact.fragment" -2:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -3:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +2:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +3:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] 0:{"rsc":["$","$1","c",{"children":[null,[["$","style",null,{"children":"\n /* The site is dark-first, and a dark color-scheme makes Chrome paint\n the printed sheet's margin area dark too. Paper is light. */\n html { color-scheme: light !important; }\n html, body { background: #ffffff !important; }\n body::after { content: none !important; }\n header, footer, .site-bg, .aura { display: none !important; }\n "}],["$","$L2",null,{"parallelRouterKey":"children","template":["$","$L3",null,{}]}]]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/print/ercf/index.html b/static-site/print/ercf/index.html index f49901aa7..52dcfd883 100644 --- a/static-site/print/ercf/index.html +++ b/static-site/print/ercf/index.html @@ -1,4 +1,4 @@ -The Evacuation Risk and Cost Framework (print edition) · NYU Ethical Tech CoLab

Publications · Academic report

The Evacuation Risk and Cost Framework

Estimating the Human and Financial Cost of Civilian Evacuation in Armed Conflict

Ethical Tech CoLab

NYU Center for Global Affairs

July 2026

Yago Rocha. Developed as part of research on International Humanitarian Law and civilian protection at the NYU Center for Global Affairs, under the Ethical Tech CoLab. The repository records the work as presented to Teresa Cantero, PhD Researcher, whose suggestions are cited in the project backlog.

A humanitarian organisation facing a besieged city has to answer a question that no manual answers well: what would it actually take to move these people, and what would it cost to keep them alive if they cannot be moved? Both questions have to be answered before a funding appeal can be written, and both are usually answered from memory, from a previous operation in a different country, or from an experienced coordinator's intuition. ERCF is a research prototype that attempts to put numbers on both and to show its working.

Key figures

7

weighted dimensions a planner scores from 1 to 5, with kinetic threat carrying a quarter of the total

31

documented historical conflicts in the corpus, sixteen of which were used to fit the mortality model

0.855

R-squared in logarithmic space, against only 44 per cent of fitted cases landing within a factor of two

90

days of planning window, beyond which the model sits outside the range of its own calibration data

01

Executive Summary

ERCF is a research prototype: experimental software, not an operational system. It helps a humanitarian planner estimate two things side by side. The first is the one-time cost of organising an evacuation. The second is the daily cost of keeping people supplied if they stay. It then compares the two.

The planner describes the situation using seven factors, called dimensions and labelled D1 to D7. Each is scored from 1 to 5. They cover the intensity of fighting, the mobility of the population, whether the armed parties have authorised movement, the state of roads and logistics, the safety of the destination, how fast the window is closing, and how good the available information is. These seven scores drive everything else the tool produces.

The seven scores are combined into a single weighted number, which is placed in one of five risk levels from Level 0 to Level 4. That level then selects the cost rates, loss rates and mortality rates used in the rest of the calculation. A planner who moves one slider sees the entire cost estimate change.

The distinctive output is the break-even analysis. Evacuation is a large one-time expense. Assistance in place is a smaller expense repeated every day. Somewhere in the future, if the crisis persists, the cumulative daily cost overtakes the one-time cost. ERCF calculates the day on which that happens, which is a question funding appeals have to answer and rarely answer with arithmetic.

ERCF also produces an estimate of deaths and injuries among a population that does not leave. This estimate was fitted against sixteen documented historical conflicts. Its author is explicit that this part of the tool is indicative and that the financial estimates are considerably more reliable than the mortality ones.

The project takes an explicit ethical position, stated in its own documentation and repeated in the interface. It estimates the cost of logistics. It does not place a monetary value on human life, and it does not tell a user whether an evacuation is worth carrying out. The labels it produces are descriptive, not instructions.

The software is substantially built rather than aspirational. The repository contains roughly eleven thousand lines of working Python, a browser interface, a reproducible calibration pipeline, a database of thirty-one documented historical conflicts, and connections to four external data services. It runs as a live public demonstration. What remains incomplete is documented openly in the project's own backlog.

02

Background and Rationale

An organisation deciding whether to support an evacuation needs three figures quickly: what evacuation would cost, what staying would cost, and where those two lines cross. Each figure exists somewhere. Sphere standards give water and food quantities per person. The World Food Programme publishes what its air service costs per passenger kilometre. Case studies record what particular operations spent. But these sit in separate documents written for separate purposes, and nobody has assembled them into a single calculation that a planner can run in an afternoon.

The project's concept document states the position directly: these estimates exist in operational manuals and case studies, but no publicly available tool integrates them with a historical calibration dataset and a framework grounded in International Humanitarian Law. ERCF is an attempt to build that integration and to publish the assumptions rather than hide them.

One finding recorded in the source code is worth restating for a policy audience. When the author searched for a published figure describing how much more expensive it is to deliver humanitarian assistance inside an active conflict than in a stable setting, no such figure was found. The International Federation of Red Cross and Red Crescent Societies confirms in a 2021 methodology paper that the per-capita cost of response is statistically significantly higher in conflict, but publishes no ratio. The World Food Programme publishes per-tonne logistics costs that imply a premium of roughly 1.23 times for Sudan in 2023 against an Eastern Africa regional average. Security costs, which the model treats as the larger component, appear only in fragments: an armed escort in Yemen at up to 3,000 US dollars per convoy on one route, security spending in Somalia in 2017 estimated at around 400 million US dollars per year against total humanitarian operations of 700 to 900 million. The code notes that a single published multiplier for active-conflict delivery does not appear to exist, and that this model may be among the first attempts to quantify it explicitly. That is an honest statement of a weakness and of a contribution at the same time.

ERCF builds the calculation anyway, using the best documented values available, and tags every single number with the strength of the evidence behind it. A planner reading the interface can see which figures are validated against a published standard, which are estimated from indirect evidence, and which are unvalidated assumptions the author could not support from any public source. That tagging is the tool's principal safeguard against being trusted more than it deserves.

03

Objectives

The tool is designed to:

  • Produce a rapid, itemised cost estimate for a civilian evacuation operation, broken into transport, fuel, personnel, food, water, shelter, medical supplies, communications and contingency.
  • Produce a parallel estimate for the cost of assisting a population that remains in the conflict zone, including the extra cost that conflict access conditions impose and the proportion of supplies that never arrives.
  • Identify the break-even point at which continuing in-zone assistance becomes more expensive than a one-time evacuation, so that funding appeals can be argued from arithmetic rather than assertion.
  • Place the current situation against documented historical operations, so that a planner can see which past conflicts a scenario most resembles.
  • Make every parameter, source and confidence level visible in the interface rather than buried in the code.
  • Keep the estimate of financial cost separate from the estimate of mortality, so that a funding calculation is never confused with a casualty prediction.

04

How ERCF Works: The Seven Dimensions

The whole tool rests on seven scores. Each is a number from 1 to 5 that the planner sets by moving a slider. The dimensions were chosen, according to the concept document, to capture the factors that field coordinators consistently cite when assessing whether an evacuation is feasible, and deliberately to require no real-time data feed or classified intelligence. A planner with a situation report and professional judgment can set all seven.

D1, Kinetic Threat. How much direct violence the civilian population is exposed to. A score of 5 means active, sustained attack. This is the dimension that drives almost everything downstream: it determines whether movement is physically survivable at all.

D2, Mobility Constraints. Also called Vulnerability. How much of the population cannot move unaided. A high score means many people who are elderly, very young, disabled, chronically ill or otherwise dependent on assisted transport. This dimension is what forces the operation to allocate medical buses and ambulances rather than ordinary buses.

D3, Authorization. Whether the armed parties have consented to civilian movement. The scale runs the same direction as the others: a high D3 score means authorization is a serious problem, that is, that consent is absent or unreliable. Without consent, an evacuation is both unlawful and practically blocked.

D4, Logistics. The state of roads, bridges, vehicles, fuel supply and the supporting infrastructure. The project's documentation observes that logistics collapse, not danger as such, has empirically been the leading cause of delayed evacuations, citing Mosul in 2016 and Aleppo in 2016.

D5, Destination. Whether the place people would be moved to is genuinely safe and able to receive them. The reason this dimension exists at all is Srebrenica in 1995, cited in the code as the canonical case in which evacuation to a nominally safe area became the site of a massacre. Moving civilians into danger is a distinct harm, not a lesser version of leaving them where they are.

D6, Urgency. How fast the window for organised movement is closing.

D7, Information. How poor the information environment is. A high score means communications failure, rumour, and an inability to coordinate.

The seven scores are combined into a single number by multiplying each by a weight and adding the results. The weights add up to 1.00, so the resulting score stays on the same 1 to 5 scale as the inputs.

DimensionFactorWeight
D1Kinetic Threat0.25
D2Mobility Constraints0.15
D3Authorization0.15
D4Logistics0.15
D5Destination0.15
D6Urgency0.10
D7Information0.05

ERCF = (D1 x 0.25) + (D2 x 0.15) + (D3 x 0.15) + (D4 x 0.15) + (D5 x 0.15) + (D6 x 0.10) + (D7 x 0.05)

The seven weights sum to 1.00.

D1 carries the largest weight, one quarter of the total, because direct physical threat is treated as the primary driver of evacuation necessity, and because it is the dimension that alone determines whether movement is safe. It is tied to the precautionary obligations in Articles 57 and 58 of Additional Protocol I.

D2, D3, D4 and D5 each carry fifteen per cent and form an equal group. The argument is that mobility, consent, logistics and destination safety are each individually capable of making an evacuation impossible, so none should dominate the others. D3 is explicitly capped at fifteen per cent rather than being set higher, because the most extreme authorization failures are already captured elsewhere in the tool by a hard trigger.

D6 carries ten per cent, deliberately below the equal group. The reasoning is that urgency is largely absorbed by D1 in the most extreme situations, and that the tool captures urgency more sharply through the hard trigger and through a separate extraction-probability floor than a linear weight would. D7 carries five per cent, the lowest. Poor information raises coordination cost and panic risk, but on its own it does not determine whether an evacuation is necessary or possible.

An important caveat is stated by the author in the code itself and reproduced in the interface as a red marker beside every weight: all seven weights are modelled estimates. No published framework in International Humanitarian Law or humanitarian operations specifies numerical weights for these factors, so there was nothing to copy. The author's own backlog lists expert-panel validation of these weights as outstanding research work. A reader should treat the weights as a reasoned starting proposal, not as an established standard.

05

From Seven Scores to a Risk Level

The weighted score is placed into one of five bands, each with a label and a NATO-doctrine equivalent used in military planning.

LevelLabelWeighted scoreNATO equivalent
Level 0Baseline and Monitoring1.5 or belowpermissive and stable
Level 1Low Risk and Advisoryabove 1.5 up to 2.5permissive but degrading
Level 2Moderate Risk and Watchfulabove 2.5 up to 3.5uncertain
Level 3High Risk and Contestedabove 3.5 up to 4.2hostile, partial
Level 4Critical and Emergencyabove 4.2hostile, imminent

There is one override. If D1 and D6 are both at 4.5 or above, that is, if violence is extreme and the window is closing at the same time, the score is floored at 4.21, which forces Level 4 regardless of how favourable the other five dimensions look. This exists because the linear weighted average would otherwise allow good logistics and a safe destination to pull an imminent massacre down into Level 3.

The level is not merely a label. It selects the numerical rates used everywhere downstream: the security escort ratio, the daily cost of assistance per person, the access multiplier, the supply loss rate, the injury rate, the mortality base rate, and the probability of emergency extraction. Almost every figure the tool produces changes when the level changes. This is a design choice worth noticing, because it means the five-band classification carries a great deal of weight and the boundaries between bands, at 1.5, 2.5, 3.5 and 4.2, are themselves modelled judgments rather than empirical findings.

Alongside the composite score, the tool computes three sub-indexes that keep distinct questions apart. This was added on the recommendation of the project's academic reviewers. Risk Severity asks how dangerous the situation is for civilians, and is built from D1, D2 and D6 only, rescaled back onto a 1 to 5 range. Feasibility asks whether people can realistically move, and is built from D3, D4 and D5, but inverted: because a high D3, D4 or D5 score means bad conditions, each is subtracted from 6 before being used, so that a high feasibility number means a genuinely open corridor. Information Quality is simply 6 minus D7, so that a high number means good situational awareness.

Keeping severity and feasibility apart matters. A situation can be extremely dangerous and simultaneously impossible to evacuate. Blending the two into one score would make that case look moderate, which is precisely the case that should escalate most urgently. The tool instead places the scenario in a four-cell matrix. High severity with high feasibility returns evacuate immediately. High severity with low feasibility returns shelter in place and negotiate a corridor urgently, with the reasoning that forced movement without a safe corridor risks greater harm than staying, and with a note that the precautionary obligations under Articles 57 and 58 of Additional Protocol I apply while negotiation continues. Low severity with high feasibility returns facilitate voluntary departure, and explicitly says do not mandate evacuation. Low severity with low feasibility returns monitor and reassess daily.

06

The Cost of Evacuating: Every Variable Explained

This is the part of the tool with the strongest evidentiary foundation. The planner supplies four things: the population at risk, the percentage of that population who are vulnerable, the distance in kilometres to safety, and the terrain quality. Everything else is derived.

The population is split into vulnerable and non-vulnerable. Non-vulnerable people are assigned to standard buses at 50 people per bus. Vulnerable people are assigned to medical buses at 20 people per bus, and ambulances at one per 150 vulnerable people. The bus capacities are marked in the code as operational assumptions not validated against field data. The ambulance ratio has a history worth recording: it was originally set at one ambulance per 40 vulnerable people, and was revised to one per 150 after the author found that no published field standard exists at all and that the original figure was three to five times above what documented practice, in particular a study of the Kosovo operation, actually showed.

The number of medical buses and ambulances is then multiplied by a factor derived from D2, the mobility dimension. At D2 of 1 the factor is 0.8, at 2 it is 1.0, at 3 it is 1.3, at 4 it is 1.8, and at 5 it is 2.5. In plain terms, a population with severe mobility constraints needs roughly two and a half times the assisted transport of a baseline population. This factor is described in the code as estimated with no primary source.

Staffing is set by ratio, and the security ratio tightens sharply as danger rises. The medical staffing figure is the Sphere Handbook 2018 standard for clinical officers in emergency settings, and it was corrected upward during development: an earlier version used one per 500, half the Sphere standard, with no documented justification.

RoleAllocationDaily rate, US dollars
Securitynone at Level 0, one per 500 civilians at Level 1, per 200 at Level 2, per 100 at Level 3, per 50 at Level 4300
Medical staffone per 250 people200
Paramedicsone per 100 people150
Driversone per vehicle50
Daily rates are total cost to the operation, not take-home pay. All four are tagged as estimated.

The code carries an unusually candid note about the rates: every rate implicitly assumes national staff or a lower-cost international non-governmental deployment at roughly the level of Médecins Sans Frontières, and full United Nations international professional staff, once daily subsistence allowance and danger pay are included, would cost three to six times more. The author records that this assumption is not stated in the interface and should be flagged.

Water is calculated at 20 litres per person per day for three days. This is the UNHCR full planning standard, chosen deliberately over the Sphere emergency minimum of 7.5 to 15 litres because an evacuation is a planned operation rather than a first-response emergency. Food is calculated at 0.45 kilograms of dry food per person per day for three days, which is the dry-weight equivalent of the Sphere minimum of 2,100 kilocalories per person per day. Tents are provided at one per five people, which at the Sphere standard of 3.5 square metres per person gives 17.5 square metres per tent and is internally consistent. Basic medical kits are provided at one per 100 people, and trauma kits at one per 50 people when the risk level is 3 or above, or one per 200 otherwise. Radios are provided at one per five vehicles plus a fixed five for coordination.

The unit costs applied to those quantities are set out below, each with the evidence the project records behind it. Fuel is the one line with a consumption figure of its own: it is calculated at 0.35 litres per kilometre per vehicle for a return journey before the per-litre price is applied.

ItemUnit costEvidence recorded in the code
Standard bus200 US dollars each
Medical bus400 US dollars eachRevised upward after research found no primary source for the earlier value; sits at the lower bound of documented field ranges for medically-equipped vehicles.
Ambulance700 US dollars eachRevised upward on the same basis as the medical bus.
Fuel1.20 US dollars per litreRevised down from 1.50 on the basis of ACAPS reporting of Yemeni consumer fuel prices in 2022.
Food3 US dollars per kilogramFlagged by the project's own parameter registry as having no citation behind it, even though the quantity does.
Water0.05 US dollars per litreField evidence gathered in 2026 found real water trucking costs of 2 to 23 US dollars per cubic metre, all below the model's baseline; the lower figures were not adopted pending further validation.
Tent380 US dollars eachRaised from an earlier 150 once that figure was found to describe a tarpaulin kit suitable for a week rather than a standard tent; 380 sits just below the 400 dollar replacement cost UNHCR gave publicly in 2022.
Basic medical kit21 US dollarsDerived from the WHO and UNICEF Interagency Emergency Health Kit costing, revised down from an earlier 50 that was roughly seven times too high for a three-day convoy.
Trauma kit200 US dollarsUnvalidated. The ICRC does not publish per-kit pricing.
Radio500 US dollars eachWithin the documented procurement range for professional handheld VHF units.

Three multipliers are applied on top. Terrain multiplies transport and fuel costs, with five levels running from 4.0 for the worst terrain down to 1.0 for the best, and 2.5, 1.7 and 1.2 in between. The lower end of this range is consistent with published road-condition cost models; the upper end of 4.0 is an expert estimate consistent with a World Food Programme figure showing per-tonne delivery costs in the Central African Republic, South Sudan and the Democratic Republic of the Congo running about five times a standard country average, though that figure mixes terrain, access and security together.

Season adjusts terrain further. If the operation starts in a month the tool classifies as a closure period for that latitude, the terrain multiplier is boosted by 50 per cent for the worst terrain, 30 per cent for the second worst and 20 per cent for middling terrain. Closure periods are set by latitude: December to March above 30 degrees north, June to September below 30 degrees south, and April to October in the tropics for wet season. The author describes these as broad regional approximations. The most useful output here is not the cost boost but a flag: the worst terrain in a closure period is marked potentially impassable, which is an operational warning rather than a number.

D4, logistics, adds 10 per cent to transport and fuel for each point above 1, so a D4 of 5 adds 40 per cent. D5, destination, changes the number of tents needed: a destination with good existing infrastructure halves the tent requirement, a destination with none doubles it. Both are marked estimated with no primary source. Finally, 15 per cent is added to the whole subtotal as contingency, which the code justifies as the lower bound of the 15 to 20 per cent standard used for high-risk projects.

A worked example is carried in the code. For 10,000 people, 20 per cent of them vulnerable, at Level 2, moving 50 kilometres: 160 standard buses, 100 medical buses and 50 ambulances, 310 vehicles in total; 50 security staff, 20 medical staff, 100 paramedics and 310 drivers; 10,850 litres of fuel, 15,000 kilograms of food, 450,000 litres of water, 2,000 tents and 67 radios. Transport comes to 92,500 US dollars, fuel to 16,275, personnel to 34,000, food to 45,000, water to 22,500, shelter to 300,000, medical to 15,000 and communications to 33,500, giving a subtotal of 558,775, a contingency of 83,816, and a total of roughly 643,000 US dollars, or about 64 dollars per person. Shelter dominates, which is a useful thing for a planner to be able to see at a glance.

07

The Cost of Staying

The second calculation estimates what it costs an organisation to keep a population alive where it is. It has three parts.

The first is supply delivery. The starting point is 3.50 US dollars per person per day, a multi-sector figure covering food, water, health, shelter and coordination. The World Food Programme's own global average for food and cash assistance alone was 42 US cents per beneficiary per day in 2023, and the code explains the difference between the two figures at some length: at the full Sphere water standard, water alone dominates supply weight, and applying the World Food Programme's per-tonne Sudan delivery cost to that weight gives 3.28 US dollars per person per day, within six per cent of the model's figure. That is a genuinely careful piece of reasoning.

The baseline is then multiplied by an access multiplier that represents how much more expensive delivery becomes as conflict intensifies. The second part of the calculation, supplies that never arrive, is then added as a loss rate representing goods destroyed, looted or simply undelivered. Both ladders are selected by the risk level.

Risk levelAccess multiplierSupply loss rate
Level 01.05 per cent
Level 11.55 per cent
Level 22.015 per cent
Level 33.630 per cent
Level 44.050 per cent

The access multipliers were revised downward during development, from earlier values of 3.0, 5.0 and 8.0 for the top three levels, after the author's research established that documented sources support only about 2.5 to 3.6 times at Level 3 and nothing above about 4 times at Level 4. Level 4 remains marked as directionally plausible but unconfirmed, and the project's backlog lists it as the outstanding unvalidated cost parameter. A further penalty is applied from D4, and the terrain and seasonal multipliers apply here too, because bad roads make delivery expensive whether people are leaving or staying.

Only the lowest of the loss rates has any published anchor: OCHA monitoring of Gaza during the ceasefire period in late 2025 recorded under two per cent of cargo looted or intercepted under active monitoring, and the model uses 5 per cent as a planning buffer including spoilage. The three higher figures are internal planning estimates with no published equivalent that the author could find. A low D3, meaning consent is absent, adds further to the loss rate, and a low D7, meaning poor information, adds a coordination overhead.

The third part is emergency extraction. The model assumes that some proportion of the population will have to be pulled out in an emergency at some point, and prices that in advance. The probability of this happening rises over time along a saturating curve: it starts near zero and approaches a ceiling. The daily rate driving that curve was fitted against historical cases and is 0.021 for Level 4, 0.010 for Level 3, 0.005 for Level 2 and 0.002 for Level 1. The Level 4 figure comes from averaging Mariupol in 2022 and Aleppo in 2016; the Level 3 figure from Mosul, Goma and the Central African Republic. The Level 2 and Level 1 figures have no historical anchor at all and are described as structurally plausible interpolations that require validation.

The curve is capped at 95 per cent for Level 4, 80 for Level 3, 60 for Level 2 and 30 for Level 1. Two modifiers then apply: a blocked corridor, meaning a low D3, raises the probability, and high urgency imposes a floor, 85 per cent when D6 is 5 and 60 per cent when D6 is 4. The floor exists because of Srebrenica, where the crisis unfolded over three days and no duration-based curve would have registered it.

The cost of extraction is anchored to the United Nations Humanitarian Air Service, which is the World Food Programme's air service flying humanitarian staff and light cargo into places commercial aviation does not serve. Its published operating cost was 2.08 US dollars per passenger kilometre in 2023. Ground extraction is priced at 30 per cent of that rate, which the code describes as an internal heuristic with no published source. Air medical evacuation is priced at three times the rate, on the reasoning that a medical flight carries far fewer passengers. At Level 4 the whole extraction cost is multiplied by 2.5 for a helicopter premium, which the author records as unconfirmed, since UNHAS does not publish separate helicopter and fixed-wing rates.

Field medical treatment is costed separately. Injuries are estimated from a per-thousand daily rate that rises with the risk level: zero at Level 0, 0.1 at Level 1, 0.5 at Level 2, 2.0 at Level 3 and 8.0 at Level 4. The code states plainly that no public source reports in-field civilian injury rates as a daily incidence per thousand, so these are unvalidated. Each injury is costed at 800 US dollars, modified upward when the destination has poor medical capacity. The 800 figure sits within a documented peer-reviewed range: 211 US dollars per surgical case at a conflict-affected mission hospital in South Sudan in 2022, and roughly 500 to 650 dollars per case at two MSF surgical trauma centres around 2009, which inflate to roughly 780 and 1,010 dollars by 2026.

08

The Break-Even Analysis

This is the calculation the tool is built around. Two options are compared. Option A is to evacuate now and then provide assistance to whoever remains behind. Option B is not to evacuate and to assist the whole population in place. Option A carries a large one-time cost plus a small daily cost. Option B carries only a daily cost, but a larger one.

The break-even day is the one-time evacuation cost divided by the difference between the two daily costs. If evacuating costs 640,000 US dollars and staying costs 4,000 US dollars a day more than the post-evacuation arrangement, the break-even is 160 days. Before that day, staying is cheaper. After it, evacuation would have been cheaper.

It is a resource-planning number. The project's concept document is emphatic that it is not an answer to whether an evacuation is worth carrying out. The distinction is that the calculation compares the cost of one logistics operation against the cost of another logistics operation. It does not weigh either against lives. The decision to move civilians, the documentation states, is a matter of legal obligation and humanitarian judgment, not financial optimisation.

The tool also checks the transport mode against the dimension scores and warns when they are inconsistent. Five patterns trigger a warning: ground transport with a D4 of 4.0 or above, which indicates logistics breakdown; ground transport with a D1 of 4.5 or above, which indicates unacceptable kinetic exposure; walking with a population above 5,000, which is infeasible at scale; walking with a D2 of 3.5 or above, which means the population cannot walk; and air transport with a D3 of 2.0 or below, which means airspace authorization is missing. These are useful precisely because a cost model will happily produce a plausible-looking figure for an operation that could never be attempted.

09

Estimating Deaths: The Mortality Model

Separately from cost, the tool estimates how many people would die and be injured if the population stayed. The author's framing is that this provides scale context for a planning decision, and both the README and the concept document state that the mortality model is indicative rather than predictive and that the financial estimates are substantially more reliable.

Each risk level has a baseline death rate expressed per 10,000 people per day. The current values are 0.777 at Level 0, 0.964 at Level 1, 3.625 at Level 2, 1.805 at Level 3 and 1.000 at Level 4. A reader will immediately notice that these are not in ascending order, which looks wrong and requires explanation. The explanation given in the code is that these are not standalone death rates but base rates that get multiplied by three further factors, and that Level 3 in the historical corpus is dominated by urban sieges and city fighting where a high proportion of the population is directly exposed, while Level 4 is dominated by large-enclave operations across much bigger populations where per-capita exposure is lower. The ordering was produced by the fitting procedure rather than assumed, and the author flags it explicitly as empirically validated but counterintuitive.

The confinement multiplier captures whether people are trapped. It is calculated as 5 minus D3, multiplied by D4, divided by 5. In plain words: when consent for movement is absent and logistics have collapsed at the same time, people cannot get out, and mortality rises sharply. The resulting score is converted into a multiplier in steps: half at the lowest, then 1, then 2, then 4, then 8 at the highest. An eightfold difference from a single factor is a very large lever, and its stepwise rather than smooth form is listed by the author as a known limitation. The anchors given are Aleppo and Kosovo at a multiplier of 2, Mosul at 1, and the Kherson evacuation, which had an open corridor, at 0.5.

confinement score = ((5 - D3) x D4) / 5

The score is then mapped to a multiplier in steps: 0.5, 1, 2, 4, 8.

The displacement protection factor applies where part of the population has already left, since the remaining exposure is lower. The tool reduces the death rate by 60 per cent of the share that has departed. The 60 per cent coefficient, rather than 100, reflects that displaced people still face risk on the road, at checkpoints and from exposure. There is then an important refinement: in a siege, defined as a D3 of 2 or below combined with a D1 of 4 or above and a population of 500,000 or fewer, the coefficient is halved to 30 per cent. The reasoning is that in an encircled city, movement itself is lethal, because civilians pass checkpoints under fire and buses have been attacked. Displacement is still safer than staying, but only half as much safer as in an open corridor.

The geographic exposure factor recognises that not everyone in a conflict area is under fire at the same time. Four conflict shapes are recognised, with the fraction of the population treated as simultaneously exposed given in brackets: urban siege such as Mariupol or Aleppo (0.85), enclave such as Gaza (0.65), city conflict where a front line moves through a city, such as Mosul or Kherson (0.40), and regional dispersed conflict such as Sudan, the Central African Republic or the eastern Democratic Republic of the Congo (0.12). When the planner has not specified a shape, the tool infers one from D1 and the population size, on the logic that higher kinetic threat means more direct fire while a larger population means more dispersion. The population term uses a logarithm raised to the power 1.4, a refinement introduced specifically because the earlier formula did not fall away fast enough for continental-scale conflicts and overestimated Sudan by a wide margin.

Cumulative deaths accumulate linearly for the first 90 days and then decelerate along a square-root curve, on the reasoning that populations adapt and survivors relocate. This also marks the outer edge of the model's intended planning window. Injuries are set at four times deaths, following the ICRC planning ratio. The code notes that a peer-reviewed systematic review implies a ratio closer to 3.3 to 1, and that frontline-specific figures from Ukraine run far higher, and retains 4 to 1 as a mid-range planning estimate.

The infrastructure-denial multiplier applies only where there is primary source documentation that survival infrastructure was deliberately destroyed. It is calculated as 1 plus 0.4251 times D1 minus 3, times D4 minus 3, and only when D1 is at least 4.5 and D4 at least 4.0. It is switched on for four documented cases: Mariupol, Aleppo, Vukovar and Huambo, giving effective multipliers between roughly 1.6 and 2.3. The supporting documentation cited is a 2024 report on starvation as a method of warfare in Mariupol, the UN Commission of Inquiry on Syria for hospital bombing in Aleppo, ICTY proceedings for Vukovar, and Human Rights Watch and Amnesty reporting on Angola for Huambo. Critically, this multiplier is switched off by default for live scenarios. The code records why: an earlier version activated it automatically whenever the dimension thresholds were met, and calibration accuracy collapsed from 80 per cent to 20 per cent. It is a finding about deliberate atrocity, not a threshold that can be inferred from slider positions.

infrastructure-denial multiplier = 1 + (0.4251 x (D1 - 3) x (D4 - 3))

Applied only where D1 is at least 4.5 and D4 at least 4.0, and only with primary source documentation. Off by default for live scenarios.

10

How the Model Was Fitted, and Against What

The repository contains thirty-one documented conflicts. Each record holds the population at risk, duration in days, recorded deaths, numbers displaced and remaining, the estimated vulnerable percentage with its own source and confidence rating, the distance to safety with its source, all seven dimension scores assigned retrospectively, corridor status and notes, key lessons, IHL issues, and a calibration block recording what the model predicted against what was recorded, across every model version.

Sixteen cases are used to fit the model: Mariupol 2022, Gaza 2023, Aleppo 2016, the Kherson evacuation 2022, Mosul 2016, eastern DRC and Goma 2024, the second battle of Grozny 1999, both battles of Fallujah 2004, the Jenin refugee camp 2002, the siege of Vukovar 1991, Gaza Cast Lead 2008, Gaza Protective Edge 2014, the siege of West Beirut 1982, the second Nagorno-Karabakh war 2020, and the siege of Huambo 1993.

Fifteen are excluded, and the reasons form a clear statement of where the model does not apply. Duration beyond the 90-day window excludes Sarajevo at 1,425 days, Kuito in Angola at over 540, Marawi at 154 and the final phase of the Sri Lankan Vanni offensive at 120. Dispersed regional conflict with no siege perimeter excludes the Central African Republic in 2014 and Sudan in 2023, the latter also on duration and on a population of six million. Deliberate massacre as the dominant mechanism of death, rather than attrition under fire, excludes Srebrenica 1995, Hue 1968, Manila 1945 and Bucha 2022. Open-corridor forced displacement, where civilians were made to move rather than prevented from moving, excludes Kosovo 1999. Recorded death counts too uncertain to fit against exclude Misrata 2011, where the contested range runs from 102 to 700.

Two exclusions are of a different character and the repository says so openly. Eastern Ghouta 2018 and Gaza Pillar of Defense 2012 are labelled challenge cases and were excluded, in the words of the data file, to preserve the version 7 metrics. They document a real structural limit: the Level 4 base rate cannot simultaneously fit a high-mortality urban siege such as Aleppo, with 31,000 deaths, and a large-enclave precision operation such as Cast Lead, with 965. The model undercounts one and overcounts the other by a wide margin. Retaining them in the corpus while removing them from the fit is a defensible research decision, and the fact that it is documented rather than concealed is to the project's credit. But a reader should understand that the headline accuracy figures are calculated on a set from which two known-difficult cases were removed.

The fitting procedure adjusts six numbers: the five base death rates and the infrastructure-denial coefficient. It searches for the values that minimise the mean squared error between the logarithm of predicted deaths and the logarithm of recorded deaths. Working in logarithms means the procedure fits proportional rather than absolute error, which is appropriate when the recorded death counts in the corpus run from 37 to 45,000. The search is constrained so that the rates for Levels 0, 1 and 2 rise in order; Levels 3 and 4 are deliberately left unconstrained relative to each other, which is how the counterintuitive ordering emerged.

On the sixteen fitted cases, the correlation between predicted and recorded deaths in logarithmic space gives an R-squared of 0.855, meaning the model captures most of the variation in order of magnitude. Seven of the sixteen cases, or 44 per cent, fall within a factor of two of the recorded figure. Those two statements sit oddly together and both should be kept in view: the model is good at ranking conflicts by scale of loss and much less good at predicting the number for any individual conflict. A battery of statistical checks reported in the README covers normality of residuals, constant variance, rank correlation and multicollinearity, and reports Aleppo as the single most influential case.

One qualification about the cross-validation figure is worth making. The reported leave-one-out R-squared of 0.807 is not produced by refitting the whole model with each case removed in turn. It fits a straight line through the remaining points in logarithmic space and predicts the held-out point from that line. It therefore measures how stable the relationship between predicted and recorded deaths is across the corpus, which is useful, rather than how well fitted parameters generalise to unseen conflicts, which is what the term usually implies.

11

Where the Data Comes From

The Uppsala Conflict Data Program, based at Uppsala University in Sweden, is the principal academic provider of data on organised violence worldwide. Its Georeferenced Event Dataset, version 26.1, records individual lethal events down to the level of a single village on a single day, with separate counts for civilian deaths and for each party, and low, best and high estimates reflecting source uncertainty. ERCF uses it to check its historical cases against an independent record and to let a planner query recorded violence for a location and date range. The tool queries the live service where a token is available and falls back to a locally stored copy of the dataset otherwise.

ACAPS is an independent humanitarian analysis organisation. Together with the European Commission's Joint Research Centre it publishes the INFORM Severity Index, a monthly assessment scoring the severity of humanitarian crises across more than thirty indicators grouped into the impact of the crisis, the conditions of the people affected, and the complexity of the response. ERCF uses these scores to colour a world map by country-level severity and to convert a national severity score onto its own five-level scale. Two things should be noted. The repository works from a stored table of roughly eighty countries with a live overlay attempted on top, and the code records that the two relevant ACAPS endpoints were consistently timing out at the time of writing, which means in practice the stored table is usually what drives the map. Users should treat the map as an orientation aid rather than a live feed.

GeoNames provides city population figures so a planner can start from a real place rather than a guess. The project's own backlog flags the obvious limitation: these are pre-conflict populations and do not reflect displacement that has already occurred, so a user must adjust them manually.

Open-Meteo provides historical weather from the European reanalysis archive. ERCF uses fourteen days of temperature, precipitation, snowfall and wind to classify the operating conditions and to set two multipliers: one on fuel and transport, one on shelter. Deep winter raises fuel costs by 35 per cent and shelter costs by 120 per cent, the latter reflecting that cold-climate shelter requires more space per person and winterised units cost several times a standard tent. A heavy rainy season raises fuel and transport by 80 per cent, reflecting documented road impassability. If the planned date is in the future, the tool steps back a year at a time to find comparable historical conditions and labels the result a planning estimate. The weather data is tagged as validated; the multipliers derived from it are tagged as estimated.

A language model is used for one narrow purpose: generating a written country briefing that summarises the crisis, access conditions, likely exit routes, neighbouring safe areas, active humanitarian actors and the applicable legal framework, and proposes starting values for the seven dimensions. If the service is unavailable the tool falls back to fixed text built from the stored severity table. This is a convenience feature for a first draft of a scenario. Nothing in the cost or mortality calculation depends on it.

For eighteen conflict-affected countries the tool suggests a vulnerable population percentage. The figure is calculated as the percentage of children under five, plus the percentage of adults over sixty, plus half the percentage of people with disabilities. The halving is to avoid double-counting people who are both elderly and disabled, and to reflect that only some disabilities create evacuation-relevant constraints. The underlying figures come from UNICEF's State of the World's Children 2023 and the UN World Population Prospects 2022, with a uniform 15 per cent disability figure drawn from the WHO World Report on Disability. The results range from 21.5 per cent for Syria, Iraq and Libya to 36.5 per cent for Ukraine, where the driver is an unusually large elderly population. Two caveats matter: the child component counts only children under five rather than all minors, and the disability figure is a global average applied uniformly, whereas conflict-affected populations are known to have higher rates. The suggestion is advisory and the planner confirms or dismisses it.

Two supplementary transport models exist. Air evacuation prices fixed-wing movement at the same UNHAS rate of 2.08 US dollars per passenger kilometre and helicopter movement from a United Nations procurement contract for a Mi-8 airframe divided by an assumed number of guaranteed flight hours, returning a low, mid and high band rather than a single figure, with a deliberately conservative 50 per cent load factor. It plans fleet size across one, three, seven, fourteen and thirty day windows. Walking evacuation interpolates group speed between 8 kilometres per hour for healthy adults and 3 kilometres per hour for a fully vulnerable group, the latter taken from published gait-speed data for adults over 65, and assumes 8 walking hours a day rather than the 15 used in the academic migration model it draws on. It contains no attrition or mortality term at all, reports only the number of days of exposure, and states explicitly that the exposure does not feed back into the danger or urgency dimensions: that remains the analyst's judgment. Both models are marked as less validated than the ground transport model, and the backlog records that walking is not yet wired into the break-even calculation.

12

Reading the Results

The interface is a single web page with several sections. The Scenario Builder holds the seven sliders and the population, vulnerability, distance and terrain inputs, and updates every figure as the sliders move. A cost breakdown itemises the evacuation estimate line by line. A decision analysis panel shows the break-even chart with the two cost lines crossing. Historical Cases lists all thirty-one conflicts with badges marking which are out of scope and which are challenge cases. Compare on Radar overlays the current scenario's seven scores against any historical case as a seven-pointed shape, alongside a table comparing population, cost, cost per evacuee, duration and deaths. Map View colours the world by severity. A references section lists the data sources, and a methodology panel documents the statistical validation.

Experienced coordinators reason by analogy, saying that a situation resembles Mosul or resembles Mariupol. The radar overlay makes that comparison explicit and checkable, by showing exactly which of the seven dimensions match and which do not. That is a modest but genuine contribution: it turns an intuition into something a colleague can disagree with.

Three real operations are provided for cost benchmarking: the Kosovo Humanitarian Evacuation Programme in 1999, which moved 60,549 people at roughly 562 US dollars each; Libya in 2011 at roughly 984 dollars; and the evacuation of about 15,000 people from Lebanon to Canada in 2006 at several thousand dollars each. The README is careful to explain why these exceed the model's own figures by a wide margin: ERCF prices the immediate field operation only, at roughly 134 US dollars per person for a Level 2 ground movement over 50 kilometres, while real operations include international transport, reception facilities and weeks of hosting. A user comparing the tool's output to a real operation without reading that note would draw the wrong conclusion.

13

Grounding in International Humanitarian Law

Each risk level is tied to a stated legal reference. At Level 1 the tool cites the right of voluntary departure under Article 35 of the Fourth Geneva Convention. At Level 2 it cites the precautionary obligations in Articles 57 and 58 of Additional Protocol I, which require parties to take feasible precautions to spare civilians, including effective advance warning, and to keep civilians away from military objectives. At Level 3 it cites Article 49 of the Fourth Geneva Convention. At Level 4 it cites the same article's prohibition on forcible transfer and the corresponding provision of the Rome Statute of the International Criminal Court.

Article 49 is the central provision and deserves setting out. It prohibits individual or mass forcible transfers and deportations of protected persons from occupied territory, regardless of motive. It then creates a narrow exception: an Occupying Power may undertake total or partial evacuation of a given area if the security of the population or imperative military reasons so demand. Such evacuations must not move people outside the occupied territory unless it is materially impossible to avoid, and those evacuated must be returned home as soon as hostilities in the area have ceased. The ICRC commentary is emphatic that imperative means an absolute constraint leaving no other choice, not military convenience or advantage.

This matters for a cost tool because the exception in Article 49 turns on necessity, not on cost or convenience. A tool that produced a number and called it a recommendation would risk implying that an evacuation which is expensive is therefore not demanded, which inverts the legal test. ERCF's stated philosophy addresses this directly: labels are descriptive, not imperative; a Level 4 classification describes a risk profile and does not order an action. The break-even day is presented as a resource-planning figure, not as a decision rule.

The framework also draws on Sphere standards throughout for its quantities. The Sphere Handbook is the humanitarian sector's most widely used set of minimum standards for response, developed collectively by humanitarian organisations, covering water and sanitation, food security, shelter and health. It is a standards document rather than a costing document, which is precisely the boundary ERCF has to cross: Sphere tells the model how many litres and how many square metres, but not what they cost.

14

The Ethical Position

A tool that computes the cost of evacuating people and the cost of not evacuating them invites an obvious objection. The repository addresses it in a notice displayed in the interface itself and repeated in the concept document: the tool estimates the operational cost of humanitarian evacuation logistics, in order to support planning and resource mobilisation, and does not place a monetary value on human life.

The structural safeguards behind that claim are worth naming, because they are design decisions rather than assertions. Cost and mortality are computed by separate functions and never combined into a single index; there is no cost-per-life-saved figure anywhere in the codebase. The mortality output is described in the documentation as contextual support only, not a target, threshold or measure of acceptable loss. The break-even calculation compares one logistics operation against another logistics operation, never against a count of deaths. And the concept document states explicitly that the tool is not intended to determine whether an evacuation is worth it or to rank the value of civilian lives.

The residual risk remains, and it is the risk that attaches to any costing tool in this domain: a figure produced for a funding appeal can be repurposed in an argument about whether to act. The separation in the software is real, but it depends on the people using the output to maintain the same separation.

15

Notable Inconsistencies Found in Review

A plain-language review is of limited value if it only reports what the documentation claims. The following discrepancies were found by reading the code against the documentation. None is fatal, and several are the ordinary residue of a fast-moving research project, but a reader relying on the repository should know about them.

The acronym is used inconsistently. The README and both design documents give ERCF as the Evacuation Risk and Cost Framework. The prompt used for generating country briefings expands it as the Evacuation Risk Classification Framework. This report follows the README.

Comments have fallen out of step with the code. The explanation attached to the in-zone assistance calculation still lists access multipliers of 1.0, 1.5, 3.0, 5.0 and 8.0, while the values actually used are 1.0, 1.5, 2.0, 3.6 and 4.0. The same explanation gives a treatment cost of 1,200 US dollars per injury, while the code applies 800. In both cases the values in force are the more conservative and better-evidenced ones, so the effect is that the documentation overstates the model's outputs rather than understating them. A summary string displayed with the evacuation results still reports the Sphere emergency minimum of 15 litres of water per person per day although the calculation uses 20.

The stated date range of the corpus is wrong. The README describes the historical corpus as covering 1991 to 2024. The earliest case in the data file is the battle of Manila in 1945, and the corpus also includes Hue in 1968 and West Beirut in 1982. The correct range is 1945 to 2024.

One calibration script is not portable. It contains a file path pointing at a directory on the original author's own computer, which means it will not run on a fresh copy of the repository without editing that line. The main calibration script runs correctly. A comment in that script still refers to twenty in-scope cases, although it counts them dynamically and in practice uses sixteen.

One country code is wrong. The demographic dataset contains a country code entry for Mali that does not match that country's actual three-letter code, so a lookup by code would fail for Mali while a lookup by name succeeds.

16

Maturity, and What Remains Unbuilt

The core of ERCF is genuinely built. The seven-dimension scoring, the evacuation cost model, the in-zone assistance model, the break-even analysis, the mortality model, the historical corpus, the calibration pipeline, the statistical validation, the transport warnings, the radar comparison and the map are all implemented and running. The calibration can be reproduced by anyone who clones the repository and runs a single command. This is not a mock-up.

What the project's own backlog lists as outstanding is instructive, and it divides into three groups. In the first are features that would improve the tool as it stands: making fuel, food and medical supply prices adjustable per country, showing the user when the external data was last updated, adding a structured flag for cases involving mass atrocity rather than leaving it in free text, and building a pipeline for adding new historical cases and re-running the fit. In the second are capabilities the tool does not currently have: the dimension scores are a static snapshot with no modelling of how a conflict escalates over time, only one evacuation route can be assessed at a time, walking evacuation is not integrated into the break-even calculation, and there is no export to a document format that a funding appeal could use directly. In the third are research questions: expert-panel validation of the seven weights, resolving the Level 4 structural limitation that the two challenge cases document, and putting confidence intervals around the cost estimates, which are currently single point figures.

That last item deserves emphasis. Every cost figure the tool produces is a single number built from parameters that the project itself classifies as validated, estimated or unvalidated. A total assembled from a validated water quantity, an uncited food price, an estimated tent price and an unvalidated access multiplier is not a number with uniform reliability, and at present the interface does not show a range around it.

A companion module, RICS, the Resettlement and Inclusion Capacity Simulator, sits in a subdirectory of the same repository. It applies the same cost and break-even machinery to a different question: the long-term cost of a protracted refugee camp against the upfront cost of local integration, with Dzaleka camp in Malawi as its worked case. Its architecture is built out, with roughly two thousand lines of Python and a working interface, and it introduces two design ideas worth noting: a requirement, enforced by the database itself rather than by policy, that no group smaller than twenty people can be recorded, so that no individual can be identified; and a decision to propagate uncertainty into the cost estimate by widening each parameter into a band according to how well evidenced it is, rather than by simulation. But RICS is scaffolding at present. Several of its key parameters are unset placeholders, its readiness parameters are inherited wholesale from another tool and tagged unvalidated, and its own design document says that its central coefficient awaits calibration against real camp data. It should be read as a design proposal with working plumbing, not as a second finished tool.

A copy of a second Ethical Tech CoLab project sits in the repository as a reference. That tool, the Evacuation Readiness and Uncertainty Simulator, asks a different question: not what an evacuation costs, but how confident an assessment of readiness can be. It scores candidate destinations across seven factors, three of which are treated as gatekeepers, meaning that security, authority consent and host willingness each individually cap readiness at twenty per cent if blocked, no matter how good everything else looks. It runs five hundred simulations per pairing under a global uncertainty setting, in order to show that identical assessed readiness produces materially different predicted outcomes as confidence degrades. Its two epistemic distinctions are worth carrying into any reading of ERCF: that a confidently identified gap is better than an unreliably assessed one, and that Unknown and Unwilling must never be conflated, because an intelligence gap can be closed while a confirmed refusal cannot. The two projects share a documentation structure and an author, and they are the two parents of RICS, which ports the cost engine from ERCF and the readiness engine from the simulator. This report treats ERCF on its own terms; the simulator is not required to understand it.

17

Limitations and Caveats

Beyond the specific issues already described, the following limitations apply to the framework as a whole.

The weights are unvalidated. All seven are the author's modelled estimates. No expert panel has reviewed them.

The dimension scores are a snapshot. There is no modelling of escalation. A conflict that will look completely different in three weeks is scored as it looks today.

The planning window is 90 days. Beyond that, mortality accumulation is subject to a saturation adjustment and the model is outside the range of its own calibration data.

The mortality model fits scale, not individual cases. Fewer than half the fitted cases fall within a factor of two of their recorded death tolls. The model is useful for indicating whether a situation is in the hundreds, the thousands or the tens of thousands, and should not be used for anything finer.

The model has documented blind spots. It does not capture deliberate massacre or genocide, where death occurs in days rather than accumulating through attrition. It does not capture famine or the collapse of healthcare under blockade, which the code identifies as a major driver of mortality in Gaza that the seven dimensions do not represent. It underestimates open-corridor forced displacement, because it assumes a population trapped under fire rather than one compelled to move.

The most consequential cost multipliers are the least evidenced. The access multiplier that governs the entire in-zone assistance calculation is unconfirmed at its highest level, and the loss rates at Levels 2, 3 and 4, which reach fifty per cent, have no published equivalent that the author could locate.

Population figures are pre-conflict. City populations retrieved automatically do not account for displacement that has already taken place, and must be adjusted by hand.

The tool models one corridor. Real evacuation planning compares routes.

The historical corpus is small. Sixteen fitted cases is enough to sanity-check a framework, not enough to support statistical generalisation, and two known-difficult cases were held out of the fit.

Above all, this is a research prototype. It was produced by one researcher with academic supervision. Its own README states that it does not constitute operational advice and that all estimates require validation against country-specific intelligence and field assessment before any operational application.

18

Intended Audience and Use

The concept document identifies five groups: humanitarian programme officers producing rapid cost estimates for funding proposals, emergency coordinators comparing scenarios before a crisis escalates, donors and fund managers benchmarking against documented operations, researchers studying civilian protection frameworks, and government civil protection agencies doing preparedness planning.

The most defensible use of the tool today is the one furthest from an active crisis: preparedness planning, funding appeal preparation, and teaching. Its value in those settings does not depend on the precision of any individual figure. It depends on forcing a planner to state, in numbers, what they are assuming about vehicle capacity, staff ratios, supply losses and access costs, and on making those assumptions available for a colleague to contest.

The tool is explicitly not a real-time conflict monitoring system, not a forecasting model, not a mass casualty prediction tool, and not a substitute for field assessment or expert judgment. The repository states each of these explicitly.

19

Conclusion

ERCF's contribution is not a set of accurate numbers. It is the assembly of scattered, separately published operational standards into one calculation that can be run, inspected and argued with, together with an unusually disciplined habit of recording where each number came from and how far it can be trusted. The source code reads in places less like software than like a research notebook, documenting failed searches, revised assumptions, and figures that were lowered when the evidence would not support them. A model that reduced its own access multipliers from eight to four because no published source went above four is behaving the way a research instrument should.

The framework's structural insight is the separation it maintains between the danger a population faces and the practicality of moving it, carried through into the separation of cost from mortality and of description from recommendation. This keeps the legal question, which turns on necessity, from being quietly answered by an operational one, which turns on difficulty.

The prototype's weakest point is also clearly visible: the parameters carrying the most weight in the in-zone assistance calculation are the ones with the least published support behind them, and the mortality model resolves scale rather than magnitude. The author says so in the code, in the README and in the concept document. A reader should take the transparency as an invitation to check the work rather than as a reason to skip checking it.

What a tool of this kind can reasonably offer today is a structured first estimate, produced in an hour rather than a week, with every assumption on the surface where a humanitarian planner, a donor or a lawyer can see it and challenge it. That is a real contribution to a decision that is currently made largely from memory, and it is offered here without the claim that the numbers themselves are yet good enough to act on.

References

Sources

  1. 01International Committee of the Red Cross. Article 49, Geneva Convention (IV) relative to the Protection of Civilian Persons in Time of War, 1949.
  2. 02International Committee of the Red Cross. Articles 57 and 58, Protocol Additional to the Geneva Conventions (Protocol I), 1977, on precautions in attack and against the effects of attacks.
  3. 03Rome Statute of the International Criminal Court, provisions on unlawful deportation and transfer.
  4. 04Sphere Association. The Sphere Handbook: Humanitarian Charter and Minimum Standards in Humanitarian Response, 2018 edition.
  5. 05Uppsala Conflict Data Program. UCDP Georeferenced Event Dataset, version 26.1. Uppsala University.
  6. 06ACAPS and the Joint Research Centre of the European Commission. INFORM Severity Index.
  7. 07World Food Programme. United Nations Humanitarian Air Service, published operating cost of 2.08 US dollars per passenger kilometre, 2023.
  8. 08Open-Meteo. Historical weather from the European reanalysis archive.
  9. 09GeoNames. City population figures.
  10. 10United Nations Department of Economic and Social Affairs. World Population Prospects, 2022 revision.
  11. 11UNICEF. The State of the World's Children 2023.
  12. 12World Health Organization and World Bank. World Report on Disability, uniform 15 per cent disability prevalence figure.
  13. 13World Health Organization and UNICEF. Interagency Emergency Health Kit, costing used for basic medical kits.
  14. 14International Federation of Red Cross and Red Crescent Societies. (2021). Methodology paper reporting that per-capita response cost is statistically significantly higher in conflict settings.
  15. 15United Nations Office for the Coordination of Humanitarian Affairs. Monitoring of aid cargo in Gaza during the ceasefire period, late 2025.
  16. 16United Nations Independent International Commission of Inquiry on the Syrian Arab Republic, reporting on hospital bombing in Aleppo.
  17. 17International Criminal Tribunal for the former Yugoslavia, proceedings concerning the siege of Vukovar.
  18. 18Human Rights Watch and Amnesty International, reporting on Angola and the siege of Huambo.
  19. 19UNHCR. Emergency water planning standard of 20 litres per person per day, and publicly stated tent replacement cost, 2022.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file +

Publications · Academic report

The Evacuation Risk and Cost Framework

Estimating the Human and Financial Cost of Civilian Evacuation in Armed Conflict

Ethical Tech CoLab

NYU Center for Global Affairs

July 2026

Yago Rocha. Developed as part of research on International Humanitarian Law and civilian protection at the NYU Center for Global Affairs, under the Ethical Tech CoLab. The repository records the work as presented to Teresa Cantero, PhD Researcher, whose suggestions are cited in the project backlog.

A humanitarian organisation facing a besieged city has to answer a question that no manual answers well: what would it actually take to move these people, and what would it cost to keep them alive if they cannot be moved? Both questions have to be answered before a funding appeal can be written, and both are usually answered from memory, from a previous operation in a different country, or from an experienced coordinator's intuition. ERCF is a research prototype that attempts to put numbers on both and to show its working.

Key figures

7

weighted dimensions a planner scores from 1 to 5, with kinetic threat carrying a quarter of the total

31

documented historical conflicts in the corpus, sixteen of which were used to fit the mortality model

0.855

R-squared in logarithmic space, against only 44 per cent of fitted cases landing within a factor of two

90

days of planning window, beyond which the model sits outside the range of its own calibration data

01

Executive Summary

ERCF is a research prototype: experimental software, not an operational system. It helps a humanitarian planner estimate two things side by side. The first is the one-time cost of organising an evacuation. The second is the daily cost of keeping people supplied if they stay. It then compares the two.

The planner describes the situation using seven factors, called dimensions and labelled D1 to D7. Each is scored from 1 to 5. They cover the intensity of fighting, the mobility of the population, whether the armed parties have authorised movement, the state of roads and logistics, the safety of the destination, how fast the window is closing, and how good the available information is. These seven scores drive everything else the tool produces.

The seven scores are combined into a single weighted number, which is placed in one of five risk levels from Level 0 to Level 4. That level then selects the cost rates, loss rates and mortality rates used in the rest of the calculation. A planner who moves one slider sees the entire cost estimate change.

The distinctive output is the break-even analysis. Evacuation is a large one-time expense. Assistance in place is a smaller expense repeated every day. Somewhere in the future, if the crisis persists, the cumulative daily cost overtakes the one-time cost. ERCF calculates the day on which that happens, which is a question funding appeals have to answer and rarely answer with arithmetic.

ERCF also produces an estimate of deaths and injuries among a population that does not leave. This estimate was fitted against sixteen documented historical conflicts. Its author is explicit that this part of the tool is indicative and that the financial estimates are considerably more reliable than the mortality ones.

The project takes an explicit ethical position, stated in its own documentation and repeated in the interface. It estimates the cost of logistics. It does not place a monetary value on human life, and it does not tell a user whether an evacuation is worth carrying out. The labels it produces are descriptive, not instructions.

The software is substantially built rather than aspirational. The repository contains roughly eleven thousand lines of working Python, a browser interface, a reproducible calibration pipeline, a database of thirty-one documented historical conflicts, and connections to four external data services. It runs as a live public demonstration. What remains incomplete is documented openly in the project's own backlog.

02

Background and Rationale

An organisation deciding whether to support an evacuation needs three figures quickly: what evacuation would cost, what staying would cost, and where those two lines cross. Each figure exists somewhere. Sphere standards give water and food quantities per person. The World Food Programme publishes what its air service costs per passenger kilometre. Case studies record what particular operations spent. But these sit in separate documents written for separate purposes, and nobody has assembled them into a single calculation that a planner can run in an afternoon.

The project's concept document states the position directly: these estimates exist in operational manuals and case studies, but no publicly available tool integrates them with a historical calibration dataset and a framework grounded in International Humanitarian Law. ERCF is an attempt to build that integration and to publish the assumptions rather than hide them.

One finding recorded in the source code is worth restating for a policy audience. When the author searched for a published figure describing how much more expensive it is to deliver humanitarian assistance inside an active conflict than in a stable setting, no such figure was found. The International Federation of Red Cross and Red Crescent Societies confirms in a 2021 methodology paper that the per-capita cost of response is statistically significantly higher in conflict, but publishes no ratio. The World Food Programme publishes per-tonne logistics costs that imply a premium of roughly 1.23 times for Sudan in 2023 against an Eastern Africa regional average. Security costs, which the model treats as the larger component, appear only in fragments: an armed escort in Yemen at up to 3,000 US dollars per convoy on one route, security spending in Somalia in 2017 estimated at around 400 million US dollars per year against total humanitarian operations of 700 to 900 million. The code notes that a single published multiplier for active-conflict delivery does not appear to exist, and that this model may be among the first attempts to quantify it explicitly. That is an honest statement of a weakness and of a contribution at the same time.

ERCF builds the calculation anyway, using the best documented values available, and tags every single number with the strength of the evidence behind it. A planner reading the interface can see which figures are validated against a published standard, which are estimated from indirect evidence, and which are unvalidated assumptions the author could not support from any public source. That tagging is the tool's principal safeguard against being trusted more than it deserves.

03

Objectives

The tool is designed to:

  • Produce a rapid, itemised cost estimate for a civilian evacuation operation, broken into transport, fuel, personnel, food, water, shelter, medical supplies, communications and contingency.
  • Produce a parallel estimate for the cost of assisting a population that remains in the conflict zone, including the extra cost that conflict access conditions impose and the proportion of supplies that never arrives.
  • Identify the break-even point at which continuing in-zone assistance becomes more expensive than a one-time evacuation, so that funding appeals can be argued from arithmetic rather than assertion.
  • Place the current situation against documented historical operations, so that a planner can see which past conflicts a scenario most resembles.
  • Make every parameter, source and confidence level visible in the interface rather than buried in the code.
  • Keep the estimate of financial cost separate from the estimate of mortality, so that a funding calculation is never confused with a casualty prediction.

04

How ERCF Works: The Seven Dimensions

The whole tool rests on seven scores. Each is a number from 1 to 5 that the planner sets by moving a slider. The dimensions were chosen, according to the concept document, to capture the factors that field coordinators consistently cite when assessing whether an evacuation is feasible, and deliberately to require no real-time data feed or classified intelligence. A planner with a situation report and professional judgment can set all seven.

D1, Kinetic Threat. How much direct violence the civilian population is exposed to. A score of 5 means active, sustained attack. This is the dimension that drives almost everything downstream: it determines whether movement is physically survivable at all.

D2, Mobility Constraints. Also called Vulnerability. How much of the population cannot move unaided. A high score means many people who are elderly, very young, disabled, chronically ill or otherwise dependent on assisted transport. This dimension is what forces the operation to allocate medical buses and ambulances rather than ordinary buses.

D3, Authorization. Whether the armed parties have consented to civilian movement. The scale runs the same direction as the others: a high D3 score means authorization is a serious problem, that is, that consent is absent or unreliable. Without consent, an evacuation is both unlawful and practically blocked.

D4, Logistics. The state of roads, bridges, vehicles, fuel supply and the supporting infrastructure. The project's documentation observes that logistics collapse, not danger as such, has empirically been the leading cause of delayed evacuations, citing Mosul in 2016 and Aleppo in 2016.

D5, Destination. Whether the place people would be moved to is genuinely safe and able to receive them. The reason this dimension exists at all is Srebrenica in 1995, cited in the code as the canonical case in which evacuation to a nominally safe area became the site of a massacre. Moving civilians into danger is a distinct harm, not a lesser version of leaving them where they are.

D6, Urgency. How fast the window for organised movement is closing.

D7, Information. How poor the information environment is. A high score means communications failure, rumour, and an inability to coordinate.

The seven scores are combined into a single number by multiplying each by a weight and adding the results. The weights add up to 1.00, so the resulting score stays on the same 1 to 5 scale as the inputs.

DimensionFactorWeight
D1Kinetic Threat0.25
D2Mobility Constraints0.15
D3Authorization0.15
D4Logistics0.15
D5Destination0.15
D6Urgency0.10
D7Information0.05

ERCF = (D1 x 0.25) + (D2 x 0.15) + (D3 x 0.15) + (D4 x 0.15) + (D5 x 0.15) + (D6 x 0.10) + (D7 x 0.05)

The seven weights sum to 1.00.

D1 carries the largest weight, one quarter of the total, because direct physical threat is treated as the primary driver of evacuation necessity, and because it is the dimension that alone determines whether movement is safe. It is tied to the precautionary obligations in Articles 57 and 58 of Additional Protocol I.

D2, D3, D4 and D5 each carry fifteen per cent and form an equal group. The argument is that mobility, consent, logistics and destination safety are each individually capable of making an evacuation impossible, so none should dominate the others. D3 is explicitly capped at fifteen per cent rather than being set higher, because the most extreme authorization failures are already captured elsewhere in the tool by a hard trigger.

D6 carries ten per cent, deliberately below the equal group. The reasoning is that urgency is largely absorbed by D1 in the most extreme situations, and that the tool captures urgency more sharply through the hard trigger and through a separate extraction-probability floor than a linear weight would. D7 carries five per cent, the lowest. Poor information raises coordination cost and panic risk, but on its own it does not determine whether an evacuation is necessary or possible.

An important caveat is stated by the author in the code itself and reproduced in the interface as a red marker beside every weight: all seven weights are modelled estimates. No published framework in International Humanitarian Law or humanitarian operations specifies numerical weights for these factors, so there was nothing to copy. The author's own backlog lists expert-panel validation of these weights as outstanding research work. A reader should treat the weights as a reasoned starting proposal, not as an established standard.

05

From Seven Scores to a Risk Level

The weighted score is placed into one of five bands, each with a label and a NATO-doctrine equivalent used in military planning.

LevelLabelWeighted scoreNATO equivalent
Level 0Baseline and Monitoring1.5 or belowpermissive and stable
Level 1Low Risk and Advisoryabove 1.5 up to 2.5permissive but degrading
Level 2Moderate Risk and Watchfulabove 2.5 up to 3.5uncertain
Level 3High Risk and Contestedabove 3.5 up to 4.2hostile, partial
Level 4Critical and Emergencyabove 4.2hostile, imminent

There is one override. If D1 and D6 are both at 4.5 or above, that is, if violence is extreme and the window is closing at the same time, the score is floored at 4.21, which forces Level 4 regardless of how favourable the other five dimensions look. This exists because the linear weighted average would otherwise allow good logistics and a safe destination to pull an imminent massacre down into Level 3.

The level is not merely a label. It selects the numerical rates used everywhere downstream: the security escort ratio, the daily cost of assistance per person, the access multiplier, the supply loss rate, the injury rate, the mortality base rate, and the probability of emergency extraction. Almost every figure the tool produces changes when the level changes. This is a design choice worth noticing, because it means the five-band classification carries a great deal of weight and the boundaries between bands, at 1.5, 2.5, 3.5 and 4.2, are themselves modelled judgments rather than empirical findings.

Alongside the composite score, the tool computes three sub-indexes that keep distinct questions apart. This was added on the recommendation of the project's academic reviewers. Risk Severity asks how dangerous the situation is for civilians, and is built from D1, D2 and D6 only, rescaled back onto a 1 to 5 range. Feasibility asks whether people can realistically move, and is built from D3, D4 and D5, but inverted: because a high D3, D4 or D5 score means bad conditions, each is subtracted from 6 before being used, so that a high feasibility number means a genuinely open corridor. Information Quality is simply 6 minus D7, so that a high number means good situational awareness.

Keeping severity and feasibility apart matters. A situation can be extremely dangerous and simultaneously impossible to evacuate. Blending the two into one score would make that case look moderate, which is precisely the case that should escalate most urgently. The tool instead places the scenario in a four-cell matrix. High severity with high feasibility returns evacuate immediately. High severity with low feasibility returns shelter in place and negotiate a corridor urgently, with the reasoning that forced movement without a safe corridor risks greater harm than staying, and with a note that the precautionary obligations under Articles 57 and 58 of Additional Protocol I apply while negotiation continues. Low severity with high feasibility returns facilitate voluntary departure, and explicitly says do not mandate evacuation. Low severity with low feasibility returns monitor and reassess daily.

06

The Cost of Evacuating: Every Variable Explained

This is the part of the tool with the strongest evidentiary foundation. The planner supplies four things: the population at risk, the percentage of that population who are vulnerable, the distance in kilometres to safety, and the terrain quality. Everything else is derived.

The population is split into vulnerable and non-vulnerable. Non-vulnerable people are assigned to standard buses at 50 people per bus. Vulnerable people are assigned to medical buses at 20 people per bus, and ambulances at one per 150 vulnerable people. The bus capacities are marked in the code as operational assumptions not validated against field data. The ambulance ratio has a history worth recording: it was originally set at one ambulance per 40 vulnerable people, and was revised to one per 150 after the author found that no published field standard exists at all and that the original figure was three to five times above what documented practice, in particular a study of the Kosovo operation, actually showed.

The number of medical buses and ambulances is then multiplied by a factor derived from D2, the mobility dimension. At D2 of 1 the factor is 0.8, at 2 it is 1.0, at 3 it is 1.3, at 4 it is 1.8, and at 5 it is 2.5. In plain terms, a population with severe mobility constraints needs roughly two and a half times the assisted transport of a baseline population. This factor is described in the code as estimated with no primary source.

Staffing is set by ratio, and the security ratio tightens sharply as danger rises. The medical staffing figure is the Sphere Handbook 2018 standard for clinical officers in emergency settings, and it was corrected upward during development: an earlier version used one per 500, half the Sphere standard, with no documented justification.

RoleAllocationDaily rate, US dollars
Securitynone at Level 0, one per 500 civilians at Level 1, per 200 at Level 2, per 100 at Level 3, per 50 at Level 4300
Medical staffone per 250 people200
Paramedicsone per 100 people150
Driversone per vehicle50
Daily rates are total cost to the operation, not take-home pay. All four are tagged as estimated.

The code carries an unusually candid note about the rates: every rate implicitly assumes national staff or a lower-cost international non-governmental deployment at roughly the level of Médecins Sans Frontières, and full United Nations international professional staff, once daily subsistence allowance and danger pay are included, would cost three to six times more. The author records that this assumption is not stated in the interface and should be flagged.

Water is calculated at 20 litres per person per day for three days. This is the UNHCR full planning standard, chosen deliberately over the Sphere emergency minimum of 7.5 to 15 litres because an evacuation is a planned operation rather than a first-response emergency. Food is calculated at 0.45 kilograms of dry food per person per day for three days, which is the dry-weight equivalent of the Sphere minimum of 2,100 kilocalories per person per day. Tents are provided at one per five people, which at the Sphere standard of 3.5 square metres per person gives 17.5 square metres per tent and is internally consistent. Basic medical kits are provided at one per 100 people, and trauma kits at one per 50 people when the risk level is 3 or above, or one per 200 otherwise. Radios are provided at one per five vehicles plus a fixed five for coordination.

The unit costs applied to those quantities are set out below, each with the evidence the project records behind it. Fuel is the one line with a consumption figure of its own: it is calculated at 0.35 litres per kilometre per vehicle for a return journey before the per-litre price is applied.

ItemUnit costEvidence recorded in the code
Standard bus200 US dollars each
Medical bus400 US dollars eachRevised upward after research found no primary source for the earlier value; sits at the lower bound of documented field ranges for medically-equipped vehicles.
Ambulance700 US dollars eachRevised upward on the same basis as the medical bus.
Fuel1.20 US dollars per litreRevised down from 1.50 on the basis of ACAPS reporting of Yemeni consumer fuel prices in 2022.
Food3 US dollars per kilogramFlagged by the project's own parameter registry as having no citation behind it, even though the quantity does.
Water0.05 US dollars per litreField evidence gathered in 2026 found real water trucking costs of 2 to 23 US dollars per cubic metre, all below the model's baseline; the lower figures were not adopted pending further validation.
Tent380 US dollars eachRaised from an earlier 150 once that figure was found to describe a tarpaulin kit suitable for a week rather than a standard tent; 380 sits just below the 400 dollar replacement cost UNHCR gave publicly in 2022.
Basic medical kit21 US dollarsDerived from the WHO and UNICEF Interagency Emergency Health Kit costing, revised down from an earlier 50 that was roughly seven times too high for a three-day convoy.
Trauma kit200 US dollarsUnvalidated. The ICRC does not publish per-kit pricing.
Radio500 US dollars eachWithin the documented procurement range for professional handheld VHF units.

Three multipliers are applied on top. Terrain multiplies transport and fuel costs, with five levels running from 4.0 for the worst terrain down to 1.0 for the best, and 2.5, 1.7 and 1.2 in between. The lower end of this range is consistent with published road-condition cost models; the upper end of 4.0 is an expert estimate consistent with a World Food Programme figure showing per-tonne delivery costs in the Central African Republic, South Sudan and the Democratic Republic of the Congo running about five times a standard country average, though that figure mixes terrain, access and security together.

Season adjusts terrain further. If the operation starts in a month the tool classifies as a closure period for that latitude, the terrain multiplier is boosted by 50 per cent for the worst terrain, 30 per cent for the second worst and 20 per cent for middling terrain. Closure periods are set by latitude: December to March above 30 degrees north, June to September below 30 degrees south, and April to October in the tropics for wet season. The author describes these as broad regional approximations. The most useful output here is not the cost boost but a flag: the worst terrain in a closure period is marked potentially impassable, which is an operational warning rather than a number.

D4, logistics, adds 10 per cent to transport and fuel for each point above 1, so a D4 of 5 adds 40 per cent. D5, destination, changes the number of tents needed: a destination with good existing infrastructure halves the tent requirement, a destination with none doubles it. Both are marked estimated with no primary source. Finally, 15 per cent is added to the whole subtotal as contingency, which the code justifies as the lower bound of the 15 to 20 per cent standard used for high-risk projects.

A worked example is carried in the code. For 10,000 people, 20 per cent of them vulnerable, at Level 2, moving 50 kilometres: 160 standard buses, 100 medical buses and 50 ambulances, 310 vehicles in total; 50 security staff, 20 medical staff, 100 paramedics and 310 drivers; 10,850 litres of fuel, 15,000 kilograms of food, 450,000 litres of water, 2,000 tents and 67 radios. Transport comes to 92,500 US dollars, fuel to 16,275, personnel to 34,000, food to 45,000, water to 22,500, shelter to 300,000, medical to 15,000 and communications to 33,500, giving a subtotal of 558,775, a contingency of 83,816, and a total of roughly 643,000 US dollars, or about 64 dollars per person. Shelter dominates, which is a useful thing for a planner to be able to see at a glance.

07

The Cost of Staying

The second calculation estimates what it costs an organisation to keep a population alive where it is. It has three parts.

The first is supply delivery. The starting point is 3.50 US dollars per person per day, a multi-sector figure covering food, water, health, shelter and coordination. The World Food Programme's own global average for food and cash assistance alone was 42 US cents per beneficiary per day in 2023, and the code explains the difference between the two figures at some length: at the full Sphere water standard, water alone dominates supply weight, and applying the World Food Programme's per-tonne Sudan delivery cost to that weight gives 3.28 US dollars per person per day, within six per cent of the model's figure. That is a genuinely careful piece of reasoning.

The baseline is then multiplied by an access multiplier that represents how much more expensive delivery becomes as conflict intensifies. The second part of the calculation, supplies that never arrive, is then added as a loss rate representing goods destroyed, looted or simply undelivered. Both ladders are selected by the risk level.

Risk levelAccess multiplierSupply loss rate
Level 01.05 per cent
Level 11.55 per cent
Level 22.015 per cent
Level 33.630 per cent
Level 44.050 per cent

The access multipliers were revised downward during development, from earlier values of 3.0, 5.0 and 8.0 for the top three levels, after the author's research established that documented sources support only about 2.5 to 3.6 times at Level 3 and nothing above about 4 times at Level 4. Level 4 remains marked as directionally plausible but unconfirmed, and the project's backlog lists it as the outstanding unvalidated cost parameter. A further penalty is applied from D4, and the terrain and seasonal multipliers apply here too, because bad roads make delivery expensive whether people are leaving or staying.

Only the lowest of the loss rates has any published anchor: OCHA monitoring of Gaza during the ceasefire period in late 2025 recorded under two per cent of cargo looted or intercepted under active monitoring, and the model uses 5 per cent as a planning buffer including spoilage. The three higher figures are internal planning estimates with no published equivalent that the author could find. A low D3, meaning consent is absent, adds further to the loss rate, and a low D7, meaning poor information, adds a coordination overhead.

The third part is emergency extraction. The model assumes that some proportion of the population will have to be pulled out in an emergency at some point, and prices that in advance. The probability of this happening rises over time along a saturating curve: it starts near zero and approaches a ceiling. The daily rate driving that curve was fitted against historical cases and is 0.021 for Level 4, 0.010 for Level 3, 0.005 for Level 2 and 0.002 for Level 1. The Level 4 figure comes from averaging Mariupol in 2022 and Aleppo in 2016; the Level 3 figure from Mosul, Goma and the Central African Republic. The Level 2 and Level 1 figures have no historical anchor at all and are described as structurally plausible interpolations that require validation.

The curve is capped at 95 per cent for Level 4, 80 for Level 3, 60 for Level 2 and 30 for Level 1. Two modifiers then apply: a blocked corridor, meaning a low D3, raises the probability, and high urgency imposes a floor, 85 per cent when D6 is 5 and 60 per cent when D6 is 4. The floor exists because of Srebrenica, where the crisis unfolded over three days and no duration-based curve would have registered it.

The cost of extraction is anchored to the United Nations Humanitarian Air Service, which is the World Food Programme's air service flying humanitarian staff and light cargo into places commercial aviation does not serve. Its published operating cost was 2.08 US dollars per passenger kilometre in 2023. Ground extraction is priced at 30 per cent of that rate, which the code describes as an internal heuristic with no published source. Air medical evacuation is priced at three times the rate, on the reasoning that a medical flight carries far fewer passengers. At Level 4 the whole extraction cost is multiplied by 2.5 for a helicopter premium, which the author records as unconfirmed, since UNHAS does not publish separate helicopter and fixed-wing rates.

Field medical treatment is costed separately. Injuries are estimated from a per-thousand daily rate that rises with the risk level: zero at Level 0, 0.1 at Level 1, 0.5 at Level 2, 2.0 at Level 3 and 8.0 at Level 4. The code states plainly that no public source reports in-field civilian injury rates as a daily incidence per thousand, so these are unvalidated. Each injury is costed at 800 US dollars, modified upward when the destination has poor medical capacity. The 800 figure sits within a documented peer-reviewed range: 211 US dollars per surgical case at a conflict-affected mission hospital in South Sudan in 2022, and roughly 500 to 650 dollars per case at two MSF surgical trauma centres around 2009, which inflate to roughly 780 and 1,010 dollars by 2026.

08

The Break-Even Analysis

This is the calculation the tool is built around. Two options are compared. Option A is to evacuate now and then provide assistance to whoever remains behind. Option B is not to evacuate and to assist the whole population in place. Option A carries a large one-time cost plus a small daily cost. Option B carries only a daily cost, but a larger one.

The break-even day is the one-time evacuation cost divided by the difference between the two daily costs. If evacuating costs 640,000 US dollars and staying costs 4,000 US dollars a day more than the post-evacuation arrangement, the break-even is 160 days. Before that day, staying is cheaper. After it, evacuation would have been cheaper.

It is a resource-planning number. The project's concept document is emphatic that it is not an answer to whether an evacuation is worth carrying out. The distinction is that the calculation compares the cost of one logistics operation against the cost of another logistics operation. It does not weigh either against lives. The decision to move civilians, the documentation states, is a matter of legal obligation and humanitarian judgment, not financial optimisation.

The tool also checks the transport mode against the dimension scores and warns when they are inconsistent. Five patterns trigger a warning: ground transport with a D4 of 4.0 or above, which indicates logistics breakdown; ground transport with a D1 of 4.5 or above, which indicates unacceptable kinetic exposure; walking with a population above 5,000, which is infeasible at scale; walking with a D2 of 3.5 or above, which means the population cannot walk; and air transport with a D3 of 2.0 or below, which means airspace authorization is missing. These are useful precisely because a cost model will happily produce a plausible-looking figure for an operation that could never be attempted.

09

Estimating Deaths: The Mortality Model

Separately from cost, the tool estimates how many people would die and be injured if the population stayed. The author's framing is that this provides scale context for a planning decision, and both the README and the concept document state that the mortality model is indicative rather than predictive and that the financial estimates are substantially more reliable.

Each risk level has a baseline death rate expressed per 10,000 people per day. The current values are 0.777 at Level 0, 0.964 at Level 1, 3.625 at Level 2, 1.805 at Level 3 and 1.000 at Level 4. A reader will immediately notice that these are not in ascending order, which looks wrong and requires explanation. The explanation given in the code is that these are not standalone death rates but base rates that get multiplied by three further factors, and that Level 3 in the historical corpus is dominated by urban sieges and city fighting where a high proportion of the population is directly exposed, while Level 4 is dominated by large-enclave operations across much bigger populations where per-capita exposure is lower. The ordering was produced by the fitting procedure rather than assumed, and the author flags it explicitly as empirically validated but counterintuitive.

The confinement multiplier captures whether people are trapped. It is calculated as 5 minus D3, multiplied by D4, divided by 5. In plain words: when consent for movement is absent and logistics have collapsed at the same time, people cannot get out, and mortality rises sharply. The resulting score is converted into a multiplier in steps: half at the lowest, then 1, then 2, then 4, then 8 at the highest. An eightfold difference from a single factor is a very large lever, and its stepwise rather than smooth form is listed by the author as a known limitation. The anchors given are Aleppo and Kosovo at a multiplier of 2, Mosul at 1, and the Kherson evacuation, which had an open corridor, at 0.5.

confinement score = ((5 - D3) x D4) / 5

The score is then mapped to a multiplier in steps: 0.5, 1, 2, 4, 8.

The displacement protection factor applies where part of the population has already left, since the remaining exposure is lower. The tool reduces the death rate by 60 per cent of the share that has departed. The 60 per cent coefficient, rather than 100, reflects that displaced people still face risk on the road, at checkpoints and from exposure. There is then an important refinement: in a siege, defined as a D3 of 2 or below combined with a D1 of 4 or above and a population of 500,000 or fewer, the coefficient is halved to 30 per cent. The reasoning is that in an encircled city, movement itself is lethal, because civilians pass checkpoints under fire and buses have been attacked. Displacement is still safer than staying, but only half as much safer as in an open corridor.

The geographic exposure factor recognises that not everyone in a conflict area is under fire at the same time. Four conflict shapes are recognised, with the fraction of the population treated as simultaneously exposed given in brackets: urban siege such as Mariupol or Aleppo (0.85), enclave such as Gaza (0.65), city conflict where a front line moves through a city, such as Mosul or Kherson (0.40), and regional dispersed conflict such as Sudan, the Central African Republic or the eastern Democratic Republic of the Congo (0.12). When the planner has not specified a shape, the tool infers one from D1 and the population size, on the logic that higher kinetic threat means more direct fire while a larger population means more dispersion. The population term uses a logarithm raised to the power 1.4, a refinement introduced specifically because the earlier formula did not fall away fast enough for continental-scale conflicts and overestimated Sudan by a wide margin.

Cumulative deaths accumulate linearly for the first 90 days and then decelerate along a square-root curve, on the reasoning that populations adapt and survivors relocate. This also marks the outer edge of the model's intended planning window. Injuries are set at four times deaths, following the ICRC planning ratio. The code notes that a peer-reviewed systematic review implies a ratio closer to 3.3 to 1, and that frontline-specific figures from Ukraine run far higher, and retains 4 to 1 as a mid-range planning estimate.

The infrastructure-denial multiplier applies only where there is primary source documentation that survival infrastructure was deliberately destroyed. It is calculated as 1 plus 0.4251 times D1 minus 3, times D4 minus 3, and only when D1 is at least 4.5 and D4 at least 4.0. It is switched on for four documented cases: Mariupol, Aleppo, Vukovar and Huambo, giving effective multipliers between roughly 1.6 and 2.3. The supporting documentation cited is a 2024 report on starvation as a method of warfare in Mariupol, the UN Commission of Inquiry on Syria for hospital bombing in Aleppo, ICTY proceedings for Vukovar, and Human Rights Watch and Amnesty reporting on Angola for Huambo. Critically, this multiplier is switched off by default for live scenarios. The code records why: an earlier version activated it automatically whenever the dimension thresholds were met, and calibration accuracy collapsed from 80 per cent to 20 per cent. It is a finding about deliberate atrocity, not a threshold that can be inferred from slider positions.

infrastructure-denial multiplier = 1 + (0.4251 x (D1 - 3) x (D4 - 3))

Applied only where D1 is at least 4.5 and D4 at least 4.0, and only with primary source documentation. Off by default for live scenarios.

10

How the Model Was Fitted, and Against What

The repository contains thirty-one documented conflicts. Each record holds the population at risk, duration in days, recorded deaths, numbers displaced and remaining, the estimated vulnerable percentage with its own source and confidence rating, the distance to safety with its source, all seven dimension scores assigned retrospectively, corridor status and notes, key lessons, IHL issues, and a calibration block recording what the model predicted against what was recorded, across every model version.

Sixteen cases are used to fit the model: Mariupol 2022, Gaza 2023, Aleppo 2016, the Kherson evacuation 2022, Mosul 2016, eastern DRC and Goma 2024, the second battle of Grozny 1999, both battles of Fallujah 2004, the Jenin refugee camp 2002, the siege of Vukovar 1991, Gaza Cast Lead 2008, Gaza Protective Edge 2014, the siege of West Beirut 1982, the second Nagorno-Karabakh war 2020, and the siege of Huambo 1993.

Fifteen are excluded, and the reasons form a clear statement of where the model does not apply. Duration beyond the 90-day window excludes Sarajevo at 1,425 days, Kuito in Angola at over 540, Marawi at 154 and the final phase of the Sri Lankan Vanni offensive at 120. Dispersed regional conflict with no siege perimeter excludes the Central African Republic in 2014 and Sudan in 2023, the latter also on duration and on a population of six million. Deliberate massacre as the dominant mechanism of death, rather than attrition under fire, excludes Srebrenica 1995, Hue 1968, Manila 1945 and Bucha 2022. Open-corridor forced displacement, where civilians were made to move rather than prevented from moving, excludes Kosovo 1999. Recorded death counts too uncertain to fit against exclude Misrata 2011, where the contested range runs from 102 to 700.

Two exclusions are of a different character and the repository says so openly. Eastern Ghouta 2018 and Gaza Pillar of Defense 2012 are labelled challenge cases and were excluded, in the words of the data file, to preserve the version 7 metrics. They document a real structural limit: the Level 4 base rate cannot simultaneously fit a high-mortality urban siege such as Aleppo, with 31,000 deaths, and a large-enclave precision operation such as Cast Lead, with 965. The model undercounts one and overcounts the other by a wide margin. Retaining them in the corpus while removing them from the fit is a defensible research decision, and the fact that it is documented rather than concealed is to the project's credit. But a reader should understand that the headline accuracy figures are calculated on a set from which two known-difficult cases were removed.

The fitting procedure adjusts six numbers: the five base death rates and the infrastructure-denial coefficient. It searches for the values that minimise the mean squared error between the logarithm of predicted deaths and the logarithm of recorded deaths. Working in logarithms means the procedure fits proportional rather than absolute error, which is appropriate when the recorded death counts in the corpus run from 37 to 45,000. The search is constrained so that the rates for Levels 0, 1 and 2 rise in order; Levels 3 and 4 are deliberately left unconstrained relative to each other, which is how the counterintuitive ordering emerged.

On the sixteen fitted cases, the correlation between predicted and recorded deaths in logarithmic space gives an R-squared of 0.855, meaning the model captures most of the variation in order of magnitude. Seven of the sixteen cases, or 44 per cent, fall within a factor of two of the recorded figure. Those two statements sit oddly together and both should be kept in view: the model is good at ranking conflicts by scale of loss and much less good at predicting the number for any individual conflict. A battery of statistical checks reported in the README covers normality of residuals, constant variance, rank correlation and multicollinearity, and reports Aleppo as the single most influential case.

One qualification about the cross-validation figure is worth making. The reported leave-one-out R-squared of 0.807 is not produced by refitting the whole model with each case removed in turn. It fits a straight line through the remaining points in logarithmic space and predicts the held-out point from that line. It therefore measures how stable the relationship between predicted and recorded deaths is across the corpus, which is useful, rather than how well fitted parameters generalise to unseen conflicts, which is what the term usually implies.

11

Where the Data Comes From

The Uppsala Conflict Data Program, based at Uppsala University in Sweden, is the principal academic provider of data on organised violence worldwide. Its Georeferenced Event Dataset, version 26.1, records individual lethal events down to the level of a single village on a single day, with separate counts for civilian deaths and for each party, and low, best and high estimates reflecting source uncertainty. ERCF uses it to check its historical cases against an independent record and to let a planner query recorded violence for a location and date range. The tool queries the live service where a token is available and falls back to a locally stored copy of the dataset otherwise.

ACAPS is an independent humanitarian analysis organisation. Together with the European Commission's Joint Research Centre it publishes the INFORM Severity Index, a monthly assessment scoring the severity of humanitarian crises across more than thirty indicators grouped into the impact of the crisis, the conditions of the people affected, and the complexity of the response. ERCF uses these scores to colour a world map by country-level severity and to convert a national severity score onto its own five-level scale. Two things should be noted. The repository works from a stored table of roughly eighty countries with a live overlay attempted on top, and the code records that the two relevant ACAPS endpoints were consistently timing out at the time of writing, which means in practice the stored table is usually what drives the map. Users should treat the map as an orientation aid rather than a live feed.

GeoNames provides city population figures so a planner can start from a real place rather than a guess. The project's own backlog flags the obvious limitation: these are pre-conflict populations and do not reflect displacement that has already occurred, so a user must adjust them manually.

Open-Meteo provides historical weather from the European reanalysis archive. ERCF uses fourteen days of temperature, precipitation, snowfall and wind to classify the operating conditions and to set two multipliers: one on fuel and transport, one on shelter. Deep winter raises fuel costs by 35 per cent and shelter costs by 120 per cent, the latter reflecting that cold-climate shelter requires more space per person and winterised units cost several times a standard tent. A heavy rainy season raises fuel and transport by 80 per cent, reflecting documented road impassability. If the planned date is in the future, the tool steps back a year at a time to find comparable historical conditions and labels the result a planning estimate. The weather data is tagged as validated; the multipliers derived from it are tagged as estimated.

A language model is used for one narrow purpose: generating a written country briefing that summarises the crisis, access conditions, likely exit routes, neighbouring safe areas, active humanitarian actors and the applicable legal framework, and proposes starting values for the seven dimensions. If the service is unavailable the tool falls back to fixed text built from the stored severity table. This is a convenience feature for a first draft of a scenario. Nothing in the cost or mortality calculation depends on it.

For eighteen conflict-affected countries the tool suggests a vulnerable population percentage. The figure is calculated as the percentage of children under five, plus the percentage of adults over sixty, plus half the percentage of people with disabilities. The halving is to avoid double-counting people who are both elderly and disabled, and to reflect that only some disabilities create evacuation-relevant constraints. The underlying figures come from UNICEF's State of the World's Children 2023 and the UN World Population Prospects 2022, with a uniform 15 per cent disability figure drawn from the WHO World Report on Disability. The results range from 21.5 per cent for Syria, Iraq and Libya to 36.5 per cent for Ukraine, where the driver is an unusually large elderly population. Two caveats matter: the child component counts only children under five rather than all minors, and the disability figure is a global average applied uniformly, whereas conflict-affected populations are known to have higher rates. The suggestion is advisory and the planner confirms or dismisses it.

Two supplementary transport models exist. Air evacuation prices fixed-wing movement at the same UNHAS rate of 2.08 US dollars per passenger kilometre and helicopter movement from a United Nations procurement contract for a Mi-8 airframe divided by an assumed number of guaranteed flight hours, returning a low, mid and high band rather than a single figure, with a deliberately conservative 50 per cent load factor. It plans fleet size across one, three, seven, fourteen and thirty day windows. Walking evacuation interpolates group speed between 8 kilometres per hour for healthy adults and 3 kilometres per hour for a fully vulnerable group, the latter taken from published gait-speed data for adults over 65, and assumes 8 walking hours a day rather than the 15 used in the academic migration model it draws on. It contains no attrition or mortality term at all, reports only the number of days of exposure, and states explicitly that the exposure does not feed back into the danger or urgency dimensions: that remains the analyst's judgment. Both models are marked as less validated than the ground transport model, and the backlog records that walking is not yet wired into the break-even calculation.

12

Reading the Results

The interface is a single web page with several sections. The Scenario Builder holds the seven sliders and the population, vulnerability, distance and terrain inputs, and updates every figure as the sliders move. A cost breakdown itemises the evacuation estimate line by line. A decision analysis panel shows the break-even chart with the two cost lines crossing. Historical Cases lists all thirty-one conflicts with badges marking which are out of scope and which are challenge cases. Compare on Radar overlays the current scenario's seven scores against any historical case as a seven-pointed shape, alongside a table comparing population, cost, cost per evacuee, duration and deaths. Map View colours the world by severity. A references section lists the data sources, and a methodology panel documents the statistical validation.

Experienced coordinators reason by analogy, saying that a situation resembles Mosul or resembles Mariupol. The radar overlay makes that comparison explicit and checkable, by showing exactly which of the seven dimensions match and which do not. That is a modest but genuine contribution: it turns an intuition into something a colleague can disagree with.

Three real operations are provided for cost benchmarking: the Kosovo Humanitarian Evacuation Programme in 1999, which moved 60,549 people at roughly 562 US dollars each; Libya in 2011 at roughly 984 dollars; and the evacuation of about 15,000 people from Lebanon to Canada in 2006 at several thousand dollars each. The README is careful to explain why these exceed the model's own figures by a wide margin: ERCF prices the immediate field operation only, at roughly 134 US dollars per person for a Level 2 ground movement over 50 kilometres, while real operations include international transport, reception facilities and weeks of hosting. A user comparing the tool's output to a real operation without reading that note would draw the wrong conclusion.

13

Grounding in International Humanitarian Law

Each risk level is tied to a stated legal reference. At Level 1 the tool cites the right of voluntary departure under Article 35 of the Fourth Geneva Convention. At Level 2 it cites the precautionary obligations in Articles 57 and 58 of Additional Protocol I, which require parties to take feasible precautions to spare civilians, including effective advance warning, and to keep civilians away from military objectives. At Level 3 it cites Article 49 of the Fourth Geneva Convention. At Level 4 it cites the same article's prohibition on forcible transfer and the corresponding provision of the Rome Statute of the International Criminal Court.

Article 49 is the central provision and deserves setting out. It prohibits individual or mass forcible transfers and deportations of protected persons from occupied territory, regardless of motive. It then creates a narrow exception: an Occupying Power may undertake total or partial evacuation of a given area if the security of the population or imperative military reasons so demand. Such evacuations must not move people outside the occupied territory unless it is materially impossible to avoid, and those evacuated must be returned home as soon as hostilities in the area have ceased. The ICRC commentary is emphatic that imperative means an absolute constraint leaving no other choice, not military convenience or advantage.

This matters for a cost tool because the exception in Article 49 turns on necessity, not on cost or convenience. A tool that produced a number and called it a recommendation would risk implying that an evacuation which is expensive is therefore not demanded, which inverts the legal test. ERCF's stated philosophy addresses this directly: labels are descriptive, not imperative; a Level 4 classification describes a risk profile and does not order an action. The break-even day is presented as a resource-planning figure, not as a decision rule.

The framework also draws on Sphere standards throughout for its quantities. The Sphere Handbook is the humanitarian sector's most widely used set of minimum standards for response, developed collectively by humanitarian organisations, covering water and sanitation, food security, shelter and health. It is a standards document rather than a costing document, which is precisely the boundary ERCF has to cross: Sphere tells the model how many litres and how many square metres, but not what they cost.

14

The Ethical Position

A tool that computes the cost of evacuating people and the cost of not evacuating them invites an obvious objection. The repository addresses it in a notice displayed in the interface itself and repeated in the concept document: the tool estimates the operational cost of humanitarian evacuation logistics, in order to support planning and resource mobilisation, and does not place a monetary value on human life.

The structural safeguards behind that claim are worth naming, because they are design decisions rather than assertions. Cost and mortality are computed by separate functions and never combined into a single index; there is no cost-per-life-saved figure anywhere in the codebase. The mortality output is described in the documentation as contextual support only, not a target, threshold or measure of acceptable loss. The break-even calculation compares one logistics operation against another logistics operation, never against a count of deaths. And the concept document states explicitly that the tool is not intended to determine whether an evacuation is worth it or to rank the value of civilian lives.

The residual risk remains, and it is the risk that attaches to any costing tool in this domain: a figure produced for a funding appeal can be repurposed in an argument about whether to act. The separation in the software is real, but it depends on the people using the output to maintain the same separation.

15

Notable Inconsistencies Found in Review

A plain-language review is of limited value if it only reports what the documentation claims. The following discrepancies were found by reading the code against the documentation. None is fatal, and several are the ordinary residue of a fast-moving research project, but a reader relying on the repository should know about them.

The acronym is used inconsistently. The README and both design documents give ERCF as the Evacuation Risk and Cost Framework. The prompt used for generating country briefings expands it as the Evacuation Risk Classification Framework. This report follows the README.

Comments have fallen out of step with the code. The explanation attached to the in-zone assistance calculation still lists access multipliers of 1.0, 1.5, 3.0, 5.0 and 8.0, while the values actually used are 1.0, 1.5, 2.0, 3.6 and 4.0. The same explanation gives a treatment cost of 1,200 US dollars per injury, while the code applies 800. In both cases the values in force are the more conservative and better-evidenced ones, so the effect is that the documentation overstates the model's outputs rather than understating them. A summary string displayed with the evacuation results still reports the Sphere emergency minimum of 15 litres of water per person per day although the calculation uses 20.

The stated date range of the corpus is wrong. The README describes the historical corpus as covering 1991 to 2024. The earliest case in the data file is the battle of Manila in 1945, and the corpus also includes Hue in 1968 and West Beirut in 1982. The correct range is 1945 to 2024.

One calibration script is not portable. It contains a file path pointing at a directory on the original author's own computer, which means it will not run on a fresh copy of the repository without editing that line. The main calibration script runs correctly. A comment in that script still refers to twenty in-scope cases, although it counts them dynamically and in practice uses sixteen.

One country code is wrong. The demographic dataset contains a country code entry for Mali that does not match that country's actual three-letter code, so a lookup by code would fail for Mali while a lookup by name succeeds.

16

Maturity, and What Remains Unbuilt

The core of ERCF is genuinely built. The seven-dimension scoring, the evacuation cost model, the in-zone assistance model, the break-even analysis, the mortality model, the historical corpus, the calibration pipeline, the statistical validation, the transport warnings, the radar comparison and the map are all implemented and running. The calibration can be reproduced by anyone who clones the repository and runs a single command. This is not a mock-up.

What the project's own backlog lists as outstanding is instructive, and it divides into three groups. In the first are features that would improve the tool as it stands: making fuel, food and medical supply prices adjustable per country, showing the user when the external data was last updated, adding a structured flag for cases involving mass atrocity rather than leaving it in free text, and building a pipeline for adding new historical cases and re-running the fit. In the second are capabilities the tool does not currently have: the dimension scores are a static snapshot with no modelling of how a conflict escalates over time, only one evacuation route can be assessed at a time, walking evacuation is not integrated into the break-even calculation, and there is no export to a document format that a funding appeal could use directly. In the third are research questions: expert-panel validation of the seven weights, resolving the Level 4 structural limitation that the two challenge cases document, and putting confidence intervals around the cost estimates, which are currently single point figures.

That last item deserves emphasis. Every cost figure the tool produces is a single number built from parameters that the project itself classifies as validated, estimated or unvalidated. A total assembled from a validated water quantity, an uncited food price, an estimated tent price and an unvalidated access multiplier is not a number with uniform reliability, and at present the interface does not show a range around it.

A companion module, RICS, the Resettlement and Inclusion Capacity Simulator, sits in a subdirectory of the same repository. It applies the same cost and break-even machinery to a different question: the long-term cost of a protracted refugee camp against the upfront cost of local integration, with Dzaleka camp in Malawi as its worked case. Its architecture is built out, with roughly two thousand lines of Python and a working interface, and it introduces two design ideas worth noting: a requirement, enforced by the database itself rather than by policy, that no group smaller than twenty people can be recorded, so that no individual can be identified; and a decision to propagate uncertainty into the cost estimate by widening each parameter into a band according to how well evidenced it is, rather than by simulation. But RICS is scaffolding at present. Several of its key parameters are unset placeholders, its readiness parameters are inherited wholesale from another tool and tagged unvalidated, and its own design document says that its central coefficient awaits calibration against real camp data. It should be read as a design proposal with working plumbing, not as a second finished tool.

A copy of a second Ethical Tech CoLab project sits in the repository as a reference. That tool, the Evacuation Readiness and Uncertainty Simulator, asks a different question: not what an evacuation costs, but how confident an assessment of readiness can be. It scores candidate destinations across seven factors, three of which are treated as gatekeepers, meaning that security, authority consent and host willingness each individually cap readiness at twenty per cent if blocked, no matter how good everything else looks. It runs five hundred simulations per pairing under a global uncertainty setting, in order to show that identical assessed readiness produces materially different predicted outcomes as confidence degrades. Its two epistemic distinctions are worth carrying into any reading of ERCF: that a confidently identified gap is better than an unreliably assessed one, and that Unknown and Unwilling must never be conflated, because an intelligence gap can be closed while a confirmed refusal cannot. The two projects share a documentation structure and an author, and they are the two parents of RICS, which ports the cost engine from ERCF and the readiness engine from the simulator. This report treats ERCF on its own terms; the simulator is not required to understand it.

17

Limitations and Caveats

Beyond the specific issues already described, the following limitations apply to the framework as a whole.

The weights are unvalidated. All seven are the author's modelled estimates. No expert panel has reviewed them.

The dimension scores are a snapshot. There is no modelling of escalation. A conflict that will look completely different in three weeks is scored as it looks today.

The planning window is 90 days. Beyond that, mortality accumulation is subject to a saturation adjustment and the model is outside the range of its own calibration data.

The mortality model fits scale, not individual cases. Fewer than half the fitted cases fall within a factor of two of their recorded death tolls. The model is useful for indicating whether a situation is in the hundreds, the thousands or the tens of thousands, and should not be used for anything finer.

The model has documented blind spots. It does not capture deliberate massacre or genocide, where death occurs in days rather than accumulating through attrition. It does not capture famine or the collapse of healthcare under blockade, which the code identifies as a major driver of mortality in Gaza that the seven dimensions do not represent. It underestimates open-corridor forced displacement, because it assumes a population trapped under fire rather than one compelled to move.

The most consequential cost multipliers are the least evidenced. The access multiplier that governs the entire in-zone assistance calculation is unconfirmed at its highest level, and the loss rates at Levels 2, 3 and 4, which reach fifty per cent, have no published equivalent that the author could locate.

Population figures are pre-conflict. City populations retrieved automatically do not account for displacement that has already taken place, and must be adjusted by hand.

The tool models one corridor. Real evacuation planning compares routes.

The historical corpus is small. Sixteen fitted cases is enough to sanity-check a framework, not enough to support statistical generalisation, and two known-difficult cases were held out of the fit.

Above all, this is a research prototype. It was produced by one researcher with academic supervision. Its own README states that it does not constitute operational advice and that all estimates require validation against country-specific intelligence and field assessment before any operational application.

18

Intended Audience and Use

The concept document identifies five groups: humanitarian programme officers producing rapid cost estimates for funding proposals, emergency coordinators comparing scenarios before a crisis escalates, donors and fund managers benchmarking against documented operations, researchers studying civilian protection frameworks, and government civil protection agencies doing preparedness planning.

The most defensible use of the tool today is the one furthest from an active crisis: preparedness planning, funding appeal preparation, and teaching. Its value in those settings does not depend on the precision of any individual figure. It depends on forcing a planner to state, in numbers, what they are assuming about vehicle capacity, staff ratios, supply losses and access costs, and on making those assumptions available for a colleague to contest.

The tool is explicitly not a real-time conflict monitoring system, not a forecasting model, not a mass casualty prediction tool, and not a substitute for field assessment or expert judgment. The repository states each of these explicitly.

19

Conclusion

ERCF's contribution is not a set of accurate numbers. It is the assembly of scattered, separately published operational standards into one calculation that can be run, inspected and argued with, together with an unusually disciplined habit of recording where each number came from and how far it can be trusted. The source code reads in places less like software than like a research notebook, documenting failed searches, revised assumptions, and figures that were lowered when the evidence would not support them. A model that reduced its own access multipliers from eight to four because no published source went above four is behaving the way a research instrument should.

The framework's structural insight is the separation it maintains between the danger a population faces and the practicality of moving it, carried through into the separation of cost from mortality and of description from recommendation. This keeps the legal question, which turns on necessity, from being quietly answered by an operational one, which turns on difficulty.

The prototype's weakest point is also clearly visible: the parameters carrying the most weight in the in-zone assistance calculation are the ones with the least published support behind them, and the mortality model resolves scale rather than magnitude. The author says so in the code, in the README and in the concept document. A reader should take the transparency as an invitation to check the work rather than as a reason to skip checking it.

What a tool of this kind can reasonably offer today is a structured first estimate, produced in an hour rather than a week, with every assumption on the surface where a humanitarian planner, a donor or a lawyer can see it and challenge it. That is a real contribution to a decision that is currently made largely from memory, and it is offered here without the claim that the numbers themselves are yet good enough to act on.

References

Sources

  1. 01International Committee of the Red Cross. Article 49, Geneva Convention (IV) relative to the Protection of Civilian Persons in Time of War, 1949.
  2. 02International Committee of the Red Cross. Articles 57 and 58, Protocol Additional to the Geneva Conventions (Protocol I), 1977, on precautions in attack and against the effects of attacks.
  3. 03Rome Statute of the International Criminal Court, provisions on unlawful deportation and transfer.
  4. 04Sphere Association. The Sphere Handbook: Humanitarian Charter and Minimum Standards in Humanitarian Response, 2018 edition.
  5. 05Uppsala Conflict Data Program. UCDP Georeferenced Event Dataset, version 26.1. Uppsala University.
  6. 06ACAPS and the Joint Research Centre of the European Commission. INFORM Severity Index.
  7. 07World Food Programme. United Nations Humanitarian Air Service, published operating cost of 2.08 US dollars per passenger kilometre, 2023.
  8. 08Open-Meteo. Historical weather from the European reanalysis archive.
  9. 09GeoNames. City population figures.
  10. 10United Nations Department of Economic and Social Affairs. World Population Prospects, 2022 revision.
  11. 11UNICEF. The State of the World's Children 2023.
  12. 12World Health Organization and World Bank. World Report on Disability, uniform 15 per cent disability prevalence figure.
  13. 13World Health Organization and UNICEF. Interagency Emergency Health Kit, costing used for basic medical kits.
  14. 14International Federation of Red Cross and Red Crescent Societies. (2021). Methodology paper reporting that per-capita response cost is statistically significantly higher in conflict settings.
  15. 15United Nations Office for the Coordination of Humanitarian Affairs. Monitoring of aid cargo in Gaza during the ceasefire period, late 2025.
  16. 16United Nations Independent International Commission of Inquiry on the Syrian Arab Republic, reporting on hospital bombing in Aleppo.
  17. 17International Criminal Tribunal for the former Yugoslavia, proceedings concerning the siege of Vukovar.
  18. 18Human Rights Watch and Amnesty International, reporting on Angola and the siege of Huambo.
  19. 19UNHCR. Emergency water planning standard of 20 litres per person per day, and publicly stated tent replacement cost, 2022.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file diff --git a/static-site/print/ercf/index.txt b/static-site/print/ercf/index.txt index 2966c90b4..02dd61633 100644 --- a/static-site/print/ercf/index.txt +++ b/static-site/print/ercf/index.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -12:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","print","ercf",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","ercf","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -14:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -16:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","print","ercf",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","ercf","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:"$Sreact.suspense" -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1c:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -29,7 +29,7 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 19:[] 10:"$W19" 11:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1e:Ta14, @page { size: 210mm 297mm; margin: 20mm 18mm; } @@ -184,6 +184,6 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 67:["$","tr","8",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Trauma kit"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"200 US dollars"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Unvalidated. The ICRC does not publish per-kit pricing."}]]}] 68:["$","tr","9",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Radio"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"500 US dollars each"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Within the documented procurement range for professional handheld VHF units."}]]}] 1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -6d:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +6d:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 18:null 1d:[["$","title","0",{"children":"The Evacuation Risk and Cost Framework (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L6d","6",{}]] diff --git a/static-site/print/erus/__next._full.txt b/static-site/print/erus/__next._full.txt index c0b129a67..fc3a6aaa5 100644 --- a/static-site/print/erus/__next._full.txt +++ b/static-site/print/erus/__next._full.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -12:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","print","erus",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","erus","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -14:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -16:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","print","erus",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","erus","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:"$Sreact.suspense" -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1c:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -29,7 +29,7 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 19:[] 10:"$W19" 11:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1e:Ta14, @page { size: 210mm 297mm; margin: 20mm 18mm; } @@ -174,6 +174,6 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 5a:["$","tr","7",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Mixed general population"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"2"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"None"}],["$","td","3",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Can wait"}]]}] 5b:["$","figcaption",null,{"className":"mt-3 text-sm text-muted","children":"The eight evacuee archetypes and their fixed properties."}] 1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -5c:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +5c:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 18:null 1d:[["$","title","0",{"children":"The Evacuation Readiness and Uncertainty Simulator (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L5c","6",{}]] diff --git a/static-site/print/erus/__next._head.txt b/static-site/print/erus/__next._head.txt index 0e43fc466..4fd0959d4 100644 --- a/static-site/print/erus/__next._head.txt +++ b/static-site/print/erus/__next._head.txt @@ -1,6 +1,6 @@ 1:"$Sreact.fragment" -2:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -3:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +2:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +3:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 4:"$Sreact.suspense" -5:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +5:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 0:{"rsc":["$","$1","h",{"children":[null,["$","$L2",null,{"children":[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]]}],["$","div",null,{"hidden":true,"children":["$","$L3",null,{"children":["$","$4",null,{"name":"Next.Metadata","children":[["$","title","0",{"children":"The Evacuation Readiness and Uncertainty Simulator (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L5","6",{}]]}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/print/erus/__next._index.txt b/static-site/print/erus/__next._index.txt index f98a4fd90..18b9d82b6 100644 --- a/static-site/print/erus/__next._index.txt +++ b/static-site/print/erus/__next._index.txt @@ -1,13 +1,13 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -d:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +d:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} 8:["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$Ld",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$Ld",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$Ld",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$Ld",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] diff --git a/static-site/print/erus/__next._tree.txt b/static-site/print/erus/__next._tree.txt index 55701894f..ed0bb3dec 100644 --- a/static-site/print/erus/__next._tree.txt +++ b/static-site/print/erus/__next._tree.txt @@ -1,7 +1,7 @@ -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] 0:{"tree":{"name":"","param":null,"prefetchHints":16,"slots":{"children":{"name":"print","param":null,"prefetchHints":0,"slots":{"children":{"name":"slug","param":{"type":"d","key":"erus","siblings":null},"prefetchHints":0,"slots":{"children":{"name":"__PAGE__","param":null,"prefetchHints":0,"slots":null}}}}}}},"staleTime":300,"buildId":"static-snapshot"} diff --git a/static-site/print/erus/__next.print.txt b/static-site/print/erus/__next.print.txt index 460afe242..fa617e997 100644 --- a/static-site/print/erus/__next.print.txt +++ b/static-site/print/erus/__next.print.txt @@ -1,4 +1,4 @@ 1:"$Sreact.fragment" -2:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -3:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +2:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +3:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] 0:{"rsc":["$","$1","c",{"children":[null,[["$","style",null,{"children":"\n /* The site is dark-first, and a dark color-scheme makes Chrome paint\n the printed sheet's margin area dark too. Paper is light. */\n html { color-scheme: light !important; }\n html, body { background: #ffffff !important; }\n body::after { content: none !important; }\n header, footer, .site-bg, .aura { display: none !important; }\n "}],["$","$L2",null,{"parallelRouterKey":"children","template":["$","$L3",null,{}]}]]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/print/erus/index.html b/static-site/print/erus/index.html index 4bd16e2b8..8269f1136 100644 --- a/static-site/print/erus/index.html +++ b/static-site/print/erus/index.html @@ -1,4 +1,4 @@ -The Evacuation Readiness and Uncertainty Simulator (print edition) · NYU Ethical Tech CoLab

Publications · Academic report

The Evacuation Readiness and Uncertainty Simulator

Showing How Poor Field Information Degrades Evacuation Decisions

Ethical Tech CoLab

NYU Center for Global Affairs

July 2026

India Clarke. Developed as part of doctoral research on civilian protection in armed conflict, under the Ethical Tech CoLab at the NYU Center for Global Affairs.

When a humanitarian organisation decides where to move civilians fleeing conflict, it is rarely choosing between a good option and a bad one. It is choosing between several options it does not fully understand. ERUS exists to make one thing visible: the same underlying reality, assessed under worse information, produces materially worse predicted outcomes. Move the uncertainty slider and predicted success falls while nothing on the ground has changed at all.

Key figures

500

Monte Carlo trials run for every pairing of an evacuee group with a destination

20%

readiness cap on any destination with a blocked gatekeeper, which sits below the forty per cent success floor

0.30

score given to an unassessed factor, set deliberately below the 0.50 given to a partial one

4

values that fully determine every output: seed, destination count, group count, and uncertainty level

00

Foreword

When a humanitarian organisation decides where to move a group of civilians fleeing conflict, it is rarely choosing between a good option and a bad one. It is choosing between several options it does not fully understand. Is the camp forty kilometres away still secure, or was that report three weeks old? Has the district authority actually agreed to receive people, or did someone assume it had? Is the host community willing, or has nobody asked?

The situation on the ground is whatever it is. What changes, hour by hour, is how much of it the people making the decision can actually see. This report concerns a piece of software built to make that second problem visible: to show that the quality of a decision can collapse while nothing on the ground has changed at all, purely because the information environment has deteriorated.

A note on the name. The repository is called India-EvacSimulation because its author is named India Clarke. It does not concern the country of India, Indian disaster management, or the National Disaster Management Authority. The tool is about civilian evacuation in armed conflict, and it uses no real place names of any kind.

01

Executive Summary

The Evacuation Readiness and Uncertainty Simulator, referred to here as ERUS, is a research instrument that runs as a single web page in an ordinary browser. It invents a set of possible evacuation destinations and a set of groups of people needing to move, scores how ready each destination is, and works out which group should go where.

Every destination, every group, and every piece of information about them is synthetic. Nothing in the tool is a real place, a real population, or a real dataset. This is deliberate and is stated repeatedly in the project's own documentation. The tool is not making claims about any actual conflict.

The central argument the tool exists to make legible can be stated in one sentence: the same underlying reality, assessed under worse information, produces materially worse predicted outcomes. The tool provides a single slider labelled Field Uncertainty. Moving it does not change any fact about any destination. It changes only how much confidence the assessment carries. Predicted success rates fall anyway. That gap between what is true and what is knowable is what the project is about.

One thing has to be said plainly about that, because a reader could otherwise mistake it for a finding. The decline is guaranteed by the arithmetic. Uncertainty enters the model at exactly one point, by scaling each factor's confidence, and a factor's score increases monotonically with that confidence. Raising uncertainty therefore lowers every unblocked factor's score, which lowers mean readiness, which pushes more trials below the success floor. The downward curve is an identity of the scoring rule, not something the tool could have failed to produce. The tool does not demonstrate that uncertainty degrades decisions. It operationalises an assumed relationship and lets a user inspect its shape and magnitude under stated parameters. What is genuinely not guaranteed, and is therefore where the interest actually lies, is the relative steepness of the curve across different group types and the ranking of which unknowns cost the most to leave unresolved.

Destinations are scored across seven factors. Three of them are treated as gatekeepers: Security, Authority consent, and Willingness. Any one of these being confirmed as blocked caps the destination's overall readiness at twenty per cent, regardless of how good everything else is. The remaining four factors, being Capacity, Shelter, Food and water, and Medical capacity, reduce a score proportionally rather than capping it.

Because each factor is uncertain, the tool does not produce a single answer. It re-runs each pairing of a group and a destination five hundred times, each time randomly nudging factor assessments in proportion to how unreliable they are, and reports the proportion of runs in which the move would have succeeded. This is a standard technique known as Monte Carlo simulation.

The model draws a distinction that most scoring systems collapse. A destination whose willingness to receive people has simply never been assessed is recorded as Unknown. A host community that has explicitly refused is recorded as Unwilling. The first is a gap that better field work could close. The second is a settled answer that no amount of further enquiry will change. Treating them identically would make a refusal look like a solvable research problem.

The software itself works and is publicly demonstrated. The project's own backlog is candid that no parameter has been calibrated against real field data. An earlier version of this report noted that the document the project named as its authoritative methodology was not actually present in the repository. That has since been fixed by replacing it with a version-controlled methodology document, checked directly against the code.

02

Background and Rationale

The problem. Humanitarian actors working in active conflict do not operate with complete information. They operate with fragments: a report from a partner agency, an assessment carried out before the front line moved, a phone call that did not connect. The project's own concept note describes incomplete, unreliable, or absent field intelligence as the normal operating condition rather than an exceptional one.

The consequence. When the information is poor, the decision is still made. Someone still has to say which convoy goes where. The danger is not that a decision-maker knows they are guessing. The danger is that a planning tool presents an estimate with the same visual confidence whether that estimate rests on a verified assessment from yesterday or an assumption from last month.

The gap this addresses. Readiness scoring in humanitarian planning tends to produce a single number per site. That number does not usually carry any indication of how fragile it is. Two sites can score identically while one score is well founded and the other is largely inference.

The response. ERUS separates two things that are normally merged: what a factor is assessed to be, and how confident the assessor is in that assessment. Every factor in the tool carries both. The scoring then combines them, so that a well-evidenced partial assessment can outscore a poorly-evidenced good one. This is the mechanical heart of the tool.

The stated research questions. The project sets itself three. How does multi-factor readiness translate into a probable outcome? How does degraded field intelligence propagate through the chain of an evacuation decision? And where would real-time, machine-assisted assessment provide the greatest gain in decision quality? The third question is the one the tool's Factor Information Value panel is built to answer.

03

Objectives

The simulator sets itself six objectives:

  1. Make the propagation of uncertainty through an evacuation decision visible and measurable, rather than asserted.
  2. Express readiness as a distribution of possible outcomes rather than as a single point estimate that conceals its own fragility.
  3. Encode the fact that certain conditions are not tradeable. A site under active threat cannot be redeemed by an excellent food supply, and the arithmetic of the model must refuse to allow that trade.
  4. Keep the distinction between an unresolved question and a settled refusal, so that an unresolvable exclusion is never presented as an intelligence gap.
  5. Identify which specific unknowns, if resolved, would most improve the predicted outcome, so that scarce assessment capacity can be pointed at the questions that matter.
  6. Make every result exactly reproducible by anyone who has four numbers, so that a claim made from the tool can be independently regenerated and checked.

04

How the Simulator Works

The tool runs the same sequence of six steps every time a scenario is generated or the uncertainty slider is moved.

Step one, the seeded number generator. Every random value in the tool comes from a single arithmetic recipe that produces a stream of numbers which look random but are entirely determined by a starting number, called the seed. The recipe used is the Park and Miller generator, a well-documented method published in 1988. The practical consequence is that the same seed always produces the same scenario, on any computer, at any time. The tool never uses the browser's own randomness, which would not be repeatable.

Step two, inventing the scenario. The generator builds a chosen number of destinations, each with seven factor assessments, a capacity, and a distance, and a chosen number of evacuee groups drawn from eight fixed population types. This happens once, when a scenario is created. It does not re-run when the uncertainty slider moves, which is what allows the tool to prove that a falling success rate is caused by information quality alone and not by a changed scenario.

Step three, scoring readiness. Each destination is reduced to a single readiness figure between zero and one hundred per cent, using the weighted formula set out in the following section.

Step four, the Monte Carlo engine. For each possible pairing of a group with a destination, the tool runs five hundred trials. In each trial it independently considers every factor, asks whether an assessment that unreliable might be wrong, and if so shifts it one step better or worse. It then re-scores readiness. The output is a success rate and a standard deviation, which is a measure of how widely the five hundred results were spread.

Step five, assignment. Groups are sorted by urgency, with immediate cases first, then urgent, then those who can wait. Each group in turn is given the highest-scoring destination that still has enough spare capacity for it. Capacity is decremented as groups are placed, so later groups genuinely compete for what earlier groups did not take. If no site with room remains, the group is marked unassigned rather than being quietly given a place that does not exist.

Step six, display. The results are drawn as destination cards, an assignment matrix, an outcome table, an alerts panel, a sensitivity chart, and the information-value panel.

05

The Variables Explained

This section is the substance of the report. Every number the tool uses is listed below in plain terms: what it represents, why it was set where it was set, and how it reaches the result.

FactorWhat it asksRole
SecurityWhether the site is under threatGatekeeper
Authority consentWhether the relevant host authority has agreed to the movementGatekeeper
WillingnessWhether the host community itself is prepared to receive peopleGatekeeper
CapacityWhether there is roomStandard
ShelterWhether there is somewhere to sleepStandard
Food and waterWhether basic provisions existStandard
Medical capacityWhether there is clinical careStandard
The seven factors by which each destination is described. The three gatekeepers are marked as such in the code.

The three gatekeepers were chosen because each represents a condition that cannot be compensated for by anything else. The project's concept note states the reasoning plainly for Security: a site under active threat cannot be made viable by good food supply. Willingness was originally an ordinary factor and was promoted to gatekeeper status during development, precisely because as an ordinary factor it allowed strong shelter, food, and medical scores to arithmetically outweigh an outright refusal.

StatusScoreMeaning
Operational1.0The factor is in good order
Partial0.5The factor is degraded but functioning
Blocked0.0The factor has failed
Unknown0.30 before any adjustmentThe factor has not been assessed
Every factor, at every destination, holds one of four statuses, which convert to numbers as follows.

The choice of 0.30 for Unknown is the model's stance on ignorance. It is deliberately below Partial. An unassessed factor is not treated as probably fine. It is treated as more likely to be a problem than not, which the documentation calls epistemic conservatism, meaning caution about what one does not know.

Confidence, and why it is applied twice. Each factor also carries a base confidence, a number between zero and one describing how well that particular thing was assessed. The generator assigns it by status. The logic is that a field team reporting a functioning clinic has usually seen it, whereas a team reporting that they do not know has by definition seen less.

StatusBase confidence
Unknown0.05 to 0.50
Blocked0.40 to 0.75
Operational or Partial0.60 to 0.95
Base confidence assigned by status.

Separately, the Field Uncertainty slider produces a single confidence multiplier applied to every factor at once. If the slider reads thirty per cent uncertainty, the multiplier is 0.70, and every factor's confidence is reduced to seventy per cent of what it was. This is the whole-environment term: how degraded is the information picture right now, across the board. The two multiply together to give an effective confidence for each factor. This is the pairing the concept note calls uncertainty twice over.

How a factor score is actually computed. A factor's contribution is its status value scaled by a confidence adjustment. The adjustment is not the raw confidence but 0.5 plus half of the effective confidence. In plain terms, a factor never loses more than half its value to uncertainty. A destination reported as Operational with perfect confidence contributes its full 1.0. The same destination reported as Operational with no confidence at all contributes 0.5, the same as a confidently-reported Partial. Blocked is the exception. It scores zero regardless of confidence, on the reasoning that a reported blockage should not be discounted merely because the report is shaky. That does not make a Blocked factor's confidence value dead weight: it still feeds the perturbation probability described further down, where a lower confidence makes a trial more likely to flip that factor to a different status.

The weights and the gatekeeper cap. Gatekeeper factors carry a weight of 2. Standard factors carry a weight of 1. Readiness is the weighted average of the seven factor scores, so the three gatekeepers between them account for six of the ten weight units, which is to say sixty per cent of the ordinary score.

On top of that average sits the hard rule. If any gatekeeper is Blocked, readiness is capped at twenty per cent no matter what the average said. This is what makes the gatekeepers non-substitutable. The weighting alone would still have permitted a very strong performance elsewhere to lift a site with a blocked gatekeeper above a mediocre but unblocked one. The cap forbids it.

Destination capacity and distance. Capacity is generated between 200 and 5,000 places. It is not drawn evenly across that range but from an exponential distribution, which produces many small values and few large ones. The stated reason is that this matches the real shape of humanitarian site provision: many small sites and a few large camps. Distance is generated between 20 and 400 kilometres. The documentation is explicit that distance is a stand-in for operational burden, and that real routing constraints such as road condition, checkpoints, and fuel are not modelled. It is the only route variable in the tool.

The evacuee groups. Groups are drawn from eight fixed population types, described in the methodology as grounded in field taxonomy and in the categories of persons afforded specific protection under international humanitarian law. Only the size of a group is randomised, between 50 and 2,000 people. Each type carries three fixed properties.

GroupVulnerabilitySpecial needUrgency
Elderly and mobility-impaired5MobilityImmediate
Wounded and medical cases5MedicalImmediate
Unaccompanied minors5MedicalImmediate
Families with children4NoneUrgent
Pregnant women4MedicalUrgent
Journalists and aid workers2NoneUrgent
Unaccompanied adults2NoneCan wait
Mixed general population2NoneCan wait
The eight evacuee archetypes and their fixed properties.

Vulnerability runs on a five-point scale. Urgency determines the order in which groups are assigned, which matters because assignment is sequential and capacity runs out. The special-needs marker determines whether a destination's medical provision is treated as decisive for that group.

The composite score, and its four weights. Readiness alone does not decide where a group goes. The tool ranks destinations for each group using a composite score built from four parts.

ComponentWeightWhat it measures
Readiness0.40The quality of protection at the site
Capacity fit0.30The site's capacity divided by the group's size, capped at 1.0, since a site twice the size of the group scores the same as one ten times its size and surplus beyond sufficiency adds nothing
Proximity0.20One minus the distance divided by the 400 kilometre maximum, so a nearby site scores near 1.0 and the furthest scores near zero
Vulnerability match0.10A binary: 1.0 if the group has a special need and the site's medical capacity is Operational, and 0.5 in every other case

The in-app methodology states the ordering these weights encode: protection quality first, physical fit second, operational burden third, population needs fourth.

One observation worth recording. Because vulnerability match is checked only against medical capacity, a group whose stated need is mobility rather than medical is rewarded by the presence of a clinic. The mobility need has no separate representation anywhere in the scoring. The obvious remedy is a second, independent accessibility factor, generated and scored the same way as the existing seven and checked against a mobility need the way medical capacity is checked against a medical one. That has not been built yet, since it would change the balance between gatekeeper and standard factors and the shape of every generated destination, which is a modelling decision rather than a small fix.

TermWhat it measures
ReadinessA single destination's own quality, independent of any particular group
Composite scoreHow well a destination fits one specific group, blending readiness with capacity fit, proximity, and vulnerability match
Success rateThe predicted outcome of assigning one group to one destination, the share of five hundred trials that clear the readiness, capacity, and gatekeeper conditions at once
Three terms this report uses that are easy to conflate, since all three describe how good something is at a different point in the pipeline.

A destination can rank first by composite score for a group and still have a low success rate, because composite score blends readiness with logistics terms that a hard gatekeeper cap does not touch. The worked example below shows this happening.

Three numbers govern the simulation of uncertainty itself:

  • Runs, set to 500 per group and destination pair. This is how many times each pairing is tested. More runs give a tighter estimate at the cost of speed. The information-value panel originally used a cheaper 100 runs, but that count produced estimates too noisy to trust, so it now uses the same 500 runs as everywhere else in the tool.
  • Maximum perturbation probability, set to 0.85. This is the ceiling on how likely a factor's assessment is to be wrong in a given trial, so that at an effective confidence of zero a factor still has a fifteen per cent chance of being reported correctly. The tool never assumes information is entirely worthless.
  • The perturbation step, fixed at one level. When a factor is judged to be misreported it moves exactly one place better or worse along the sequence, with an even chance of each direction. The documentation acknowledges this as a simplification, noting that a real mis-assessment could jump multiple levels, such as an Operational site being reported as Blocked.

Factors already recorded as Unknown are never perturbed. Uncertainty about an unknown is already expressed in its low score and low confidence.

The success threshold. A single trial counts as a success only if three conditions hold together: readiness reaches at least forty per cent, the destination's capacity is at least the group's size, and no gatekeeper is blocked in that trial. An earlier version of the tool's own source table attributed the forty per cent floor to the UNHCR Handbook for Emergencies. That attribution has since been withdrawn as incorrect: readiness is a construct internal to this tool, a weighted average of seven invented factor scores, so no external handbook can define a minimum readiness threshold for it. The floor is now labelled what it actually is, an unsourced modelling assumption, uncalibrated and chosen by the authors. By contrast the two Sphere figures cited later, fifteen litres of water per person per day and 3.5 square metres of covered space, are accurate and are correctly described as informing which factors exist rather than validating any weight.

Note that a destination whose gatekeeper is blocked is capped at twenty per cent readiness, which is below the forty per cent floor. The cap and the threshold are therefore consistent by construction: a blocked gatekeeper cannot produce a successful trial.

06

Reading the Results

Destination cards. Each site shows its readiness percentage and its seven factor statuses. Where a gatekeeper is blocked, the card names which one rather than showing a general warning, so that three distinct failure modes are never collapsed into a single badge.

The assignment matrix. A grid of every group against every destination, showing the composite scores that produced the ranking.

Outcomes. For each group, the assigned destination, the alternative if one exists, the predicted success rate from the five hundred trials, and the standard deviation. A high success rate with a high standard deviation should be read differently from the same rate with a low one: the first is a fragile prediction, the second a stable one.

The riskiest factor. For each assignment the tool identifies the single lowest-scoring factor at the assigned destination, which is the thing most worth checking before moving anyone.

The uncertainty sensitivity chart. For each group, predicted success rate is plotted at eleven uncertainty levels from zero to one hundred per cent. This is the tool's core demonstration in a single picture: the underlying scenario is identical at every point on the line, and only the information quality differs.

The Factor Information Value panel. For each of the seven factors, the tool asks a counterfactual question: if every Unknown instance of this factor were resolved to Operational, how much would the average success rate improve? It answers by genuinely re-running the assignment and the simulation with that change applied, not by estimating. The factors are then ranked by improvement. In practical terms this tells an assessment team which single question is worth the trip.

Alerts. Warnings surface unassigned groups and destinations excluded by a named gatekeeper.

A worked example. Everything above describes what the tool computes. This report also ran it, at the exact citation the reproducibility section recommends: seed 42, eight destinations, three groups, thirty per cent uncertainty. Anyone can regenerate the same scenario at that address on the live site. Of the eight generated destinations, five carry a blocked gatekeeper and are capped at twenty per cent readiness, matching what the project's backlog reports for this seed. The three evacuee groups drawn are Unaccompanied minors, 632 people, immediate; Elderly and mobility-impaired, 378 people, immediate; and Mixed general population, 483 people, can wait.

GroupAssigned destinationComposite scoreReadinessPredicted successRiskiest factor
Unaccompanied minorsZone Golf66.7%20.0%, gatekeeper capped20.2%Willingness, Unwilling
Elderly and mobility-impairedStation Hotel72.4%47.1%48.0%Capacity, Unknown
Mixed general populationZone Golf61.7%20.0%, gatekeeper capped17.8%Willingness, Unwilling
The assignment this scenario actually produces.

Two of the three groups were sent to Zone Golf, the one destination in the pool whose Willingness gatekeeper is blocked. This is what the weighting in the composite score does when a capped but close and roomy site outranks a viable but farther one. Zone Golf sits 26 kilometres away with ample spare capacity, so its capacity fit and proximity terms offset its capped readiness in the blended score. The alternative available to the minors group scored lower on the composite, 63.3% against Zone Golf's 66.7%, but carried a predicted success rate of 49%, more than double. The alternative available to the mixed population group scored 57.4% but carried a success rate of 73.2%, roughly four times higher.

This is the answer this report can now give to the first of the project's three research questions. Multi-factor readiness does not translate into a probable outcome by simple ranking. The composite score the tool actually assigns on blends readiness with logistics terms that can outweigh a hard readiness cap, so the top ranked destination and the destination most likely to succeed are not always the same place, and in this scenario were not the same place for two of the three groups.

The same scenario at higher uncertainty. Re-running the identical destinations and groups, nothing regenerated, at eighty per cent uncertainty instead of thirty gives a clean test of the second research question.

GroupAssigned destinationSuccess at 30% uncertaintySuccess at 80% uncertaintyFall
Unaccompanied minorsZone Golf, unchanged20.2%12.6%38% relative
Elderly and mobility-impairedStation Hotel, unchanged48.0%14.8%69% relative
Mixed general populationZone Golf, unchanged17.8%13.8%22% relative
Predicted success for the same three groups at two uncertainty levels.

The assigned destination for every group is identical at both uncertainty levels. Nothing about the scenario moved. Yet predicted success fell for all three, by very different amounts. Elderly and mobility-impaired lost more than two thirds of its predicted success, Mixed general population lost about a fifth. The unevenness, not just the direction, is the finding the project's earlier peer review said the report was missing.

The sensitivity chart, corrected. The chart plots predicted success against uncertainty from zero to one hundred per cent by re-selecting whichever destination currently scores highest at each of eleven levels, rather than holding one destination fixed. An earlier pass through this exact scenario found that this re-selection step could pick a destination whose capacity was too small for the group, which forced a flat, misleading zero per cent success rate at low uncertainty for one group before the chart jumped once a viable destination became the top scorer. That has since been fixed: the chart now applies the same capacity check the actual assignment step already used. The corrected curve for that group is noisier than a clean decline, wobbling within the range ordinary Monte Carlo variance produces at 500 runs, but the systematic jump is gone.

The Factor Information Value ranking, corrected. For each factor, the panel asks what would happen to mean predicted success if every Unknown instance of that factor were resolved to Operational. Run at its original 100 trial count against this exact scenario, three of the seven factors showed a negative estimated gain, as large as almost seven percentage points, which is not possible in the underlying model, since resolving a factor toward Operational can only raise or hold its score, never lower it. The panel's trial count has since been raised to match the tool's main 500 run engine, and every delta in this scenario is now within about a percentage point of zero. The corrected finding is smaller but more honest: in this particular scenario, none of the seven factors' unresolved instances would move predicted success by more than about a point, and the panel can now say so with reasonable confidence instead of implying an effect that was sampling noise.

Answering the three research questions. Read together, the worked example above does what the project's earlier peer review said the report needed to do. It answers, rather than restates, all three questions the project sets itself. How does multi-factor readiness translate into a probable outcome. Not by simple ranking, since the composite score can send a group to a hard-capped destination over a viable alternative with several times its predicted success. How does degraded field intelligence propagate through an evacuation decision. Unevenly, with the same fixed scenario losing between a fifth and two thirds of its predicted success across three different groups, depending on which factor was driving that group's score. Where would real-time assessment help most. At this scenario's default run count, the honest answer is that the difference is not resolvable from noise, which is itself a finding about how much assessment volume the tool's own panel needs before its ranking should be acted on.

07

The Two Distinctions the Model Insists On

Known unknown against unknown unknown. A factor recorded as Unknown with high confidence, meaning the team is sure it does not know, scores higher than the same status with low confidence, meaning the team is not even sure its ignorance is accurate. This falls directly out of the confidence adjustment. The reasoning offered is that a field team which explicitly flags a gap is providing more actionable information than one whose own reporting cannot be trusted.

Unknown against Unwilling. When the Willingness factor is Blocked, the tool displays the word Unwilling rather than the generic Blocked, throughout the interface and in generated text. The point is not cosmetic. Unknown is an intelligence gap that better assessment can close. Unwilling is a confirmed exclusion that no further enquiry will change. A tool whose purpose is to identify where more information would help must not present a settled refusal as an open question, or it commits the very error it exists to expose.

The project's backlog is honest that this change had consequences. Because gatekeeper factors are generated with a higher probability of being blocked than standard factors, promoting Willingness to gatekeeper status raised the number of excluded destinations in the default scenario from roughly two in eight to roughly five in eight. That rate has since been checked across 500 different seeds rather than just the one the project cites: the mean exclusion count is 4.48 of 8, with 5 in 8 sitting inside the two most common outcomes rather than at an extreme, confirming the rate is the model's typical behaviour rather than an artefact of that particular seed.

08

Grounding in Humanitarian Standards and Law

The tool's own source table is explicit about a boundary that matters. No real dataset is used anywhere. The sources listed inform the design of the model, meaning which factors exist, which are gatekeepers, and where thresholds come from. They do not supply data.

The sources the tool names are as follows.

  • The Sphere Handbook, fourth edition, 2018, behind the Shelter and Food and water factors. Sphere is a voluntary set of minimum standards for humanitarian response, developed by a coalition of humanitarian organisations. The two figures the tool cites from it are accurate: a minimum of 15 litres of water per person per day, and 3.5 square metres of covered living space per person.
  • The UNHCR Handbook for Emergencies, third edition, 2007. An earlier version of the tool cited this as the origin of the forty per cent minimum readiness threshold. That attribution has since been withdrawn as incorrect, since readiness is a construct internal to the tool. The handbook is retained for what it does support, informing which readiness factors are worth modelling, and the threshold itself is treated in this report as an unsourced modelling assumption.
  • Inter-Agency Standing Committee guidance from 2007, cited behind the Authority consent gatekeeper. The Inter-Agency Standing Committee is the main coordination forum of the United Nations and non-governmental humanitarian system. Its guidance in this area establishes that humanitarian operations require the consent of the host government, which is the principle the gatekeeper encodes.
  • Additional Protocol I to the Geneva Conventions, 1977, Articles 12 and 58, behind the Medical capacity and Security factors. Article 12 protects medical units from attack. Article 58 requires parties to a conflict to take precautions to protect civilians under their own control from the effects of attacks.
  • An ICRC publication from 2013 on violence and the use of force, cited as grounding the Security gatekeeper. The ICRC is the International Committee of the Red Cross, the body with a specific mandate under the Geneva Conventions.
  • Park and Miller, 1988, for the random number generator, which is a computing citation rather than a humanitarian one.

A reader should treat these citations as the origin of design choices rather than as validation of the numbers. The tool does not claim that Sphere endorses a weight of 2 for gatekeepers, and it should not be read as claiming so.

09

Reproducibility

Every output of the tool is fully determined by four values: the seed, the number of destinations, the number of evacuee groups, and the uncertainty level. Nothing else varies.

The tool writes those four values into the web address of the page. A Copy link button places that address on the clipboard. Anyone who opens that address regenerates the identical scenario, on any machine. The project recommends citing results in the form seed 42, eight destinations, three groups, thirty per cent uncertainty.

This is a genuine strength and is unusual in tools of this kind. A claim made from this simulator can be checked by a reviewer in a few seconds rather than taken on trust.

10

Relationship to Other Ethical Tech CoLab Projects

Three projects in the same organisation address civilian evacuation and share a common documentation pattern of a concept note, an architecture note, and a backlog. They are separate pieces of software with different aims and no shared code.

ERCF, the Evacuation Risk and Cost Framework, is a server-based tool that estimates the human and financial cost of an evacuation and compares the one-off cost of moving people against the accumulating cost of sustaining them where they are. Unlike ERUS it uses real external data, including conflict event records and country-level severity indices, and a set of documented historical cases.

Evac-Sim-Melanie is an agent-based model of how individual households decide to leave, following a population through stages from unaware to departed, and modelling how alerts spread between neighbours.

The three occupy different levels. ERCF asks what an evacuation would cost. Evac-Sim-Melanie asks how a population behaves once told to move. ERUS asks a narrower and prior question: given a set of possible destinations, how much does the quality of our information about them determine the outcome? Readers interested in any of the others should consult those repositories directly, as this report describes only ERUS.

11

Practical Nature and Maturity of the Tool

What is built and works. The simulator is complete and functional. It is one self-contained web page with no installation, no build step, and no external dependencies, which is a deliberate choice so that it will still run years from now without a toolchain to maintain. It is published as a live public demonstration through GitHub Pages. All of the mechanics described above are implemented in the code and were verified against it for this report.

What was not built, and has since been added. Earlier versions of this report and of the project's own backlog stated there were no automated tests and no licence file. Both gaps have since been closed. The scoring, perturbation, simulation, and assignment functions now live in a standalone engine file, with a regression suite that pins known seed to known output results, including the exact figures in the worked example above. The repository carries an MIT licence. What remains genuinely absent is calibration against real data.

A documentation problem, now resolved. Earlier versions of this report and of the README named a Word document as the authoritative source for formula derivations and citations, to be preferred over the in-app text if the two ever disagreed. That document was excluded from version control, so a reader who obtained a copy of the repository could not consult the document the repository told them was authoritative, and it was separately behind the code in several respects, including the number of simulation runs and the promotion of Willingness to gatekeeper status. This has been fixed by retiring that document and replacing it with a version-controlled methodology file that makes no claim to being auto-generated, is checked directly against the code, and carries its own changelog recording exactly what it corrects relative to the old one.

Which claims in this report can be checked, and which cannot. The distinction matters more than it might appear, because it is not uniform across the report. Everything describing what the tool computes was read off the source: the scoring rule, the confidence adjustment, the gatekeeper cap, the perturbation probability, the composite weights, the success conditions, and the seeded generator. A reader with the repository open can verify all of it. Everything explaining why a parameter takes the value it does rests on the project documents, and the authoritative one among them, the project's methodology document, is now version-controlled and can be consulted directly rather than taken on trust. That includes the reasoning for promoting Willingness to gatekeeper status, the field taxonomy said to ground the eight population types, and the derivation of the 0.30 score for Unknown.

History of drift. The backlog records that two copies of the tool existed in this repository for a period, with new features landing only in the one that was not being published, so that the public demonstration was stale. This has been resolved by making one file canonical and turning the other into a redirect. It is recorded here because it illustrates the class of problem the backlog is mostly concerned with: the code is sound, and the things around the code have repeatedly fallen out of step with it.

12

Limitations and Caveats

Nothing is calibrated. No parameter in the tool has been fitted to or tested against empirical field data. The forty per cent success threshold, the 0.85 perturbation ceiling, the 0.30 score for Unknown, the twenty per cent gatekeeper cap, the doubling of gatekeeper weights, and the four composite weights are all modelling assumptions. The project describes them as open to challenge and says so repeatedly.

Factors are treated as independent. In reality security and authority consent plausibly move together, since a deteriorating security situation often accompanies a withdrawal of consent. The model assumes no such relationship, which will tend to understate how badly things fail together. The project identifies adding factor correlation as the highest-value next step for realism.

There is no time. The model is a single snapshot. Nothing deteriorates or improves over the course of an evacuation, and no journey takes any time to complete.

Routes are represented by one number. Distance stands in for the entire operational burden of a journey. Road condition, checkpoints, fuel availability, seasonal access, and the presence of mines are not modelled at all. A reader should not take the proximity term as a statement about whether a route is passable.

The assignment is greedy, not optimal. Groups are placed one at a time in urgency order, each taking the best remaining site. This is realistic in that urgent cases are handled first, but it means an early group can take a site that would have been a much better match for a later one. The tool does not search for the best overall allocation.

Mis-assessment is gentle. A factor can only be wrong by one level. Real reporting failures can be larger, and the documentation says so.

The population model is coarse. Eight archetypes with fixed vulnerability values, of which only three distinct values are used, and a single binary check against medical capacity. Group composition, disability other than mobility, language, and legal status are not represented. The specific fix for the mobility case, a dedicated accessibility factor scored the way medical capacity is scored today, is a proposed but not yet made change, since it would shift the balance between gatekeeper and standard factors.

The outputs mean nothing about the real world. This follows from the fact that the inputs are invented. The tool demonstrates a relationship between information quality and decision quality. It does not predict, and cannot predict, the outcome of any actual evacuation.

The tool's own statement of its status. It is unambiguous and should be respected: a conceptual demonstration tool for thesis research, not an operational decision-support system, whose outputs require empirical calibration before any real-world use.

13

Intended Audience and Use

The tool is built for an academic argument. Its stated purpose is to make an uncertainty-propagation argument legible to thesis examiners and stakeholders. It is well suited to that, and to teaching, where the sensitivity chart makes an abstract point about information quality concrete in a way that prose does not.

It has a secondary use in prioritising assessment effort. The information-value panel answers a question humanitarian teams face regularly, which is which of several unanswered questions is worth the journey to answer. Even with invented data, the structure of that reasoning is transferable.

It is not for planning. No part of this tool should inform a decision about moving actual people. The project says this itself, in every document it contains, and this report repeats it without qualification.

14

Conclusion

The most useful thing about this simulator is a relationship it makes easy to inspect. Move the uncertainty slider and predicted outcomes deteriorate while the scenario behind them is provably unchanged, because the scenario is regenerated only on demand and not when the slider moves. The degradation is entirely an artefact of what is knowable. That the direction of the effect is built into the scoring rule rather than discovered by it is stated above and should not be forgotten; what the tool contributes is the magnitude, the differing slopes across population types, and the ranking of which unknowns are worth resolving first, shown with real numbers in the worked example above. Making those visible, and reproducible from a seed, is a more honest contribution than a tool that produced confident numbers would be.

The two epistemic distinctions are the other substantive contribution. Separating a confident admission of ignorance from an unreliable one, and separating an open question from a settled refusal, are both choices that a simpler scoring system would have flattened. The decision to promote Willingness to a gatekeeper, and to name it Unwilling rather than Blocked, came from noticing that the arithmetic was allowing a good food supply to outvote a community's refusal. That is the kind of error that is easy to make and hard to see once made.

The candour of the project's own backlog is worth noting as a practice in itself. Across this project's history it has recorded an absent methodology document, since replaced with a version-controlled one; a stale published build, since fixed; previously missing tests, since added; and the possibility that its own most recent modelling change produced an exclusion rate that is an artefact rather than an intention, since checked across 500 seeds and confirmed as typical rather than artefactual. What remains open is calibration against real data, which the backlog names honestly as the one gap tests and documentation cannot close. A research instrument that catalogues its own weaknesses this specifically, and then closes them out in the same place it recorded them, is easier to trust about what it still claims.

What remains, before any of this could be more than a demonstration, is calibration. Every threshold in the tool is a considered guess. The tool is transparent about which guesses they are and where each one lives in the code, which is the necessary precondition for someone eventually replacing them with measurements.

References

Sources

  1. 01Sphere Association. The Sphere Handbook: Humanitarian Charter and Minimum Standards in Humanitarian Response, fourth edition, 2018. Behind the Shelter and Food and water factors.
  2. 02United Nations High Commissioner for Refugees. Handbook for Emergencies, third edition, 2007. An earlier version of the tool's documentation cited this as the origin of the forty per cent minimum readiness threshold. That attribution has been withdrawn. The handbook is retained here for informing which readiness factors are worth modelling.
  3. 03Inter-Agency Standing Committee. Guidance on humanitarian operations requiring the consent of the host government, 2007. Behind the Authority consent gatekeeper.
  4. 04Protocol Additional to the Geneva Conventions of 12 August 1949, and relating to the Protection of Victims of International Armed Conflicts (Protocol I), 1977, Articles 12 and 58. Behind the Medical capacity and Security factors.
  5. 05International Committee of the Red Cross. Publication on violence and the use of force, 2013. Cited as grounding the Security gatekeeper.
  6. 06Park, S.K. and Miller, K.W. (1988). Random Number Generators: Good Ones Are Hard To Find. Communications of the ACM, volume 31, issue 10. The generator used for all seeded randomness in the tool.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file +

Publications · Academic report

The Evacuation Readiness and Uncertainty Simulator

Showing How Poor Field Information Degrades Evacuation Decisions

Ethical Tech CoLab

NYU Center for Global Affairs

July 2026

India Clarke. Developed as part of doctoral research on civilian protection in armed conflict, under the Ethical Tech CoLab at the NYU Center for Global Affairs.

When a humanitarian organisation decides where to move civilians fleeing conflict, it is rarely choosing between a good option and a bad one. It is choosing between several options it does not fully understand. ERUS exists to make one thing visible: the same underlying reality, assessed under worse information, produces materially worse predicted outcomes. Move the uncertainty slider and predicted success falls while nothing on the ground has changed at all.

Key figures

500

Monte Carlo trials run for every pairing of an evacuee group with a destination

20%

readiness cap on any destination with a blocked gatekeeper, which sits below the forty per cent success floor

0.30

score given to an unassessed factor, set deliberately below the 0.50 given to a partial one

4

values that fully determine every output: seed, destination count, group count, and uncertainty level

00

Foreword

When a humanitarian organisation decides where to move a group of civilians fleeing conflict, it is rarely choosing between a good option and a bad one. It is choosing between several options it does not fully understand. Is the camp forty kilometres away still secure, or was that report three weeks old? Has the district authority actually agreed to receive people, or did someone assume it had? Is the host community willing, or has nobody asked?

The situation on the ground is whatever it is. What changes, hour by hour, is how much of it the people making the decision can actually see. This report concerns a piece of software built to make that second problem visible: to show that the quality of a decision can collapse while nothing on the ground has changed at all, purely because the information environment has deteriorated.

A note on the name. The repository is called India-EvacSimulation because its author is named India Clarke. It does not concern the country of India, Indian disaster management, or the National Disaster Management Authority. The tool is about civilian evacuation in armed conflict, and it uses no real place names of any kind.

01

Executive Summary

The Evacuation Readiness and Uncertainty Simulator, referred to here as ERUS, is a research instrument that runs as a single web page in an ordinary browser. It invents a set of possible evacuation destinations and a set of groups of people needing to move, scores how ready each destination is, and works out which group should go where.

Every destination, every group, and every piece of information about them is synthetic. Nothing in the tool is a real place, a real population, or a real dataset. This is deliberate and is stated repeatedly in the project's own documentation. The tool is not making claims about any actual conflict.

The central argument the tool exists to make legible can be stated in one sentence: the same underlying reality, assessed under worse information, produces materially worse predicted outcomes. The tool provides a single slider labelled Field Uncertainty. Moving it does not change any fact about any destination. It changes only how much confidence the assessment carries. Predicted success rates fall anyway. That gap between what is true and what is knowable is what the project is about.

One thing has to be said plainly about that, because a reader could otherwise mistake it for a finding. The decline is guaranteed by the arithmetic. Uncertainty enters the model at exactly one point, by scaling each factor's confidence, and a factor's score increases monotonically with that confidence. Raising uncertainty therefore lowers every unblocked factor's score, which lowers mean readiness, which pushes more trials below the success floor. The downward curve is an identity of the scoring rule, not something the tool could have failed to produce. The tool does not demonstrate that uncertainty degrades decisions. It operationalises an assumed relationship and lets a user inspect its shape and magnitude under stated parameters. What is genuinely not guaranteed, and is therefore where the interest actually lies, is the relative steepness of the curve across different group types and the ranking of which unknowns cost the most to leave unresolved.

Destinations are scored across seven factors. Three of them are treated as gatekeepers: Security, Authority consent, and Willingness. Any one of these being confirmed as blocked caps the destination's overall readiness at twenty per cent, regardless of how good everything else is. The remaining four factors, being Capacity, Shelter, Food and water, and Medical capacity, reduce a score proportionally rather than capping it.

Because each factor is uncertain, the tool does not produce a single answer. It re-runs each pairing of a group and a destination five hundred times, each time randomly nudging factor assessments in proportion to how unreliable they are, and reports the proportion of runs in which the move would have succeeded. This is a standard technique known as Monte Carlo simulation.

The model draws a distinction that most scoring systems collapse. A destination whose willingness to receive people has simply never been assessed is recorded as Unknown. A host community that has explicitly refused is recorded as Unwilling. The first is a gap that better field work could close. The second is a settled answer that no amount of further enquiry will change. Treating them identically would make a refusal look like a solvable research problem.

The software itself works and is publicly demonstrated. The project's own backlog is candid that no parameter has been calibrated against real field data. An earlier version of this report noted that the document the project named as its authoritative methodology was not actually present in the repository. That has since been fixed by replacing it with a version-controlled methodology document, checked directly against the code.

02

Background and Rationale

The problem. Humanitarian actors working in active conflict do not operate with complete information. They operate with fragments: a report from a partner agency, an assessment carried out before the front line moved, a phone call that did not connect. The project's own concept note describes incomplete, unreliable, or absent field intelligence as the normal operating condition rather than an exceptional one.

The consequence. When the information is poor, the decision is still made. Someone still has to say which convoy goes where. The danger is not that a decision-maker knows they are guessing. The danger is that a planning tool presents an estimate with the same visual confidence whether that estimate rests on a verified assessment from yesterday or an assumption from last month.

The gap this addresses. Readiness scoring in humanitarian planning tends to produce a single number per site. That number does not usually carry any indication of how fragile it is. Two sites can score identically while one score is well founded and the other is largely inference.

The response. ERUS separates two things that are normally merged: what a factor is assessed to be, and how confident the assessor is in that assessment. Every factor in the tool carries both. The scoring then combines them, so that a well-evidenced partial assessment can outscore a poorly-evidenced good one. This is the mechanical heart of the tool.

The stated research questions. The project sets itself three. How does multi-factor readiness translate into a probable outcome? How does degraded field intelligence propagate through the chain of an evacuation decision? And where would real-time, machine-assisted assessment provide the greatest gain in decision quality? The third question is the one the tool's Factor Information Value panel is built to answer.

03

Objectives

The simulator sets itself six objectives:

  1. Make the propagation of uncertainty through an evacuation decision visible and measurable, rather than asserted.
  2. Express readiness as a distribution of possible outcomes rather than as a single point estimate that conceals its own fragility.
  3. Encode the fact that certain conditions are not tradeable. A site under active threat cannot be redeemed by an excellent food supply, and the arithmetic of the model must refuse to allow that trade.
  4. Keep the distinction between an unresolved question and a settled refusal, so that an unresolvable exclusion is never presented as an intelligence gap.
  5. Identify which specific unknowns, if resolved, would most improve the predicted outcome, so that scarce assessment capacity can be pointed at the questions that matter.
  6. Make every result exactly reproducible by anyone who has four numbers, so that a claim made from the tool can be independently regenerated and checked.

04

How the Simulator Works

The tool runs the same sequence of six steps every time a scenario is generated or the uncertainty slider is moved.

Step one, the seeded number generator. Every random value in the tool comes from a single arithmetic recipe that produces a stream of numbers which look random but are entirely determined by a starting number, called the seed. The recipe used is the Park and Miller generator, a well-documented method published in 1988. The practical consequence is that the same seed always produces the same scenario, on any computer, at any time. The tool never uses the browser's own randomness, which would not be repeatable.

Step two, inventing the scenario. The generator builds a chosen number of destinations, each with seven factor assessments, a capacity, and a distance, and a chosen number of evacuee groups drawn from eight fixed population types. This happens once, when a scenario is created. It does not re-run when the uncertainty slider moves, which is what allows the tool to prove that a falling success rate is caused by information quality alone and not by a changed scenario.

Step three, scoring readiness. Each destination is reduced to a single readiness figure between zero and one hundred per cent, using the weighted formula set out in the following section.

Step four, the Monte Carlo engine. For each possible pairing of a group with a destination, the tool runs five hundred trials. In each trial it independently considers every factor, asks whether an assessment that unreliable might be wrong, and if so shifts it one step better or worse. It then re-scores readiness. The output is a success rate and a standard deviation, which is a measure of how widely the five hundred results were spread.

Step five, assignment. Groups are sorted by urgency, with immediate cases first, then urgent, then those who can wait. Each group in turn is given the highest-scoring destination that still has enough spare capacity for it. Capacity is decremented as groups are placed, so later groups genuinely compete for what earlier groups did not take. If no site with room remains, the group is marked unassigned rather than being quietly given a place that does not exist.

Step six, display. The results are drawn as destination cards, an assignment matrix, an outcome table, an alerts panel, a sensitivity chart, and the information-value panel.

05

The Variables Explained

This section is the substance of the report. Every number the tool uses is listed below in plain terms: what it represents, why it was set where it was set, and how it reaches the result.

FactorWhat it asksRole
SecurityWhether the site is under threatGatekeeper
Authority consentWhether the relevant host authority has agreed to the movementGatekeeper
WillingnessWhether the host community itself is prepared to receive peopleGatekeeper
CapacityWhether there is roomStandard
ShelterWhether there is somewhere to sleepStandard
Food and waterWhether basic provisions existStandard
Medical capacityWhether there is clinical careStandard
The seven factors by which each destination is described. The three gatekeepers are marked as such in the code.

The three gatekeepers were chosen because each represents a condition that cannot be compensated for by anything else. The project's concept note states the reasoning plainly for Security: a site under active threat cannot be made viable by good food supply. Willingness was originally an ordinary factor and was promoted to gatekeeper status during development, precisely because as an ordinary factor it allowed strong shelter, food, and medical scores to arithmetically outweigh an outright refusal.

StatusScoreMeaning
Operational1.0The factor is in good order
Partial0.5The factor is degraded but functioning
Blocked0.0The factor has failed
Unknown0.30 before any adjustmentThe factor has not been assessed
Every factor, at every destination, holds one of four statuses, which convert to numbers as follows.

The choice of 0.30 for Unknown is the model's stance on ignorance. It is deliberately below Partial. An unassessed factor is not treated as probably fine. It is treated as more likely to be a problem than not, which the documentation calls epistemic conservatism, meaning caution about what one does not know.

Confidence, and why it is applied twice. Each factor also carries a base confidence, a number between zero and one describing how well that particular thing was assessed. The generator assigns it by status. The logic is that a field team reporting a functioning clinic has usually seen it, whereas a team reporting that they do not know has by definition seen less.

StatusBase confidence
Unknown0.05 to 0.50
Blocked0.40 to 0.75
Operational or Partial0.60 to 0.95
Base confidence assigned by status.

Separately, the Field Uncertainty slider produces a single confidence multiplier applied to every factor at once. If the slider reads thirty per cent uncertainty, the multiplier is 0.70, and every factor's confidence is reduced to seventy per cent of what it was. This is the whole-environment term: how degraded is the information picture right now, across the board. The two multiply together to give an effective confidence for each factor. This is the pairing the concept note calls uncertainty twice over.

How a factor score is actually computed. A factor's contribution is its status value scaled by a confidence adjustment. The adjustment is not the raw confidence but 0.5 plus half of the effective confidence. In plain terms, a factor never loses more than half its value to uncertainty. A destination reported as Operational with perfect confidence contributes its full 1.0. The same destination reported as Operational with no confidence at all contributes 0.5, the same as a confidently-reported Partial. Blocked is the exception. It scores zero regardless of confidence, on the reasoning that a reported blockage should not be discounted merely because the report is shaky. That does not make a Blocked factor's confidence value dead weight: it still feeds the perturbation probability described further down, where a lower confidence makes a trial more likely to flip that factor to a different status.

The weights and the gatekeeper cap. Gatekeeper factors carry a weight of 2. Standard factors carry a weight of 1. Readiness is the weighted average of the seven factor scores, so the three gatekeepers between them account for six of the ten weight units, which is to say sixty per cent of the ordinary score.

On top of that average sits the hard rule. If any gatekeeper is Blocked, readiness is capped at twenty per cent no matter what the average said. This is what makes the gatekeepers non-substitutable. The weighting alone would still have permitted a very strong performance elsewhere to lift a site with a blocked gatekeeper above a mediocre but unblocked one. The cap forbids it.

Destination capacity and distance. Capacity is generated between 200 and 5,000 places. It is not drawn evenly across that range but from an exponential distribution, which produces many small values and few large ones. The stated reason is that this matches the real shape of humanitarian site provision: many small sites and a few large camps. Distance is generated between 20 and 400 kilometres. The documentation is explicit that distance is a stand-in for operational burden, and that real routing constraints such as road condition, checkpoints, and fuel are not modelled. It is the only route variable in the tool.

The evacuee groups. Groups are drawn from eight fixed population types, described in the methodology as grounded in field taxonomy and in the categories of persons afforded specific protection under international humanitarian law. Only the size of a group is randomised, between 50 and 2,000 people. Each type carries three fixed properties.

GroupVulnerabilitySpecial needUrgency
Elderly and mobility-impaired5MobilityImmediate
Wounded and medical cases5MedicalImmediate
Unaccompanied minors5MedicalImmediate
Families with children4NoneUrgent
Pregnant women4MedicalUrgent
Journalists and aid workers2NoneUrgent
Unaccompanied adults2NoneCan wait
Mixed general population2NoneCan wait
The eight evacuee archetypes and their fixed properties.

Vulnerability runs on a five-point scale. Urgency determines the order in which groups are assigned, which matters because assignment is sequential and capacity runs out. The special-needs marker determines whether a destination's medical provision is treated as decisive for that group.

The composite score, and its four weights. Readiness alone does not decide where a group goes. The tool ranks destinations for each group using a composite score built from four parts.

ComponentWeightWhat it measures
Readiness0.40The quality of protection at the site
Capacity fit0.30The site's capacity divided by the group's size, capped at 1.0, since a site twice the size of the group scores the same as one ten times its size and surplus beyond sufficiency adds nothing
Proximity0.20One minus the distance divided by the 400 kilometre maximum, so a nearby site scores near 1.0 and the furthest scores near zero
Vulnerability match0.10A binary: 1.0 if the group has a special need and the site's medical capacity is Operational, and 0.5 in every other case

The in-app methodology states the ordering these weights encode: protection quality first, physical fit second, operational burden third, population needs fourth.

One observation worth recording. Because vulnerability match is checked only against medical capacity, a group whose stated need is mobility rather than medical is rewarded by the presence of a clinic. The mobility need has no separate representation anywhere in the scoring. The obvious remedy is a second, independent accessibility factor, generated and scored the same way as the existing seven and checked against a mobility need the way medical capacity is checked against a medical one. That has not been built yet, since it would change the balance between gatekeeper and standard factors and the shape of every generated destination, which is a modelling decision rather than a small fix.

TermWhat it measures
ReadinessA single destination's own quality, independent of any particular group
Composite scoreHow well a destination fits one specific group, blending readiness with capacity fit, proximity, and vulnerability match
Success rateThe predicted outcome of assigning one group to one destination, the share of five hundred trials that clear the readiness, capacity, and gatekeeper conditions at once
Three terms this report uses that are easy to conflate, since all three describe how good something is at a different point in the pipeline.

A destination can rank first by composite score for a group and still have a low success rate, because composite score blends readiness with logistics terms that a hard gatekeeper cap does not touch. The worked example below shows this happening.

Three numbers govern the simulation of uncertainty itself:

  • Runs, set to 500 per group and destination pair. This is how many times each pairing is tested. More runs give a tighter estimate at the cost of speed. The information-value panel originally used a cheaper 100 runs, but that count produced estimates too noisy to trust, so it now uses the same 500 runs as everywhere else in the tool.
  • Maximum perturbation probability, set to 0.85. This is the ceiling on how likely a factor's assessment is to be wrong in a given trial, so that at an effective confidence of zero a factor still has a fifteen per cent chance of being reported correctly. The tool never assumes information is entirely worthless.
  • The perturbation step, fixed at one level. When a factor is judged to be misreported it moves exactly one place better or worse along the sequence, with an even chance of each direction. The documentation acknowledges this as a simplification, noting that a real mis-assessment could jump multiple levels, such as an Operational site being reported as Blocked.

Factors already recorded as Unknown are never perturbed. Uncertainty about an unknown is already expressed in its low score and low confidence.

The success threshold. A single trial counts as a success only if three conditions hold together: readiness reaches at least forty per cent, the destination's capacity is at least the group's size, and no gatekeeper is blocked in that trial. An earlier version of the tool's own source table attributed the forty per cent floor to the UNHCR Handbook for Emergencies. That attribution has since been withdrawn as incorrect: readiness is a construct internal to this tool, a weighted average of seven invented factor scores, so no external handbook can define a minimum readiness threshold for it. The floor is now labelled what it actually is, an unsourced modelling assumption, uncalibrated and chosen by the authors. By contrast the two Sphere figures cited later, fifteen litres of water per person per day and 3.5 square metres of covered space, are accurate and are correctly described as informing which factors exist rather than validating any weight.

Note that a destination whose gatekeeper is blocked is capped at twenty per cent readiness, which is below the forty per cent floor. The cap and the threshold are therefore consistent by construction: a blocked gatekeeper cannot produce a successful trial.

06

Reading the Results

Destination cards. Each site shows its readiness percentage and its seven factor statuses. Where a gatekeeper is blocked, the card names which one rather than showing a general warning, so that three distinct failure modes are never collapsed into a single badge.

The assignment matrix. A grid of every group against every destination, showing the composite scores that produced the ranking.

Outcomes. For each group, the assigned destination, the alternative if one exists, the predicted success rate from the five hundred trials, and the standard deviation. A high success rate with a high standard deviation should be read differently from the same rate with a low one: the first is a fragile prediction, the second a stable one.

The riskiest factor. For each assignment the tool identifies the single lowest-scoring factor at the assigned destination, which is the thing most worth checking before moving anyone.

The uncertainty sensitivity chart. For each group, predicted success rate is plotted at eleven uncertainty levels from zero to one hundred per cent. This is the tool's core demonstration in a single picture: the underlying scenario is identical at every point on the line, and only the information quality differs.

The Factor Information Value panel. For each of the seven factors, the tool asks a counterfactual question: if every Unknown instance of this factor were resolved to Operational, how much would the average success rate improve? It answers by genuinely re-running the assignment and the simulation with that change applied, not by estimating. The factors are then ranked by improvement. In practical terms this tells an assessment team which single question is worth the trip.

Alerts. Warnings surface unassigned groups and destinations excluded by a named gatekeeper.

A worked example. Everything above describes what the tool computes. This report also ran it, at the exact citation the reproducibility section recommends: seed 42, eight destinations, three groups, thirty per cent uncertainty. Anyone can regenerate the same scenario at that address on the live site. Of the eight generated destinations, five carry a blocked gatekeeper and are capped at twenty per cent readiness, matching what the project's backlog reports for this seed. The three evacuee groups drawn are Unaccompanied minors, 632 people, immediate; Elderly and mobility-impaired, 378 people, immediate; and Mixed general population, 483 people, can wait.

GroupAssigned destinationComposite scoreReadinessPredicted successRiskiest factor
Unaccompanied minorsZone Golf66.7%20.0%, gatekeeper capped20.2%Willingness, Unwilling
Elderly and mobility-impairedStation Hotel72.4%47.1%48.0%Capacity, Unknown
Mixed general populationZone Golf61.7%20.0%, gatekeeper capped17.8%Willingness, Unwilling
The assignment this scenario actually produces.

Two of the three groups were sent to Zone Golf, the one destination in the pool whose Willingness gatekeeper is blocked. This is what the weighting in the composite score does when a capped but close and roomy site outranks a viable but farther one. Zone Golf sits 26 kilometres away with ample spare capacity, so its capacity fit and proximity terms offset its capped readiness in the blended score. The alternative available to the minors group scored lower on the composite, 63.3% against Zone Golf's 66.7%, but carried a predicted success rate of 49%, more than double. The alternative available to the mixed population group scored 57.4% but carried a success rate of 73.2%, roughly four times higher.

This is the answer this report can now give to the first of the project's three research questions. Multi-factor readiness does not translate into a probable outcome by simple ranking. The composite score the tool actually assigns on blends readiness with logistics terms that can outweigh a hard readiness cap, so the top ranked destination and the destination most likely to succeed are not always the same place, and in this scenario were not the same place for two of the three groups.

The same scenario at higher uncertainty. Re-running the identical destinations and groups, nothing regenerated, at eighty per cent uncertainty instead of thirty gives a clean test of the second research question.

GroupAssigned destinationSuccess at 30% uncertaintySuccess at 80% uncertaintyFall
Unaccompanied minorsZone Golf, unchanged20.2%12.6%38% relative
Elderly and mobility-impairedStation Hotel, unchanged48.0%14.8%69% relative
Mixed general populationZone Golf, unchanged17.8%13.8%22% relative
Predicted success for the same three groups at two uncertainty levels.

The assigned destination for every group is identical at both uncertainty levels. Nothing about the scenario moved. Yet predicted success fell for all three, by very different amounts. Elderly and mobility-impaired lost more than two thirds of its predicted success, Mixed general population lost about a fifth. The unevenness, not just the direction, is the finding the project's earlier peer review said the report was missing.

The sensitivity chart, corrected. The chart plots predicted success against uncertainty from zero to one hundred per cent by re-selecting whichever destination currently scores highest at each of eleven levels, rather than holding one destination fixed. An earlier pass through this exact scenario found that this re-selection step could pick a destination whose capacity was too small for the group, which forced a flat, misleading zero per cent success rate at low uncertainty for one group before the chart jumped once a viable destination became the top scorer. That has since been fixed: the chart now applies the same capacity check the actual assignment step already used. The corrected curve for that group is noisier than a clean decline, wobbling within the range ordinary Monte Carlo variance produces at 500 runs, but the systematic jump is gone.

The Factor Information Value ranking, corrected. For each factor, the panel asks what would happen to mean predicted success if every Unknown instance of that factor were resolved to Operational. Run at its original 100 trial count against this exact scenario, three of the seven factors showed a negative estimated gain, as large as almost seven percentage points, which is not possible in the underlying model, since resolving a factor toward Operational can only raise or hold its score, never lower it. The panel's trial count has since been raised to match the tool's main 500 run engine, and every delta in this scenario is now within about a percentage point of zero. The corrected finding is smaller but more honest: in this particular scenario, none of the seven factors' unresolved instances would move predicted success by more than about a point, and the panel can now say so with reasonable confidence instead of implying an effect that was sampling noise.

Answering the three research questions. Read together, the worked example above does what the project's earlier peer review said the report needed to do. It answers, rather than restates, all three questions the project sets itself. How does multi-factor readiness translate into a probable outcome. Not by simple ranking, since the composite score can send a group to a hard-capped destination over a viable alternative with several times its predicted success. How does degraded field intelligence propagate through an evacuation decision. Unevenly, with the same fixed scenario losing between a fifth and two thirds of its predicted success across three different groups, depending on which factor was driving that group's score. Where would real-time assessment help most. At this scenario's default run count, the honest answer is that the difference is not resolvable from noise, which is itself a finding about how much assessment volume the tool's own panel needs before its ranking should be acted on.

07

The Two Distinctions the Model Insists On

Known unknown against unknown unknown. A factor recorded as Unknown with high confidence, meaning the team is sure it does not know, scores higher than the same status with low confidence, meaning the team is not even sure its ignorance is accurate. This falls directly out of the confidence adjustment. The reasoning offered is that a field team which explicitly flags a gap is providing more actionable information than one whose own reporting cannot be trusted.

Unknown against Unwilling. When the Willingness factor is Blocked, the tool displays the word Unwilling rather than the generic Blocked, throughout the interface and in generated text. The point is not cosmetic. Unknown is an intelligence gap that better assessment can close. Unwilling is a confirmed exclusion that no further enquiry will change. A tool whose purpose is to identify where more information would help must not present a settled refusal as an open question, or it commits the very error it exists to expose.

The project's backlog is honest that this change had consequences. Because gatekeeper factors are generated with a higher probability of being blocked than standard factors, promoting Willingness to gatekeeper status raised the number of excluded destinations in the default scenario from roughly two in eight to roughly five in eight. That rate has since been checked across 500 different seeds rather than just the one the project cites: the mean exclusion count is 4.48 of 8, with 5 in 8 sitting inside the two most common outcomes rather than at an extreme, confirming the rate is the model's typical behaviour rather than an artefact of that particular seed.

08

Grounding in Humanitarian Standards and Law

The tool's own source table is explicit about a boundary that matters. No real dataset is used anywhere. The sources listed inform the design of the model, meaning which factors exist, which are gatekeepers, and where thresholds come from. They do not supply data.

The sources the tool names are as follows.

  • The Sphere Handbook, fourth edition, 2018, behind the Shelter and Food and water factors. Sphere is a voluntary set of minimum standards for humanitarian response, developed by a coalition of humanitarian organisations. The two figures the tool cites from it are accurate: a minimum of 15 litres of water per person per day, and 3.5 square metres of covered living space per person.
  • The UNHCR Handbook for Emergencies, third edition, 2007. An earlier version of the tool cited this as the origin of the forty per cent minimum readiness threshold. That attribution has since been withdrawn as incorrect, since readiness is a construct internal to the tool. The handbook is retained for what it does support, informing which readiness factors are worth modelling, and the threshold itself is treated in this report as an unsourced modelling assumption.
  • Inter-Agency Standing Committee guidance from 2007, cited behind the Authority consent gatekeeper. The Inter-Agency Standing Committee is the main coordination forum of the United Nations and non-governmental humanitarian system. Its guidance in this area establishes that humanitarian operations require the consent of the host government, which is the principle the gatekeeper encodes.
  • Additional Protocol I to the Geneva Conventions, 1977, Articles 12 and 58, behind the Medical capacity and Security factors. Article 12 protects medical units from attack. Article 58 requires parties to a conflict to take precautions to protect civilians under their own control from the effects of attacks.
  • An ICRC publication from 2013 on violence and the use of force, cited as grounding the Security gatekeeper. The ICRC is the International Committee of the Red Cross, the body with a specific mandate under the Geneva Conventions.
  • Park and Miller, 1988, for the random number generator, which is a computing citation rather than a humanitarian one.

A reader should treat these citations as the origin of design choices rather than as validation of the numbers. The tool does not claim that Sphere endorses a weight of 2 for gatekeepers, and it should not be read as claiming so.

09

Reproducibility

Every output of the tool is fully determined by four values: the seed, the number of destinations, the number of evacuee groups, and the uncertainty level. Nothing else varies.

The tool writes those four values into the web address of the page. A Copy link button places that address on the clipboard. Anyone who opens that address regenerates the identical scenario, on any machine. The project recommends citing results in the form seed 42, eight destinations, three groups, thirty per cent uncertainty.

This is a genuine strength and is unusual in tools of this kind. A claim made from this simulator can be checked by a reviewer in a few seconds rather than taken on trust.

10

Relationship to Other Ethical Tech CoLab Projects

Three projects in the same organisation address civilian evacuation and share a common documentation pattern of a concept note, an architecture note, and a backlog. They are separate pieces of software with different aims and no shared code.

ERCF, the Evacuation Risk and Cost Framework, is a server-based tool that estimates the human and financial cost of an evacuation and compares the one-off cost of moving people against the accumulating cost of sustaining them where they are. Unlike ERUS it uses real external data, including conflict event records and country-level severity indices, and a set of documented historical cases.

Evac-Sim-Melanie is an agent-based model of how individual households decide to leave, following a population through stages from unaware to departed, and modelling how alerts spread between neighbours.

The three occupy different levels. ERCF asks what an evacuation would cost. Evac-Sim-Melanie asks how a population behaves once told to move. ERUS asks a narrower and prior question: given a set of possible destinations, how much does the quality of our information about them determine the outcome? Readers interested in any of the others should consult those repositories directly, as this report describes only ERUS.

11

Practical Nature and Maturity of the Tool

What is built and works. The simulator is complete and functional. It is one self-contained web page with no installation, no build step, and no external dependencies, which is a deliberate choice so that it will still run years from now without a toolchain to maintain. It is published as a live public demonstration through GitHub Pages. All of the mechanics described above are implemented in the code and were verified against it for this report.

What was not built, and has since been added. Earlier versions of this report and of the project's own backlog stated there were no automated tests and no licence file. Both gaps have since been closed. The scoring, perturbation, simulation, and assignment functions now live in a standalone engine file, with a regression suite that pins known seed to known output results, including the exact figures in the worked example above. The repository carries an MIT licence. What remains genuinely absent is calibration against real data.

A documentation problem, now resolved. Earlier versions of this report and of the README named a Word document as the authoritative source for formula derivations and citations, to be preferred over the in-app text if the two ever disagreed. That document was excluded from version control, so a reader who obtained a copy of the repository could not consult the document the repository told them was authoritative, and it was separately behind the code in several respects, including the number of simulation runs and the promotion of Willingness to gatekeeper status. This has been fixed by retiring that document and replacing it with a version-controlled methodology file that makes no claim to being auto-generated, is checked directly against the code, and carries its own changelog recording exactly what it corrects relative to the old one.

Which claims in this report can be checked, and which cannot. The distinction matters more than it might appear, because it is not uniform across the report. Everything describing what the tool computes was read off the source: the scoring rule, the confidence adjustment, the gatekeeper cap, the perturbation probability, the composite weights, the success conditions, and the seeded generator. A reader with the repository open can verify all of it. Everything explaining why a parameter takes the value it does rests on the project documents, and the authoritative one among them, the project's methodology document, is now version-controlled and can be consulted directly rather than taken on trust. That includes the reasoning for promoting Willingness to gatekeeper status, the field taxonomy said to ground the eight population types, and the derivation of the 0.30 score for Unknown.

History of drift. The backlog records that two copies of the tool existed in this repository for a period, with new features landing only in the one that was not being published, so that the public demonstration was stale. This has been resolved by making one file canonical and turning the other into a redirect. It is recorded here because it illustrates the class of problem the backlog is mostly concerned with: the code is sound, and the things around the code have repeatedly fallen out of step with it.

12

Limitations and Caveats

Nothing is calibrated. No parameter in the tool has been fitted to or tested against empirical field data. The forty per cent success threshold, the 0.85 perturbation ceiling, the 0.30 score for Unknown, the twenty per cent gatekeeper cap, the doubling of gatekeeper weights, and the four composite weights are all modelling assumptions. The project describes them as open to challenge and says so repeatedly.

Factors are treated as independent. In reality security and authority consent plausibly move together, since a deteriorating security situation often accompanies a withdrawal of consent. The model assumes no such relationship, which will tend to understate how badly things fail together. The project identifies adding factor correlation as the highest-value next step for realism.

There is no time. The model is a single snapshot. Nothing deteriorates or improves over the course of an evacuation, and no journey takes any time to complete.

Routes are represented by one number. Distance stands in for the entire operational burden of a journey. Road condition, checkpoints, fuel availability, seasonal access, and the presence of mines are not modelled at all. A reader should not take the proximity term as a statement about whether a route is passable.

The assignment is greedy, not optimal. Groups are placed one at a time in urgency order, each taking the best remaining site. This is realistic in that urgent cases are handled first, but it means an early group can take a site that would have been a much better match for a later one. The tool does not search for the best overall allocation.

Mis-assessment is gentle. A factor can only be wrong by one level. Real reporting failures can be larger, and the documentation says so.

The population model is coarse. Eight archetypes with fixed vulnerability values, of which only three distinct values are used, and a single binary check against medical capacity. Group composition, disability other than mobility, language, and legal status are not represented. The specific fix for the mobility case, a dedicated accessibility factor scored the way medical capacity is scored today, is a proposed but not yet made change, since it would shift the balance between gatekeeper and standard factors.

The outputs mean nothing about the real world. This follows from the fact that the inputs are invented. The tool demonstrates a relationship between information quality and decision quality. It does not predict, and cannot predict, the outcome of any actual evacuation.

The tool's own statement of its status. It is unambiguous and should be respected: a conceptual demonstration tool for thesis research, not an operational decision-support system, whose outputs require empirical calibration before any real-world use.

13

Intended Audience and Use

The tool is built for an academic argument. Its stated purpose is to make an uncertainty-propagation argument legible to thesis examiners and stakeholders. It is well suited to that, and to teaching, where the sensitivity chart makes an abstract point about information quality concrete in a way that prose does not.

It has a secondary use in prioritising assessment effort. The information-value panel answers a question humanitarian teams face regularly, which is which of several unanswered questions is worth the journey to answer. Even with invented data, the structure of that reasoning is transferable.

It is not for planning. No part of this tool should inform a decision about moving actual people. The project says this itself, in every document it contains, and this report repeats it without qualification.

14

Conclusion

The most useful thing about this simulator is a relationship it makes easy to inspect. Move the uncertainty slider and predicted outcomes deteriorate while the scenario behind them is provably unchanged, because the scenario is regenerated only on demand and not when the slider moves. The degradation is entirely an artefact of what is knowable. That the direction of the effect is built into the scoring rule rather than discovered by it is stated above and should not be forgotten; what the tool contributes is the magnitude, the differing slopes across population types, and the ranking of which unknowns are worth resolving first, shown with real numbers in the worked example above. Making those visible, and reproducible from a seed, is a more honest contribution than a tool that produced confident numbers would be.

The two epistemic distinctions are the other substantive contribution. Separating a confident admission of ignorance from an unreliable one, and separating an open question from a settled refusal, are both choices that a simpler scoring system would have flattened. The decision to promote Willingness to a gatekeeper, and to name it Unwilling rather than Blocked, came from noticing that the arithmetic was allowing a good food supply to outvote a community's refusal. That is the kind of error that is easy to make and hard to see once made.

The candour of the project's own backlog is worth noting as a practice in itself. Across this project's history it has recorded an absent methodology document, since replaced with a version-controlled one; a stale published build, since fixed; previously missing tests, since added; and the possibility that its own most recent modelling change produced an exclusion rate that is an artefact rather than an intention, since checked across 500 seeds and confirmed as typical rather than artefactual. What remains open is calibration against real data, which the backlog names honestly as the one gap tests and documentation cannot close. A research instrument that catalogues its own weaknesses this specifically, and then closes them out in the same place it recorded them, is easier to trust about what it still claims.

What remains, before any of this could be more than a demonstration, is calibration. Every threshold in the tool is a considered guess. The tool is transparent about which guesses they are and where each one lives in the code, which is the necessary precondition for someone eventually replacing them with measurements.

References

Sources

  1. 01Sphere Association. The Sphere Handbook: Humanitarian Charter and Minimum Standards in Humanitarian Response, fourth edition, 2018. Behind the Shelter and Food and water factors.
  2. 02United Nations High Commissioner for Refugees. Handbook for Emergencies, third edition, 2007. An earlier version of the tool's documentation cited this as the origin of the forty per cent minimum readiness threshold. That attribution has been withdrawn. The handbook is retained here for informing which readiness factors are worth modelling.
  3. 03Inter-Agency Standing Committee. Guidance on humanitarian operations requiring the consent of the host government, 2007. Behind the Authority consent gatekeeper.
  4. 04Protocol Additional to the Geneva Conventions of 12 August 1949, and relating to the Protection of Victims of International Armed Conflicts (Protocol I), 1977, Articles 12 and 58. Behind the Medical capacity and Security factors.
  5. 05International Committee of the Red Cross. Publication on violence and the use of force, 2013. Cited as grounding the Security gatekeeper.
  6. 06Park, S.K. and Miller, K.W. (1988). Random Number Generators: Good Ones Are Hard To Find. Communications of the ACM, volume 31, issue 10. The generator used for all seeded randomness in the tool.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file diff --git a/static-site/print/erus/index.txt b/static-site/print/erus/index.txt index c0b129a67..fc3a6aaa5 100644 --- a/static-site/print/erus/index.txt +++ b/static-site/print/erus/index.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -12:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","print","erus",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","erus","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -14:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -16:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","print","erus",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","erus","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:"$Sreact.suspense" -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1c:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -29,7 +29,7 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 19:[] 10:"$W19" 11:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1e:Ta14, @page { size: 210mm 297mm; margin: 20mm 18mm; } @@ -174,6 +174,6 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 5a:["$","tr","7",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Mixed general population"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"2"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"None"}],["$","td","3",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Can wait"}]]}] 5b:["$","figcaption",null,{"className":"mt-3 text-sm text-muted","children":"The eight evacuee archetypes and their fixed properties."}] 1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -5c:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +5c:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 18:null 1d:[["$","title","0",{"children":"The Evacuation Readiness and Uncertainty Simulator (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L5c","6",{}]] diff --git a/static-site/print/evacuation-inform-index/__next._full.txt b/static-site/print/evacuation-inform-index/__next._full.txt index 793ce2c54..879a02255 100644 --- a/static-site/print/evacuation-inform-index/__next._full.txt +++ b/static-site/print/evacuation-inform-index/__next._full.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -12:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","print","evacuation-inform-index",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","evacuation-inform-index","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -14:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -16:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","print","evacuation-inform-index",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","evacuation-inform-index","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:"$Sreact.suspense" -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1c:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -29,7 +29,7 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 19:[] 10:"$W19" 11:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1e:Ta14, @page { size: 210mm 297mm; margin: 20mm 18mm; } @@ -126,6 +126,6 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 2d:T45e,Stated plainly, the proxy overstates evacuation risk for lower intensity crises inside conflict affected states, because national conflict conditions inflate complexity independently of the hazard being scored. It understates evacuation risk where one specific corridor is dangerous but the country as a whole is permissive, because complexity has no corridor level resolution at all. A third pattern emerged that no reviewer anticipated: across the seventeen most severe crises the ratio ranges only from 0.83 to 1.13, clustering so tightly around 1.0 that it discriminates almost nothing. Every large ratio in the index comes from a low or mid severity crisis. The headline number is most confident precisely where it is least meaningful, which is why the tool now warns at the point of reading whenever both components sit near the top of the scale. The substitution is kept because no public dataset scores civilian evacuation difficulty across 104 crises and complexity is the closest construct with real coverage and a published methodology, but it should be read as a screening prompt rather than a measurement.2b:["$","p","7",{"children":"$2d"}] 2c:["$","p","8",{"children":"Indicators combine by weighted geometric mean rather than arithmetic mean. The practical consequence is that a catastrophic score on one component cannot be averaged away by comfortable scores elsewhere. If every evacuation route is closed, no amount of favourable weather or food security compensates for it. This is the same non compensatory logic used by the Human Development Index, and it reflects a judgement about the world rather than a mathematical preference: some failures are absolute."}] 1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -2e:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +2e:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 18:null 1d:[["$","title","0",{"children":"The Evacuation Inform Index (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L2e","6",{}]] diff --git a/static-site/print/evacuation-inform-index/__next._head.txt b/static-site/print/evacuation-inform-index/__next._head.txt index 91aeceed9..4143c80f3 100644 --- a/static-site/print/evacuation-inform-index/__next._head.txt +++ b/static-site/print/evacuation-inform-index/__next._head.txt @@ -1,6 +1,6 @@ 1:"$Sreact.fragment" -2:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -3:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +2:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +3:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 4:"$Sreact.suspense" -5:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +5:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 0:{"rsc":["$","$1","h",{"children":[null,["$","$L2",null,{"children":[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]]}],["$","div",null,{"hidden":true,"children":["$","$L3",null,{"children":["$","$4",null,{"name":"Next.Metadata","children":[["$","title","0",{"children":"The Evacuation Inform Index (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L5","6",{}]]}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/print/evacuation-inform-index/__next._index.txt b/static-site/print/evacuation-inform-index/__next._index.txt index f98a4fd90..18b9d82b6 100644 --- a/static-site/print/evacuation-inform-index/__next._index.txt +++ b/static-site/print/evacuation-inform-index/__next._index.txt @@ -1,13 +1,13 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -d:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +d:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} 8:["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$Ld",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$Ld",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$Ld",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$Ld",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] diff --git a/static-site/print/evacuation-inform-index/__next._tree.txt b/static-site/print/evacuation-inform-index/__next._tree.txt index f014a1343..8b258dbb4 100644 --- a/static-site/print/evacuation-inform-index/__next._tree.txt +++ b/static-site/print/evacuation-inform-index/__next._tree.txt @@ -1,7 +1,7 @@ -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] 0:{"tree":{"name":"","param":null,"prefetchHints":16,"slots":{"children":{"name":"print","param":null,"prefetchHints":0,"slots":{"children":{"name":"slug","param":{"type":"d","key":"evacuation-inform-index","siblings":null},"prefetchHints":0,"slots":{"children":{"name":"__PAGE__","param":null,"prefetchHints":0,"slots":null}}}}}}},"staleTime":300,"buildId":"static-snapshot"} diff --git a/static-site/print/evacuation-inform-index/__next.print.txt b/static-site/print/evacuation-inform-index/__next.print.txt index 460afe242..fa617e997 100644 --- a/static-site/print/evacuation-inform-index/__next.print.txt +++ b/static-site/print/evacuation-inform-index/__next.print.txt @@ -1,4 +1,4 @@ 1:"$Sreact.fragment" -2:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -3:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +2:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +3:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] 0:{"rsc":["$","$1","c",{"children":[null,[["$","style",null,{"children":"\n /* The site is dark-first, and a dark color-scheme makes Chrome paint\n the printed sheet's margin area dark too. Paper is light. */\n html { color-scheme: light !important; }\n html, body { background: #ffffff !important; }\n body::after { content: none !important; }\n header, footer, .site-bg, .aura { display: none !important; }\n "}],["$","$L2",null,{"parallelRouterKey":"children","template":["$","$L3",null,{}]}]]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/print/evacuation-inform-index/index.html b/static-site/print/evacuation-inform-index/index.html index 6203619d7..8b847f3ce 100644 --- a/static-site/print/evacuation-inform-index/index.html +++ b/static-site/print/evacuation-inform-index/index.html @@ -1,4 +1,4 @@ -The Evacuation Inform Index (print edition) · NYU Ethical Tech CoLab

Publications · Academic report

The Evacuation Inform Index

Weighing the Risk of Leaving Against the Risk of Staying

Ethical Tech CoLab

NYU Center for Global Affairs

July 2026

Carolina Morón. Prepared as masters research at the NYU Center for Global Affairs.

Humanitarian practice tends to treat evacuation as the obvious good. In practice the journey carries its own risk, and a crisis can be both catastrophic to endure and impossible to escape. The Evacuation Inform Index scores those two risks separately, side by side, so that the comparison between them becomes visible and can be argued with.

Key figures

104

active crises scored on both the risk of staying and the risk of leaving

75%

endangerment threshold, drawn from the obligation to protect civilians in danger

12

vulnerability factors, ten that raise assessed risk and two that lower it

10

limitations the prototype states about itself, including that its weights are unvalidated

01

The Question Nobody Asks

Every displacement crisis contains a question that is rarely asked out loud: is leaving actually safer than staying? Roads are mined or blockaded. Checkpoints separate families. The elderly and the sick do not survive transit that a healthy adult would manage. A destination that looked open on Monday has closed by Thursday.

Existing humanitarian indices measure how bad a situation is. They rank crises by severity so that funding and attention can be allocated. What they do not do is model the alternative. A crisis scored as extremely severe tells a planner that people are suffering; it does not tell them whether moving those people would reduce or increase the harm.

That gap matters because the two risks move independently. A siege can be catastrophic to endure and simultaneously impossible to escape. A drought can be survivable in place while the routes out remain open and safe. Treating severity as a proxy for a decision to evacuate conflates conditions that need to be distinguished.

International humanitarian law makes the distinction consequential rather than merely analytical. Under Article 49 of the Fourth Geneva Convention, an Occupying Power may undertake total or partial evacuation of a given area where the security of the population or imperative military reasons so demand, must return those evacuated to their homes once hostilities have ceased, and must ensure proper accommodation, satisfactory conditions of hygiene, health, safety and nutrition, and that members of the same family are not separated. Evacuation is a regulated act with conditions attached, not a self evidently benign one.

02

What the Index Is

The Evacuation Inform Index is a browser based research prototype. It presents a world map of active crises, a panel of live developments for each one, a methodology section, a catalogue of data sources, and a reference list.

It scores 104 active humanitarian crises on two dimensions: the risk of remaining in place, and the risk of attempting to leave. It expresses the relationship between them as a single ratio, while always displaying the two component scores that produced it.

The map can be filtered to a crisis type. INFORM's driver labels are grouped into four families: environmental, covering floods, drought, cyclone and earthquake; conflict and violence; international displacement; and political or economic crisis. Crises routinely carry several drivers, so the groups overlap deliberately and a crisis appears whenever any of its drivers is selected. All four are selected by default and every crisis belongs to at least one, so the unfiltered map is the whole dataset, and whenever a filter narrows the view the legend says how many crises are being shown.

Two further layers address the roads. A transparent road network overlay can be switched on over any base layer, so streets and highways read against the satellite imagery rather than replacing it. Separately, the road access reports gathered for each crisis can be pinned on the map. A pin marks the crisis a report belongs to and never the blockage itself, because news prose carries no coordinates: a report that a named road has been cut identifies no point that can be plotted.

Every legal provision cited in the methodology section is clickable and opens an explanation written for a reader with no legal training. Each has four parts: what the provision requires in plain words, its operative text so that the plain language rendering can be checked against the source, why this index invokes it, and how far that invocation is justified. The fourth part is the reason the feature exists. A citation that is only ever displayed reads as authority the model has not earned, and several of these do not survive the examination. The endangerment threshold is labelled with Article 49 of the Fourth Geneva Convention on every crisis, including floods and droughts where an article governing occupied territory has no application, and Article 17 of Additional Protocol II is the more relevant rule for most of the armed conflicts shown and goes unused.

It is not deployed in any operational setting, has no institutional mandate, and produces no output that any organisation is obliged to act upon. The distinction between what is implemented and what is planned is marked throughout the interface rather than left to the reader to infer.

That distinction is worth stating precisely, because the full methodology specifies weights for components that do not yet exist, and the specificity can read as a report of a running system. What is live: the INFORM mapping across all 104 crises, the ratio with its floor and its mandatory two component display, the endangerment and feasibility split with its protection gap flag, conflict data with a three month trajectory, news retrieval, route weather, and the road access layer. What is designed but not built: the three layer architecture and its weights, the seven dimension decomposition, corridor and checkpoint dynamics, and destination readiness gatekeepers. What has not started: the expert elicitation and the analytic hierarchy process validation that would turn any of these weights into something more than one researcher's estimate. The vulnerability profile sits between the two, live as an interactive display but resting entirely on unvalidated numbers. No weight in this work has been validated by anyone.

03

How It Works

The index expresses its result as a ratio: the risk score for evacuating divided by the risk score for staying. A value above 1.0 indicates that evacuation carries more risk than remaining. A value below 1.0 indicates the reverse. A value near 1.0 indicates that the two courses of action carry comparable risk, which is itself a meaningful finding rather than an absence of one.

Ratios become unstable as the denominator approaches zero, so a floor is applied to the staying score. On the current data that floor never engages, since no crisis scores low enough to reach it. And the ratio is never shown alone: both component scores accompany it wherever it appears.

Dividing one score by another assumes both are quantities with a true zero, where twice as much means something. The INFORM severity bands underneath do not clearly meet that standard. There is no defensible sense in which a conditions score of 4 is exactly twice the severity of a 2, and so no fully defensible sense in which a ratio of 2.6 means evacuation is 2.6 times riskier than staying. The division is a comparison device rather than a measurement. It supports ordering and sign, whether evacuation scores worse than staying and roughly how far apart the two sit, and it does not support arithmetic on the result: differences of a few hundredths carry no meaning, and the ratio should never be averaged across crises or fed into a further calculation.

There is a tension here worth naming rather than leaving to the reader. The index argues at length that danger and feasibility must never be collapsed into one score, and then leads with a single ratio. The ratio is a pointer that directs attention to crises where the comparison is unusual, not a verdict on either dimension. It cannot perform the collapse the argument forbids, because both components stay visible on its face and neither is absorbed into it. The collapse that matters is the one that would let operational difficulty quietly reduce apparent danger and so soften an obligation, and a number that always carries its own numerator and denominator cannot do that.

The component scores are not invented for the prototype. They derive from the INFORM Severity Index, the monthly crisis severity model produced by ACAPS with the Joint Research Centre of the European Commission. INFORM assembles 31 core indicators into three weighted dimensions: the impact of the crisis at 20 per cent, the conditions of the affected people at 50 per cent, and the complexity of the situation at 30 per cent.

The index maps two of those dimensions onto its own question. The conditions of affected people become the risk of staying, measuring how severe it is to remain in place. The complexity of the crisis, which covers access constraints, safety, and the operating environment, becomes the risk of evacuating, measuring how difficult and dangerous it is to move. The substitution is defensible on its face, since complexity measures precisely the conditions that obstruct movement, and the prototype labels it as a substitution rather than a measurement.

Labelling a substitution is not the same as testing it, and this is the load bearing question for the whole instrument. Across the 104 crises the correlation between the two dimensions is 0.62, a real association that still leaves well over half the variance unshared. The gap has a direction. Complexity scores the operating environment that responders face, which is largely a property of the country and largely driven by conflict, while conditions track the affected population of one particular crisis. When a low intensity crisis sits inside a high conflict country the two come apart hard. The clearest case in the data is the 2026 floods in Yemen, which carry the highest ratio in the entire index at 2.60: the lowest conditions score in the dataset combined with a complexity score that reflects Yemen's war. A family deciding whether to move away from floodwater is not facing that much evacuation danger from the flood. The number is real, the inputs are real, and the reading the index invites is wrong.

Stated plainly, the proxy overstates evacuation risk for lower intensity crises inside conflict affected states, because national conflict conditions inflate complexity independently of the hazard being scored. It understates evacuation risk where one specific corridor is dangerous but the country as a whole is permissive, because complexity has no corridor level resolution at all. A third pattern emerged that no reviewer anticipated: across the seventeen most severe crises the ratio ranges only from 0.83 to 1.13, clustering so tightly around 1.0 that it discriminates almost nothing. Every large ratio in the index comes from a low or mid severity crisis. The headline number is most confident precisely where it is least meaningful, which is why the tool now warns at the point of reading whenever both components sit near the top of the scale. The substitution is kept because no public dataset scores civilian evacuation difficulty across 104 crises and complexity is the closest construct with real coverage and a published methodology, but it should be read as a screening prompt rather than a measurement.

Indicators combine by weighted geometric mean rather than arithmetic mean. The practical consequence is that a catastrophic score on one component cannot be averaged away by comfortable scores elsewhere. If every evacuation route is closed, no amount of favourable weather or food security compensates for it. This is the same non compensatory logic used by the Human Development Index, and it reflects a judgement about the world rather than a mathematical preference: some failures are absolute.

04

Keeping Danger and Feasibility Apart

The prototype adopts a structural argument from the CERAI framework, which holds that danger and feasibility must never be collapsed into one score.

The reasoning is legal as much as analytical. The obligation to protect civilians arises from the danger they face. Operational difficulty does not extinguish that obligation. A single blended number would allow low feasibility to drag down the composite, making a desperate situation appear less urgent precisely because it is hard to resolve.

Keeping the two visible produces the opposite effect. High danger combined with low feasibility raises a protection gap flag, signalling that the problem has moved beyond operational reach and requires political engagement rather than a logistical answer.

Endangerment is drawn from the INFORM conditions score and expressed as a percentage, with a 75 per cent obligation threshold marked and a trajectory drawn from recent conflict fatality trends. Feasibility inverts the complexity score and reduces it by live route weather. Note that feasibility and the risk of evacuating are the same quantity under two names, pointing in opposite directions: evacuation risk is approximately one minus feasibility. They are not two independent readings of movement, they are one input read twice, and they carry every limitation of the complexity proxy described above. A future version should either derive feasibility from an independent source or report a single quantity.

A trajectory indicator compares fatalities in the most recent three months against the preceding three, answering whether a situation is deteriorating or stabilising, which a static severity score cannot convey.

05

Who Is Exposed

A crisis score describes a population. It does not describe a person within it. The prototype therefore offers an interactive profile of twelve factors that a user can switch on to represent a particular group, bounded between a multiplier of 0.7 and 1.3.

That bound has a consequence the design did not intend. Each factor shifts the multiplier by 0.06, so five upward toggles reach the ceiling of 1.3 and the sixth through the tenth change nothing at all. Ten of the twelve factors raise risk. A household carrying the most compounded vulnerability, an elderly and disabled and undocumented member of a targeted minority who cannot read the language, is precisely where the model stops discriminating between cases. The factors are described as cumulative and they stop being cumulative exactly where the need is greatest.

The multiplier raises endangerment and lowers feasibility, on the reasoning that these are two causally distinct pathways rather than one effect measured twice. A personal characteristic does not add a fixed quantity of risk; it scales the risk already present in the environment. Being elderly in a stable region and being elderly under bombardment are not the same increment.

The twelve factors extend well beyond mobility. Young children, the elderly, the disabled and medically dependent, the wounded and acutely sick, and pregnant women and new mothers all bear on the physical capacity to travel. Unaccompanied and separated minors, people facing gendered violence risk, members of targeted ethnic, religious or political groups, undocumented people with identity gaps, and linguistic minorities with low literacy face a different kind of exposure: they are endangered by targeting, detention, or exclusion rather than by the difficulty of walking. Two factors work in the opposite direction, reducing assessed risk: prior evacuation experience and available financial resources.

The distinction matters. A demographic list alone would treat vulnerability as a question of who moves slowly. The protection based additions treat it as a question of who is hunted, who is turned back, and who cannot read the sign telling them where to go.

06

How the Index Treats Evidence

The laboratory hosting this work has a lineage in supply chain traceability and forced labour mitigation, and several practices carry over into how the index treats evidence.

A two witness standard separates a fact corroborated by two or more independent reports from a single unverified claim, turning source credibility into a graded ladder rather than a binary.

The score is deterministic rather than generated. Every number comes from a fixed, auditable formula. Where a language model is involved, it supplies source grounded facts and never overrides the arithmetic. A reader can reconstruct any score by hand.

An evidence gap is treated as a risk signal rather than a neutral absence. An unverified corridor segment is penalised, not assumed safe. This inverts the common default in which missing information quietly reads as an acceptable situation. Where the tool falls back to background knowledge rather than a live verified source, that fallback is labelled, capped below verified status, and flagged.

The optional satellite damage layer follows the same discipline. It stays empty until a user connects their own deployment, and the interface reports the real state of that connection rather than assuming it: green only once imagery actually arrives, failure when none does, and partial coverage distinguished from failure. Nothing about the connection is asserted before it is observed.

07

What the Index Does Not Do

The repository states ten limitations. The most consequential are reproduced here, because they are the most important part of the document for any reader considering what weight to give the tool.

Proxy construct. Endangerment and feasibility derive from INFORM sub scores supplemented by conflict and weather data. They should be read as a faithful architectural proxy, not a validated instrument.

Researcher assigned weights. Every weight is a best estimate pending expert validation. This places them at the same evidentiary level as INFORM's own initial weights, which ACAPS likewise describes as a current best estimate to be refined by expert analysis.

No ground truth calibration. Independent ground truth for evacuation decisions does not exist. There is no register of correct historical calls against which the index could be scored.

Population level vulnerability. The demographic profile is a scenario the user sets, not measured household data. It cannot capture individual circumstance.

Static snapshot. The endangerment and feasibility pair has no corridor dynamics, ceasefire windows, or modelling of misinformation. When hosted as a static site, news and conflict data are captured rather than live.

No modelling of political will. The index cannot represent actor behaviour, negotiation status, sudden shifts in belligerent intent, or the granting and withdrawal of consent. In many real evacuations these are the determining factors.

Road access is searched everywhere, but read from news prose. All 104 crises have now been searched, so an absence of road reports means one thing rather than two: the search ran and found nothing. Fourteen crises carry reports. What remains is the harder limit. The signal is read from news prose by keyword, which cannot tell what a sentence is about, so a report of an airstrike can be classified as a road blockage; items are filtered for recency and for whether the publisher is a newsroom, but nothing checks whether the road described is the road a reader cares about. The pin marks the crisis and never the blockage, because news prose carries no coordinates. Treat the layer as a prompt to look, not as a map of which roads are open.

Correlation, not causation. The index prioritises attention. It is decision support and must not be the sole basis for an evacuation decision.

08

What It Contributes

The contribution of this prototype is not a novel algorithm. It is a reframing. By insisting that the risk of leaving be scored separately from the risk of staying, and that danger be scored separately from feasibility, it makes visible a comparison that humanitarian assessment usually leaves implicit.

Its second contribution is its treatment of its own uncertainty. The weights are labelled as unvalidated. The proxies are labelled as proxies. The absent capabilities are listed rather than omitted. The satellite overlay reports honestly that it is not connected. For a tool that touches decisions about who moves and who remains, this reticence is not a weakness in the work; it is a substantial part of what the work is demonstrating.

The index does not know whether anyone should evacuate. It is built to make the question harder to answer carelessly.

References

Sources

  1. 01ACAPS and the Joint Research Centre of the European Commission. INFORM Severity Index, April 2026 release, 104 crises.
  2. 02International Committee of the Red Cross. Article 49, Geneva Convention (IV) relative to the Protection of Civilian Persons in Time of War, 1949.
  3. 03ACLED. Armed Conflict Location and Event Data, conflict events and fatalities.
  4. 04Open-Meteo. Route weather and daylight data.
  5. 05Microsoft and Planet. HASTE, satellite damage assessment framework.
  6. 06International Organization for Migration. Risk Index for Climate Displacement, two tier macro and micro model.
  7. 07Centers for Disease Control and Prevention. Social Vulnerability Index.
  8. 08Fund for Peace. Fragile States Index.
  9. 09OECD and the Joint Research Centre. (2008). Handbook on Constructing Composite Indicators. OECD Publishing.
  10. 10Saaty, T.L. (1990). How to Make a Decision: The Analytic Hierarchy Process. European Journal of Operational Research, volume 48, issue 1, pages 9 to 26.
  11. 11Beccari, B. (2016). A Comparative Analysis of Disaster Risk, Vulnerability and Resilience Composite Indicators. PLoS Currents Disasters.
  12. 12Al Fozaie, M. (2022). A Guide to Integrating Expert Opinion and Fuzzy AHP When Generating Weights for Composite Indices. Advances in Fuzzy Systems.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file +

Publications · Academic report

The Evacuation Inform Index

Weighing the Risk of Leaving Against the Risk of Staying

Ethical Tech CoLab

NYU Center for Global Affairs

July 2026

Carolina Morón. Prepared as masters research at the NYU Center for Global Affairs.

Humanitarian practice tends to treat evacuation as the obvious good. In practice the journey carries its own risk, and a crisis can be both catastrophic to endure and impossible to escape. The Evacuation Inform Index scores those two risks separately, side by side, so that the comparison between them becomes visible and can be argued with.

Key figures

104

active crises scored on both the risk of staying and the risk of leaving

75%

endangerment threshold, drawn from the obligation to protect civilians in danger

12

vulnerability factors, ten that raise assessed risk and two that lower it

10

limitations the prototype states about itself, including that its weights are unvalidated

01

The Question Nobody Asks

Every displacement crisis contains a question that is rarely asked out loud: is leaving actually safer than staying? Roads are mined or blockaded. Checkpoints separate families. The elderly and the sick do not survive transit that a healthy adult would manage. A destination that looked open on Monday has closed by Thursday.

Existing humanitarian indices measure how bad a situation is. They rank crises by severity so that funding and attention can be allocated. What they do not do is model the alternative. A crisis scored as extremely severe tells a planner that people are suffering; it does not tell them whether moving those people would reduce or increase the harm.

That gap matters because the two risks move independently. A siege can be catastrophic to endure and simultaneously impossible to escape. A drought can be survivable in place while the routes out remain open and safe. Treating severity as a proxy for a decision to evacuate conflates conditions that need to be distinguished.

International humanitarian law makes the distinction consequential rather than merely analytical. Under Article 49 of the Fourth Geneva Convention, an Occupying Power may undertake total or partial evacuation of a given area where the security of the population or imperative military reasons so demand, must return those evacuated to their homes once hostilities have ceased, and must ensure proper accommodation, satisfactory conditions of hygiene, health, safety and nutrition, and that members of the same family are not separated. Evacuation is a regulated act with conditions attached, not a self evidently benign one.

02

What the Index Is

The Evacuation Inform Index is a browser based research prototype. It presents a world map of active crises, a panel of live developments for each one, a methodology section, a catalogue of data sources, and a reference list.

It scores 104 active humanitarian crises on two dimensions: the risk of remaining in place, and the risk of attempting to leave. It expresses the relationship between them as a single ratio, while always displaying the two component scores that produced it.

The map can be filtered to a crisis type. INFORM's driver labels are grouped into four families: environmental, covering floods, drought, cyclone and earthquake; conflict and violence; international displacement; and political or economic crisis. Crises routinely carry several drivers, so the groups overlap deliberately and a crisis appears whenever any of its drivers is selected. All four are selected by default and every crisis belongs to at least one, so the unfiltered map is the whole dataset, and whenever a filter narrows the view the legend says how many crises are being shown.

Two further layers address the roads. A transparent road network overlay can be switched on over any base layer, so streets and highways read against the satellite imagery rather than replacing it. Separately, the road access reports gathered for each crisis can be pinned on the map. A pin marks the crisis a report belongs to and never the blockage itself, because news prose carries no coordinates: a report that a named road has been cut identifies no point that can be plotted.

Every legal provision cited in the methodology section is clickable and opens an explanation written for a reader with no legal training. Each has four parts: what the provision requires in plain words, its operative text so that the plain language rendering can be checked against the source, why this index invokes it, and how far that invocation is justified. The fourth part is the reason the feature exists. A citation that is only ever displayed reads as authority the model has not earned, and several of these do not survive the examination. The endangerment threshold is labelled with Article 49 of the Fourth Geneva Convention on every crisis, including floods and droughts where an article governing occupied territory has no application, and Article 17 of Additional Protocol II is the more relevant rule for most of the armed conflicts shown and goes unused.

It is not deployed in any operational setting, has no institutional mandate, and produces no output that any organisation is obliged to act upon. The distinction between what is implemented and what is planned is marked throughout the interface rather than left to the reader to infer.

That distinction is worth stating precisely, because the full methodology specifies weights for components that do not yet exist, and the specificity can read as a report of a running system. What is live: the INFORM mapping across all 104 crises, the ratio with its floor and its mandatory two component display, the endangerment and feasibility split with its protection gap flag, conflict data with a three month trajectory, news retrieval, route weather, and the road access layer. What is designed but not built: the three layer architecture and its weights, the seven dimension decomposition, corridor and checkpoint dynamics, and destination readiness gatekeepers. What has not started: the expert elicitation and the analytic hierarchy process validation that would turn any of these weights into something more than one researcher's estimate. The vulnerability profile sits between the two, live as an interactive display but resting entirely on unvalidated numbers. No weight in this work has been validated by anyone.

03

How It Works

The index expresses its result as a ratio: the risk score for evacuating divided by the risk score for staying. A value above 1.0 indicates that evacuation carries more risk than remaining. A value below 1.0 indicates the reverse. A value near 1.0 indicates that the two courses of action carry comparable risk, which is itself a meaningful finding rather than an absence of one.

Ratios become unstable as the denominator approaches zero, so a floor is applied to the staying score. On the current data that floor never engages, since no crisis scores low enough to reach it. And the ratio is never shown alone: both component scores accompany it wherever it appears.

Dividing one score by another assumes both are quantities with a true zero, where twice as much means something. The INFORM severity bands underneath do not clearly meet that standard. There is no defensible sense in which a conditions score of 4 is exactly twice the severity of a 2, and so no fully defensible sense in which a ratio of 2.6 means evacuation is 2.6 times riskier than staying. The division is a comparison device rather than a measurement. It supports ordering and sign, whether evacuation scores worse than staying and roughly how far apart the two sit, and it does not support arithmetic on the result: differences of a few hundredths carry no meaning, and the ratio should never be averaged across crises or fed into a further calculation.

There is a tension here worth naming rather than leaving to the reader. The index argues at length that danger and feasibility must never be collapsed into one score, and then leads with a single ratio. The ratio is a pointer that directs attention to crises where the comparison is unusual, not a verdict on either dimension. It cannot perform the collapse the argument forbids, because both components stay visible on its face and neither is absorbed into it. The collapse that matters is the one that would let operational difficulty quietly reduce apparent danger and so soften an obligation, and a number that always carries its own numerator and denominator cannot do that.

The component scores are not invented for the prototype. They derive from the INFORM Severity Index, the monthly crisis severity model produced by ACAPS with the Joint Research Centre of the European Commission. INFORM assembles 31 core indicators into three weighted dimensions: the impact of the crisis at 20 per cent, the conditions of the affected people at 50 per cent, and the complexity of the situation at 30 per cent.

The index maps two of those dimensions onto its own question. The conditions of affected people become the risk of staying, measuring how severe it is to remain in place. The complexity of the crisis, which covers access constraints, safety, and the operating environment, becomes the risk of evacuating, measuring how difficult and dangerous it is to move. The substitution is defensible on its face, since complexity measures precisely the conditions that obstruct movement, and the prototype labels it as a substitution rather than a measurement.

Labelling a substitution is not the same as testing it, and this is the load bearing question for the whole instrument. Across the 104 crises the correlation between the two dimensions is 0.62, a real association that still leaves well over half the variance unshared. The gap has a direction. Complexity scores the operating environment that responders face, which is largely a property of the country and largely driven by conflict, while conditions track the affected population of one particular crisis. When a low intensity crisis sits inside a high conflict country the two come apart hard. The clearest case in the data is the 2026 floods in Yemen, which carry the highest ratio in the entire index at 2.60: the lowest conditions score in the dataset combined with a complexity score that reflects Yemen's war. A family deciding whether to move away from floodwater is not facing that much evacuation danger from the flood. The number is real, the inputs are real, and the reading the index invites is wrong.

Stated plainly, the proxy overstates evacuation risk for lower intensity crises inside conflict affected states, because national conflict conditions inflate complexity independently of the hazard being scored. It understates evacuation risk where one specific corridor is dangerous but the country as a whole is permissive, because complexity has no corridor level resolution at all. A third pattern emerged that no reviewer anticipated: across the seventeen most severe crises the ratio ranges only from 0.83 to 1.13, clustering so tightly around 1.0 that it discriminates almost nothing. Every large ratio in the index comes from a low or mid severity crisis. The headline number is most confident precisely where it is least meaningful, which is why the tool now warns at the point of reading whenever both components sit near the top of the scale. The substitution is kept because no public dataset scores civilian evacuation difficulty across 104 crises and complexity is the closest construct with real coverage and a published methodology, but it should be read as a screening prompt rather than a measurement.

Indicators combine by weighted geometric mean rather than arithmetic mean. The practical consequence is that a catastrophic score on one component cannot be averaged away by comfortable scores elsewhere. If every evacuation route is closed, no amount of favourable weather or food security compensates for it. This is the same non compensatory logic used by the Human Development Index, and it reflects a judgement about the world rather than a mathematical preference: some failures are absolute.

04

Keeping Danger and Feasibility Apart

The prototype adopts a structural argument from the CERAI framework, which holds that danger and feasibility must never be collapsed into one score.

The reasoning is legal as much as analytical. The obligation to protect civilians arises from the danger they face. Operational difficulty does not extinguish that obligation. A single blended number would allow low feasibility to drag down the composite, making a desperate situation appear less urgent precisely because it is hard to resolve.

Keeping the two visible produces the opposite effect. High danger combined with low feasibility raises a protection gap flag, signalling that the problem has moved beyond operational reach and requires political engagement rather than a logistical answer.

Endangerment is drawn from the INFORM conditions score and expressed as a percentage, with a 75 per cent obligation threshold marked and a trajectory drawn from recent conflict fatality trends. Feasibility inverts the complexity score and reduces it by live route weather. Note that feasibility and the risk of evacuating are the same quantity under two names, pointing in opposite directions: evacuation risk is approximately one minus feasibility. They are not two independent readings of movement, they are one input read twice, and they carry every limitation of the complexity proxy described above. A future version should either derive feasibility from an independent source or report a single quantity.

A trajectory indicator compares fatalities in the most recent three months against the preceding three, answering whether a situation is deteriorating or stabilising, which a static severity score cannot convey.

05

Who Is Exposed

A crisis score describes a population. It does not describe a person within it. The prototype therefore offers an interactive profile of twelve factors that a user can switch on to represent a particular group, bounded between a multiplier of 0.7 and 1.3.

That bound has a consequence the design did not intend. Each factor shifts the multiplier by 0.06, so five upward toggles reach the ceiling of 1.3 and the sixth through the tenth change nothing at all. Ten of the twelve factors raise risk. A household carrying the most compounded vulnerability, an elderly and disabled and undocumented member of a targeted minority who cannot read the language, is precisely where the model stops discriminating between cases. The factors are described as cumulative and they stop being cumulative exactly where the need is greatest.

The multiplier raises endangerment and lowers feasibility, on the reasoning that these are two causally distinct pathways rather than one effect measured twice. A personal characteristic does not add a fixed quantity of risk; it scales the risk already present in the environment. Being elderly in a stable region and being elderly under bombardment are not the same increment.

The twelve factors extend well beyond mobility. Young children, the elderly, the disabled and medically dependent, the wounded and acutely sick, and pregnant women and new mothers all bear on the physical capacity to travel. Unaccompanied and separated minors, people facing gendered violence risk, members of targeted ethnic, religious or political groups, undocumented people with identity gaps, and linguistic minorities with low literacy face a different kind of exposure: they are endangered by targeting, detention, or exclusion rather than by the difficulty of walking. Two factors work in the opposite direction, reducing assessed risk: prior evacuation experience and available financial resources.

The distinction matters. A demographic list alone would treat vulnerability as a question of who moves slowly. The protection based additions treat it as a question of who is hunted, who is turned back, and who cannot read the sign telling them where to go.

06

How the Index Treats Evidence

The laboratory hosting this work has a lineage in supply chain traceability and forced labour mitigation, and several practices carry over into how the index treats evidence.

A two witness standard separates a fact corroborated by two or more independent reports from a single unverified claim, turning source credibility into a graded ladder rather than a binary.

The score is deterministic rather than generated. Every number comes from a fixed, auditable formula. Where a language model is involved, it supplies source grounded facts and never overrides the arithmetic. A reader can reconstruct any score by hand.

An evidence gap is treated as a risk signal rather than a neutral absence. An unverified corridor segment is penalised, not assumed safe. This inverts the common default in which missing information quietly reads as an acceptable situation. Where the tool falls back to background knowledge rather than a live verified source, that fallback is labelled, capped below verified status, and flagged.

The optional satellite damage layer follows the same discipline. It stays empty until a user connects their own deployment, and the interface reports the real state of that connection rather than assuming it: green only once imagery actually arrives, failure when none does, and partial coverage distinguished from failure. Nothing about the connection is asserted before it is observed.

07

What the Index Does Not Do

The repository states ten limitations. The most consequential are reproduced here, because they are the most important part of the document for any reader considering what weight to give the tool.

Proxy construct. Endangerment and feasibility derive from INFORM sub scores supplemented by conflict and weather data. They should be read as a faithful architectural proxy, not a validated instrument.

Researcher assigned weights. Every weight is a best estimate pending expert validation. This places them at the same evidentiary level as INFORM's own initial weights, which ACAPS likewise describes as a current best estimate to be refined by expert analysis.

No ground truth calibration. Independent ground truth for evacuation decisions does not exist. There is no register of correct historical calls against which the index could be scored.

Population level vulnerability. The demographic profile is a scenario the user sets, not measured household data. It cannot capture individual circumstance.

Static snapshot. The endangerment and feasibility pair has no corridor dynamics, ceasefire windows, or modelling of misinformation. When hosted as a static site, news and conflict data are captured rather than live.

No modelling of political will. The index cannot represent actor behaviour, negotiation status, sudden shifts in belligerent intent, or the granting and withdrawal of consent. In many real evacuations these are the determining factors.

Road access is searched everywhere, but read from news prose. All 104 crises have now been searched, so an absence of road reports means one thing rather than two: the search ran and found nothing. Fourteen crises carry reports. What remains is the harder limit. The signal is read from news prose by keyword, which cannot tell what a sentence is about, so a report of an airstrike can be classified as a road blockage; items are filtered for recency and for whether the publisher is a newsroom, but nothing checks whether the road described is the road a reader cares about. The pin marks the crisis and never the blockage, because news prose carries no coordinates. Treat the layer as a prompt to look, not as a map of which roads are open.

Correlation, not causation. The index prioritises attention. It is decision support and must not be the sole basis for an evacuation decision.

08

What It Contributes

The contribution of this prototype is not a novel algorithm. It is a reframing. By insisting that the risk of leaving be scored separately from the risk of staying, and that danger be scored separately from feasibility, it makes visible a comparison that humanitarian assessment usually leaves implicit.

Its second contribution is its treatment of its own uncertainty. The weights are labelled as unvalidated. The proxies are labelled as proxies. The absent capabilities are listed rather than omitted. The satellite overlay reports honestly that it is not connected. For a tool that touches decisions about who moves and who remains, this reticence is not a weakness in the work; it is a substantial part of what the work is demonstrating.

The index does not know whether anyone should evacuate. It is built to make the question harder to answer carelessly.

References

Sources

  1. 01ACAPS and the Joint Research Centre of the European Commission. INFORM Severity Index, April 2026 release, 104 crises.
  2. 02International Committee of the Red Cross. Article 49, Geneva Convention (IV) relative to the Protection of Civilian Persons in Time of War, 1949.
  3. 03ACLED. Armed Conflict Location and Event Data, conflict events and fatalities.
  4. 04Open-Meteo. Route weather and daylight data.
  5. 05Microsoft and Planet. HASTE, satellite damage assessment framework.
  6. 06International Organization for Migration. Risk Index for Climate Displacement, two tier macro and micro model.
  7. 07Centers for Disease Control and Prevention. Social Vulnerability Index.
  8. 08Fund for Peace. Fragile States Index.
  9. 09OECD and the Joint Research Centre. (2008). Handbook on Constructing Composite Indicators. OECD Publishing.
  10. 10Saaty, T.L. (1990). How to Make a Decision: The Analytic Hierarchy Process. European Journal of Operational Research, volume 48, issue 1, pages 9 to 26.
  11. 11Beccari, B. (2016). A Comparative Analysis of Disaster Risk, Vulnerability and Resilience Composite Indicators. PLoS Currents Disasters.
  12. 12Al Fozaie, M. (2022). A Guide to Integrating Expert Opinion and Fuzzy AHP When Generating Weights for Composite Indices. Advances in Fuzzy Systems.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file diff --git a/static-site/print/evacuation-inform-index/index.txt b/static-site/print/evacuation-inform-index/index.txt index 793ce2c54..879a02255 100644 --- a/static-site/print/evacuation-inform-index/index.txt +++ b/static-site/print/evacuation-inform-index/index.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -12:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","print","evacuation-inform-index",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","evacuation-inform-index","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -14:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -16:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","print","evacuation-inform-index",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","evacuation-inform-index","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:"$Sreact.suspense" -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1c:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -29,7 +29,7 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 19:[] 10:"$W19" 11:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1e:Ta14, @page { size: 210mm 297mm; margin: 20mm 18mm; } @@ -126,6 +126,6 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 2d:T45e,Stated plainly, the proxy overstates evacuation risk for lower intensity crises inside conflict affected states, because national conflict conditions inflate complexity independently of the hazard being scored. It understates evacuation risk where one specific corridor is dangerous but the country as a whole is permissive, because complexity has no corridor level resolution at all. A third pattern emerged that no reviewer anticipated: across the seventeen most severe crises the ratio ranges only from 0.83 to 1.13, clustering so tightly around 1.0 that it discriminates almost nothing. Every large ratio in the index comes from a low or mid severity crisis. The headline number is most confident precisely where it is least meaningful, which is why the tool now warns at the point of reading whenever both components sit near the top of the scale. The substitution is kept because no public dataset scores civilian evacuation difficulty across 104 crises and complexity is the closest construct with real coverage and a published methodology, but it should be read as a screening prompt rather than a measurement.2b:["$","p","7",{"children":"$2d"}] 2c:["$","p","8",{"children":"Indicators combine by weighted geometric mean rather than arithmetic mean. The practical consequence is that a catastrophic score on one component cannot be averaged away by comfortable scores elsewhere. If every evacuation route is closed, no amount of favourable weather or food security compensates for it. This is the same non compensatory logic used by the Human Development Index, and it reflects a judgement about the world rather than a mathematical preference: some failures are absolute."}] 1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -2e:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +2e:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 18:null 1d:[["$","title","0",{"children":"The Evacuation Inform Index (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L2e","6",{}]] diff --git a/static-site/print/evacuation-simulation/__next._full.txt b/static-site/print/evacuation-simulation/__next._full.txt index 1959d59b8..8ef4b2eef 100644 --- a/static-site/print/evacuation-simulation/__next._full.txt +++ b/static-site/print/evacuation-simulation/__next._full.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -12:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","print","evacuation-simulation",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","evacuation-simulation","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -14:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -16:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","print","evacuation-simulation",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","evacuation-simulation","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:"$Sreact.suspense" -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1c:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -29,7 +29,7 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 19:[] 10:"$W19" 11:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1e:Ta14, @page { size: 210mm 297mm; margin: 20mm 18mm; } @@ -171,6 +171,6 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 57:["$","li","21",{"children":[["$","span",null,{"className":"print-ref-num","children":"22"}],["$","span",null,{"children":"International Committee of the Red Cross. Customary International Humanitarian Law Study, Rule 131. Treatment of displaced persons."}]]}] 58:["$","li","22",{"children":[["$","span",null,{"className":"print-ref-num","children":"23"}],["$","span",null,{"children":"International Committee of the Red Cross. Customary International Humanitarian Law Study, Rule 138. The elderly, disabled and infirm."}]]}] 1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -59:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +59:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 18:null 1d:[["$","title","0",{"children":"The Evacuation Simulator (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L59","6",{}]] diff --git a/static-site/print/evacuation-simulation/__next._head.txt b/static-site/print/evacuation-simulation/__next._head.txt index d059d2191..0a0cb325a 100644 --- a/static-site/print/evacuation-simulation/__next._head.txt +++ b/static-site/print/evacuation-simulation/__next._head.txt @@ -1,6 +1,6 @@ 1:"$Sreact.fragment" -2:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -3:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +2:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +3:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 4:"$Sreact.suspense" -5:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +5:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 0:{"rsc":["$","$1","h",{"children":[null,["$","$L2",null,{"children":[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]]}],["$","div",null,{"hidden":true,"children":["$","$L3",null,{"children":["$","$4",null,{"name":"Next.Metadata","children":[["$","title","0",{"children":"The Evacuation Simulator (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L5","6",{}]]}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/print/evacuation-simulation/__next._index.txt b/static-site/print/evacuation-simulation/__next._index.txt index f98a4fd90..18b9d82b6 100644 --- a/static-site/print/evacuation-simulation/__next._index.txt +++ b/static-site/print/evacuation-simulation/__next._index.txt @@ -1,13 +1,13 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -d:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +d:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} 8:["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$Ld",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$Ld",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$Ld",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$Ld",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] diff --git a/static-site/print/evacuation-simulation/__next._tree.txt b/static-site/print/evacuation-simulation/__next._tree.txt index e1dc980ec..07b9e6415 100644 --- a/static-site/print/evacuation-simulation/__next._tree.txt +++ b/static-site/print/evacuation-simulation/__next._tree.txt @@ -1,7 +1,7 @@ -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] 0:{"tree":{"name":"","param":null,"prefetchHints":16,"slots":{"children":{"name":"print","param":null,"prefetchHints":0,"slots":{"children":{"name":"slug","param":{"type":"d","key":"evacuation-simulation","siblings":null},"prefetchHints":0,"slots":{"children":{"name":"__PAGE__","param":null,"prefetchHints":0,"slots":null}}}}}}},"staleTime":300,"buildId":"static-snapshot"} diff --git a/static-site/print/evacuation-simulation/__next.print.txt b/static-site/print/evacuation-simulation/__next.print.txt index 460afe242..fa617e997 100644 --- a/static-site/print/evacuation-simulation/__next.print.txt +++ b/static-site/print/evacuation-simulation/__next.print.txt @@ -1,4 +1,4 @@ 1:"$Sreact.fragment" -2:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -3:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +2:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +3:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] 0:{"rsc":["$","$1","c",{"children":[null,[["$","style",null,{"children":"\n /* The site is dark-first, and a dark color-scheme makes Chrome paint\n the printed sheet's margin area dark too. Paper is light. */\n html { color-scheme: light !important; }\n html, body { background: #ffffff !important; }\n body::after { content: none !important; }\n header, footer, .site-bg, .aura { display: none !important; }\n "}],["$","$L2",null,{"parallelRouterKey":"children","template":["$","$L3",null,{}]}]]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/print/evacuation-simulation/index.html b/static-site/print/evacuation-simulation/index.html index 6292c70e1..a16896516 100644 --- a/static-site/print/evacuation-simulation/index.html +++ b/static-site/print/evacuation-simulation/index.html @@ -1,4 +1,4 @@ -The Evacuation Simulator (print edition) · NYU Ethical Tech CoLab

Publications · Academic report

The Evacuation Simulator

An Agent-Based Model of How Civilians Decide to Leave During Armed Conflict

Ethical Tech CoLab

NYU Center for Global Affairs

July 2026

Melanie MacKew. Developed as part of masters research at the NYU Center for Global Affairs, under the Ethical Tech CoLab.

Warnings are issued more often than they are believed. Most people do not run when an alert goes out; they telephone a relative, they step outside to see whether the family next door is loading a car, they wait for a second source to say the same thing. That interval between the warning and the departure is where humanitarian evacuations are won or lost, and EvacSim is a research prototype that makes it visible: a small simulated community of six households, each person moving through hearing, believing, preparing, and leaving, with the timing driven by who they are and what information reaches them.

Key figures

6

households in the simulated community, kept small so every person stays individually visible and traceable

4

information channels supplying confirmations: official broadcast, humanitarian actor, neighbours, and hostile misinformation

5.5

the identical travel speed every category reaches by car, where on foot a child moves at 1.5 and an adult at 2.6

10

limitations the report states about the prototype, beginning with the fact that nobody in the model can be harmed

01

Executive Summary

EvacSim is an interactive simulation, a piece of software that plays out an imagined situation on screen so that a user can watch it unfold and change the conditions. It models a community of six households receiving an evacuation warning during an armed conflict, and it shows who leaves, who is delayed, and who does not get out at all.

The tool is built on a simple idea drawn from disaster sociology: people do not act on a single warning. Each simulated person needs a certain number of independent confirmations before they will start preparing to leave. Everything else in the model follows from how quickly those confirmations arrive and how long preparation and travel then take.

Confirmations can arrive through four channels: an official broadcast, a humanitarian aid organisation, neighbours who can be seen already moving, and a hostile misinformation channel that supplies confirmations which are convincing but false. Which channel actually drove each household to move is recorded and reported at the end of every run.

The model treats vulnerability as a matter of timing rather than a score. Elders, children under five, pregnant women, and unaccompanied minors are each given specific delays in preparation, movement speed, and the number of confirmations required. The consequences are then allowed to play out. When a corridor closes on a schedule, it is these categories that are most often left behind, and the tool shows exactly that.

Escape is routed through corridors: four named gates on the edges of the map that can be opened, closed, or timed to open and close mid-run. This models a negotiated humanitarian passage with a fixed validity window. When all gates are shut, people who are ready to move are marked as trapped, which is the simulation's representation of siege.

The prototype is delivered as a web page that runs in an ordinary browser and is published as a public demonstration. It carries an extensive built-in guide explaining its own mechanics and their legal grounding.

The tool is honest, in its own documentation, that its timings are qualitative approximations rather than measured real-world durations. This report endorses that caution and adds one further caveat: several of the International Humanitarian Law citations in the repository are inaccurate and should be corrected before the tool is used in teaching. Section 09 sets out which ones and what the correct provisions are.

02

Background and Rationale

The problem. Evacuation planning in armed conflict is usually discussed in terms of capacity: how many buses, how many kilometres of road, how many hours of ceasefire. These are the easy quantities to count. They are also the least predictive. A corridor with ample capacity fails if the population does not believe the corridor exists, does not trust the party announcing it, or cannot physically reach it in the time allowed.

The gap. The behavioural side of evacuation is well documented in disaster sociology, but that literature grew out of hurricanes, floods, and industrial accidents. It assumes a benign movement environment: the warning is issued in good faith, the roads are open, and nobody is trying to deceive you. Armed conflict inverts all three assumptions. The threat is deliberate, the information is contested, and movement itself may be obstructed or compelled.

The response. EvacSim takes the established behavioural model, the confirmation-seeking household that moves as a unit, and places it inside a conflict environment. It adds the things that make conflict evacuation different: corridors that can close, checkpoints that impose delay, telecommunications that degrade under bombardment, deliberate misinformation about routes, and coercion that removes the household's choice entirely.

The repository is explicit that the behavioural foundations come from two named bodies of work. Enrico Quarantelli, of the Disaster Research Center at the University of Delaware, established that people rarely panic in disasters and instead seek confirmation from multiple sources before accepting that a threat is real. Thomas Drabek, of the University of Denver, documented that families evacuate as units rather than as individuals, waiting until all members are present before departing. These two findings are the direct source of the model's two central mechanics: the confirmation counter and the household hub that waits for its slowest member.

03

Objectives

The tool is designed to:

  1. Show how the quality and source of information, rather than the severity of the threat alone, determines how quickly a population moves.
  2. Make the cost of vulnerability legible in units of time, so that the delay imposed by an elder or an unaccompanied child can be compared directly against the length of a corridor window.
  3. Demonstrate the asymmetry between modes of transport, showing that the same obstruction which is a minor inconvenience to a household with a car may be fatal to a household on foot.
  4. Represent the operational conditions that International Humanitarian Law regulates, including corridor access, checkpoint obstruction, attacks on communications infrastructure, perfidious misinformation, and forced displacement, so that legal rules can be seen as behavioural consequences rather than abstract prohibitions.
  5. Support comparison between runs, so that a user can change one condition, run the scenario again, and see what that single change cost the population.
  6. Serve as a teaching instrument for students and humanitarian practitioners rather than as an operational planning tool.

04

How the Simulation Works

The community. The simulation places six households on a rectangular map, given the family names Rivera, Kim, Okafor, Hassan, Novak, and Tanaka. Each household has one member designated as its hub, drawn at the centre of the cluster, with the remaining members arranged around it. The hub represents the household as a coordinating unit. It does not depart until the slowest person in the household is ready, which is how the model implements Drabek's finding that families move together.

Households are linked to one another in a social network. Each family is connected to the family on either side of it in a ring, and also to the family two positions away. The result is that every household can see four of the other five. Only these linked households can influence one another.

Every individual passes through five stages in order:

  1. Unaware. The person has not yet heard anything.
  2. Seeking. The person has heard an alert and is now looking for corroboration.
  3. Milling. The person believes the alert and is preparing to leave: gathering family members, packing, securing the house.
  4. Evacuating. The person is physically moving toward an exit.
  5. Evacuated. The person has reached safety and is out of the simulation.

Each transition is probabilistic, meaning it is decided by a weighted chance each time the clock advances rather than by a fixed rule. Two people in identical circumstances will not necessarily move at the same moment.

The clock. Time in the simulation is measured in ticks. A tick is a deliberately abstract unit. When the simulation is running at normal speed one tick elapses every 200 milliseconds of real time, roughly five ticks per second, but a tick is not claimed to correspond to any particular number of real-world minutes. The documentation states the reason plainly: the timing values were chosen to produce realistic relative behaviour, not calibrated against measured durations, and labelling ticks as minutes would imply a precision the model does not have. This is an unusually candid design choice and it should be respected when interpreting any result.

A person in the Seeking stage accumulates confirmations, and the model records which source supplied the final one that tipped the person into Milling. Confirmations arrive through four channels.

  • The official broadcast. A single information node at the centre of the map, representing a government or military announcement, with its reliability set directly by the information clarity setting.
  • The humanitarian actor. A separate node in the upper left of the map, representing an aid organisation such as the ICRC. Its reliability is fixed at 0.75, higher than a government broadcast at typical clarity settings, but it reaches only a fraction of households, determined by the humanitarian access setting. This encodes the operational reality that a neutral organisation is more trusted but must negotiate for physical access.
  • Neighbours. If any linked household has a member who is already milling or evacuating, a person in the Seeking stage may take that as a confirmation. This is social contagion, and it is the mechanism by which an evacuation cascades through a community.
  • Misinformation. A hostile channel that supplies confirmations which count toward the threshold exactly as genuine ones do, but which then send the person to the wrong exit.

05

The Variables Explained

Every setting a user can change is listed here with what it represents, why it takes the range it does, and how it reaches the outcome. Where a number is quoted, it is the number in the source code.

Threat level. A whole number from 1 to 10, set to 6 by default. This represents how severe and how visible the danger is. It works by setting the chance, each tick, that any given unaware person hears an alert for the first time. The formula is the threat level divided by ten, multiplied by 0.35, with a small additional term that rises with elapsed time, and the result held between 2 and 55 per cent. At the default setting of 6 an unaware person has roughly a one in five chance of hearing something in any given tick at the start of a run, rising slowly as the run continues. The important point is that threat level in this model governs how fast news travels, not how much damage is done. Nobody is injured or killed in this simulation.

Threat rise rate. A whole number from 0 to 20, set to 0 by default. At zero the threat level stays fixed for the entire run. Above zero the threat climbs as the run proceeds, adding one level of threat every hundred ticks divided by the setting, and the climb stops at the maximum of 10. This models the compressing decision window that characterises conflict displacement: a household that hesitates finds that the situation has deteriorated around it while it was deciding.

Information clarity. A whole number from 1 to 10, set to 5 by default. This is the single most consequential setting in the model, because it acts in two places at once. First, it sets the reliability of the official broadcast node, which is simply clarity divided by ten, held between 0.1 and 0.95. That reliability then drives the per-tick chance that a seeking person receives a confirmation. Second, and more sharply, if clarity is below 4 then every person in the community is given one or two extra confirmations to obtain before they will move. Poor information is therefore doubly punishing: each confirmation is slower to arrive, and more of them are needed. Populations that have been given vague, contradictory, or previously false warnings become more sceptical, and their scepticism is rational.

Humanitarian access. A percentage from 0 to 100, set to 40 by default. This represents how far an aid organisation has been permitted to operate. At the moment the simulation is built, each individual is independently assigned as reachable or not reachable by the humanitarian node, with the probability equal to the access setting. Someone who is not reachable never receives anything from that channel, no matter how long the run lasts. Those who are reachable receive confirmations from it at a fixed chance of 37.5 per cent per tick. Setting this to zero removes the humanitarian actor entirely and is the simulation's model of denied access.

The four armed conflict settings are all set to zero by default, meaning the simulation begins in a relatively benign state which the user then degrades deliberately.

Checkpoint delay. A whole number of ticks from 0 to 15. When above zero, each person has a 55 per cent chance of being stopped at the moment they finish preparing and begin to move, with a random delay added to their travel time. This represents document checks and security screening. The design choice worth noting is that the delay is applied to everyone equally, which means its humanitarian effect is not equal: a household that was already close to the edge of the corridor window is pushed over it, while a fast household absorbs the same delay without consequence.

Misinformation. A percentage from 0 to 100. When above zero, a person in the Seeking stage has a per-tick chance equal to the setting multiplied by 0.35 of receiving a false confirmation, so a setting of 60 per cent produces a 21 per cent chance per tick. The false confirmation counts toward the threshold exactly as a genuine one does, which is the point: the household cannot tell the difference. The consequence appears later. If the confirmation that finally tipped a person into preparing came from the misinformation channel, then when that person begins to move they are routed not to their nearest open gate but to a randomly chosen different open gate. If only one gate is open there is no wrong way to send them, and the misdirection has no effect, which is a small but real artefact of the implementation.

Infrastructure damage. A whole number from 0 to 20, modelling the destruction of telecommunications. When above zero, the effective information clarity falls every tick, reduced by the elapsed ticks multiplied by the setting and divided by fifteen, with a floor of 1. Both the broadcast node's reliability and the per-tick confirmation chance fall with it. The households punished by this setting are precisely the slowest ones, because they are still seeking confirmation at the point when the information environment collapses.

Two observations about that setting. First, it is far more aggressive than the guide describes. The built-in guide states that at a setting of 10 a starting clarity of 7 will have fallen to around 4 by tick 45. The formula in the source code does not produce that result. A setting of 10 reduces clarity by two thirds of a point every tick, so a starting clarity of 7 reaches the floor of 1 in around nine ticks, not forty-five. Even the setting labelled light damage in the interface collapses clarity from 7 to the floor within about thirteen ticks. Users should treat any non-zero value here as modelling a rapid communications blackout rather than a gradual decline. Second, the extra confirmations imposed by low clarity are assigned once when the run is built and are not increased retrospectively as clarity degrades, so degradation makes each confirmation harder to obtain but does not raise the threshold that must be reached.

Coercion risk. A percentage from 0 to 100, representing forced displacement. When above zero, an unaware person who has not received any alert in a given tick has a chance of being pushed directly into the Milling stage without ever seeking or receiving confirmation. Readers should note how small this is in practice: even at maximum threat and maximum coercion the per-tick chance is 0.8 per cent, well below the cap of 6 per cent, so coercion accumulates slowly over a long run rather than sweeping through the community. Coerced individuals are flagged, drawn with a red ring, logged individually, and counted separately in the end-of-run summary. They may well reach safety faster than their neighbours, because they skipped the entire confirmation process, and the tool is right to insist that this speed is not a good outcome. They left without preparation, without verifying the route, and without choosing to go.

Average family size. A whole number from 1 to 7, set to 3 by default. The actual size of each household is drawn at random between one below and one above this figure, so households vary. Larger households are slower for a structural reason: the hub adopts the longest preparation time and the longest travel time of any member, so every additional person is another chance of drawing a slow one.

Elder ratio. A percentage from 0 to 60, set to 20 by default. Each non-hub member of each household is independently tested against this percentage to determine whether they are an elder. An elder receives one additional required confirmation, extra preparation time, and a reduced movement speed. The extra confirmation is the mechanically interesting part. It encodes the finding that older people are more likely to want a second or third source before accepting a warning, particularly where there is attachment to a home and scepticism toward the announcing authority.

Children under five. A percentage from 0 to 50, set to 20 by default, applied to those not already assigned as elders. Young children receive the longest preparation penalty of the two original categories, reflecting the work of gathering and packing for a small child, and the slowest movement speed of any category on foot. They do not receive extra required confirmations, because a small child is not the one deciding.

Pregnant women. A percentage from 0 to 30, set to 0 by default. Preparation delay is comparable to an elder and movement speed sits between an elder and an adult. No extra confirmations are required. This category was added specifically to represent a group that International Humanitarian Law singles out for protection.

Unaccompanied minors. A percentage from 0 to 20, set to 0 by default. This is the most severely penalised category in the model and deliberately so. An unaccompanied minor requires two additional confirmations, more than any other category, and receives by far the largest preparation delay: six to twelve extra ticks on foot, against two to five for an elder. They then move at the slow speed of a child. A separated child cannot simply leave. Somebody must be found who is authorised to take responsibility for them, and family tracing and reunification take time that no other category incurs. In a run with a timed corridor window, this is the category that gets left behind, and demonstrating that is the point of including it.

One important structural detail. These four categories are assigned in strict priority order and are mutually exclusive: a person is tested for elder first, then child, then pregnant, then unaccompanied minor, and once assigned cannot also be something else. A pregnant elder or a child who is also unaccompanied cannot exist in this model. Household hubs are never assigned any vulnerability category at all. The practical effect is that raising the elder percentage quietly suppresses the number of people available to be assigned to the later categories, so the sliders are not independent of one another.

Neighbour influence. A percentage from 0 to 100, set to 55 by default. Each tick, if any of a household's four linked neighbours has a member visibly milling or evacuating, every seeking member of that household has a chance equal to this setting of gaining a confirmation from that observation alone. At high settings the community behaves as a cascade, where one household departing pulls the rest out behind it. At zero, social observation counts for nothing and only official and humanitarian channels matter.

Three quantities are not exposed as settings but are generated for each individual when the run is built, and they are what the settings above actually act upon. Confirmations needed is a random whole number from 1 to 3, plus one or two more if information clarity is below 4, plus a category penalty. It is the single number that most determines who leaves early and who leaves late. Preparation time is drawn at random from a range that depends on the scenario, with an additional draw added for the person's vulnerability category. Travel time is drawn the same way, and speed is a fixed value per category per scenario, measured in screen pixels per tick.

CategoryExtra confirmationsExtra preparationSpeed on foot
AdultNoneNone2.6
Elder12 to 5 ticks1.8
Child under fiveNone3 to 6 ticks1.5
Pregnant womanNone2 to 5 ticks2.0
Unaccompanied minor26 to 12 ticks1.5
How each category is penalised in the pedestrian scenario. The base preparation range is 2 to 4 ticks for everyone, and the category delay is added on top.

The scenario setting. The user chooses one of three modes of evacuation, and this choice rewrites all the preparation and travel figures at once. The three modes are not simply faster or slower versions of each other. They differ in shape.

ModeBase preparationTravel speed
Pedestrian2 to 4 ticks1.5 for a child up to 2.6 for an adult
Car1 to 3 ticks5.5 for every category
Train4 to 8 ticks8.0 for every category

Pedestrian is the slowest and the most unequal, and vulnerability matters more there than in any other mode. Car equalises mobility almost completely, and this is the model's clearest single finding: access to transport does not merely make evacuation faster, it makes vulnerability stop mattering. Train is the reverse profile, where the long preparation represents the wait for a departure but the journey itself is the fastest in the model. The train scenario has no corridor gates at all; passengers leave by the rail line.

06

Corridors: How People Actually Get Out

In the pedestrian and car scenarios, escape is not in all directions. Four gates are placed at the midpoints of the four edges of the map and named North, South, East, and West. When a person finishes preparing, they identify the nearest gate that is currently open and move toward it. The train scenario has no gates.

Each gate has three settings.

  • It can be open or closed at the start.
  • It can be given a closing tick, at which it shuts.
  • It can be given an opening tick, at which it opens, in which case it begins the run closed.

Combining an opening tick with a closing tick produces a humanitarian window: a gate that opens at tick 15 and closes at tick 35 gives the population a twenty-tick passage. This is the mechanic that makes every preparation delay described above consequential rather than merely descriptive. An eight-tick elder delay is invisible in a run with no time limit and decisive in a run with a twenty-tick window.

When a gate closes mid-run, anyone already travelling toward it recalculates and heads for the nearest remaining open gate, and the reroute is logged individually. Rerouting costs time, and the cost is much higher on foot than by car, which is the asymmetry the documentation identifies as a research finding in its own right. The same closure that mildly inconveniences a household with a vehicle can leave a household on foot unable to reach any alternative.

If no gate is open when a person is ready to move, they are marked as trapped. Trapped individuals stop, are logged by name, and are counted in the final summary. A run ends when every person is either evacuated or trapped. This is the simulation's representation of encirclement and siege, and it is the condition the tool is most clearly designed to make visible: the visual and numerical gap between the households that got out and the households that were still preparing when the gate shut.

07

Reading the Results

While a run is in progress the map is the primary output. Individuals change colour as they pass through the five stages, from grey when unaware, through blue when seeking, amber when milling, red when evacuating, to green when evacuated. Each population category has its own shape, so an elder, a child, a pregnant woman, and an unaccompanied minor can be distinguished at a glance. Lines drawn between the information nodes and individuals show confirmations arriving, colour-coded by which channel supplied them, with the hostile misinformation channel drawn in red.

Below the map, three rows of figures update every tick: how many people are in each of the five stages, how many belong to each vulnerability category, and a single prominent percentage showing how much of the population has completed evacuation.

An event log records everything in plain sentences as it happens: who received an alert, how many confirmations they now hold, who was held at a checkpoint and for how long, which corridor closed and who is rerouting, who was coerced, and who was trapped. The log is capped at the most recent eighty entries. It is the most useful output for teaching, because it converts the statistical outcome back into individual stories.

When a run ends, a summary panel reports the total number of ticks, the average time the population spent in each of the seeking, milling, and evacuating stages, which household finished last, and what the model believes caused that household to be slowest, expressed as its composition. It also reports the channel split, along with counts of those coerced, those held at checkpoints, and those trapped.

The channel split is the analytically richest output. It answers a question that matters operationally: was this evacuation driven by the authorities or by the community? A run in which social confirmation dominates is a run in which official communication failed and the population compensated for it. A run in which the humanitarian channel dominates is a run in which access mattered more than broadcasting. A run in which misinformation dominates is a run in which the population moved decisively in the wrong direction.

The last five runs are retained, and one can be pinned for direct comparison. This is the intended way to use the tool. A single run tells the user very little, because so much is decided by random draws. The signal appears in the difference between two runs that were identical except for the one thing the user changed.

08

Methodological Choices

Why time is abstract. The decision to measure time in ticks rather than minutes is the most defensible choice in the project. The preparation and travel figures were selected to produce plausible relative behaviour, and there is no empirical calibration behind them. Converting them to minutes would manufacture a precision that does not exist. The documentation says so directly and points readers toward calibrated evacuation timing studies for anyone who needs real durations.

Why vulnerability is expressed as delay rather than as a risk score. The model never assigns anyone a number representing how much danger they are in. Instead it gives them extra ticks. This is a meaningful choice. It means vulnerability only produces a bad outcome when it interacts with a constraint, such as a closing corridor or a degrading information environment. In a run with no time pressure, a household full of elders arrives late and arrives safely. That is a more honest representation of how vulnerability actually operates than a standing risk score would be.

Why the household waits. The hub takes the maximum preparation and travel time of anyone in the household, so the whole family moves at the pace of its slowest member. This directly implements Drabek's finding and it is the mechanism by which one unaccompanied minor or one elder slows an entire household rather than only themselves.

Why everything is probabilistic. Almost every transition in the model is decided by a random draw. Two runs with identical settings will produce different results. This is deliberate and correct, since it reflects genuine uncertainty in warning response, but it has a consequence for how the tool should be used. Conclusions should never be drawn from a single run.

Why the population is small. Six households and typically fifteen to twenty individuals is a very small population. This keeps every individual visible and traceable on screen, which serves the teaching purpose, but it means the results are statistically noisy and no percentage produced by a single run should be treated as an estimate of anything.

09

Grounding in International Humanitarian Law

The repository connects each of its mechanics to a provision of International Humanitarian Law, the body of law that regulates the conduct of hostilities and protects those not taking part in the fighting. This is a valuable design feature: it turns legal rules into observable behaviour.

The following provisions are cited in the repository and are cited correctly.

  • Additional Protocol II, Article 17(1), which provides that the displacement of the civilian population shall not be ordered for reasons related to the conflict unless the security of the civilians involved or imperative military reasons so demand, and that where displacement does occur all possible measures shall be taken so that the population is received under satisfactory conditions. Article 17(2) provides that civilians shall not be compelled to leave their own territory. This is the correct basis for the coercion mechanic.
  • Customary International Humanitarian Law Rule 129, which prohibits parties from deporting or forcibly transferring the civilian population of an occupied territory, and from ordering the displacement of the civilian population in non-international armed conflict, subject to the same two exceptions. Rule 131 requires that displaced civilians be received in satisfactory conditions and that families not be separated.
  • Additional Protocol I, Article 37, which prohibits killing, injuring, or capturing an adversary by perfidy, and Article 38, which prohibits improper use of the red cross and red crescent emblems, the flag of truce, and the emblem of the United Nations. These are the bases the repository gives for the misinformation mechanic, subject to the framing point noted at the end of this section.
  • Additional Protocol I, Article 52, which provides that civilian objects shall not be the object of attack, that attacks shall be limited strictly to military objectives, and that where there is doubt whether a normally civilian object is being used militarily it shall be presumed not to be. This is the correct basis for the infrastructure damage mechanic.
  • Additional Protocol I, Article 58, which requires parties to endeavour to remove civilians from the vicinity of military objectives and to take other necessary precautions to protect civilians under their control. Its opening words expressly preserve Article 49 of the Fourth Geneva Convention, which matters: the duty to move civilians away from danger cannot be used to justify displacement that Article 49 forbids.
  • Customary Rule 53 and Additional Protocol I, Article 54, which prohibit starvation of the civilian population as a method of warfare. Article 54(2) is particularly apposite to the trapped mechanic, because it prohibits destroying or rendering useless objects indispensable to civilian survival whatever the motive, expressly including the motive of causing civilians to move away.
  • Additional Protocol I, Article 70 and Customary Rule 55, which require parties to allow and facilitate rapid and unimpeded passage of impartial humanitarian relief for civilians in need, subject to a right of control. This is the correct basis for the humanitarian access setting.
  • Additional Protocol I, Article 57(2)(c) and Customary Rule 20, which require each party to give effective advance warning of attacks which may affect the civilian population unless circumstances do not permit. A high-threat, low-clarity run is a fair representation of a warning that was issued but was too vague to act on.
  • Customary Rule 138, which provides that the elderly, disabled, and infirm affected by armed conflict are entitled to special respect and protection. This supports the elder mechanic.

Four citations in the repository documentation are inaccurate and should be corrected. This is not a criticism of the project's substance, which is sound. It is a matter of the accuracy that a teaching tool on legal subjects requires.

Customary Rule 99. It is cited repeatedly, in the corridor documentation, the checkpoint documentation, and the built-in guide, as the rule requiring parties to allow civilians to leave and prohibiting disproportionate impediment to civilian movement. Rule 99 does not say this. Rule 99 provides that arbitrary deprivation of liberty is prohibited. It is a detention rule, situated in the chapter of the ICRC Customary Law Study on fundamental guarantees. There is in fact no single customary rule creating a general civilian right to leave a besieged area. The nearest provision is Article 17 of the Fourth Geneva Convention, which requires parties to endeavour to conclude local agreements for the removal from besieged or encircled areas of the wounded, sick, infirm, and aged, of children, and of maternity cases. Two limits of that article are instructive for this project: it is only an obligation to endeavour, and it covers specific vulnerable categories rather than the whole civilian population.

Article 16 of Additional Protocol I. It is cited in the population settings as the provision giving special protection to pregnant women. It is not. Article 16 of Additional Protocol I is titled General protection of medical duties and concerns the protection of medical personnel from being punished or compelled to act against medical ethics. The intended citation is almost certainly Article 16 of the Fourth Geneva Convention, which does provide that the wounded and sick, the infirm, and expectant mothers shall be the object of particular protection and respect. The stronger provision for this purpose is Article 8(a) of Additional Protocol I, which brings maternity cases, newborn babies, and expectant mothers within the definition of the wounded and sick and therefore within that entire protective regime.

Article 78 of Additional Protocol I. It is cited as the basis for the unaccompanied minors mechanic. The article is real and its content is relevant, but its scope is narrower than the repository implies. Article 78 governs the evacuation of children who are not a party's own nationals to a foreign country, and requires that such evacuation be temporary, justified by compelling reasons, consented to in writing by parents or legal guardians, and supervised by a Protecting Power. It does not regulate the internal evacuation of a state's own children, which is what the simulation depicts. For the protection of children generally, Article 77 of Additional Protocol I and Articles 24 and 50 of the Fourth Geneva Convention are the appropriate provisions. The reunification delay the model applies is nonetheless well founded: it reflects the tracing and reunification work that Article 26 of the Fourth Geneva Convention and Article 74 of Additional Protocol I require states to facilitate.

Customary Rule 9. It is cited alongside Article 52 of Additional Protocol I for the prohibition on attacking civilian objects. Rule 9 is a definition rather than a prohibition: it states that civilian objects are all objects that are not military objectives. The protective proposition is Rule 10, which provides that civilian objects are protected against attack unless and for such time as they are military objectives. Citing Rules 9 and 10 together would be correct.

One further point is a matter of framing rather than a citation error. The repository presents misinformation about evacuation routes as falling squarely within the prohibition on perfidy in Article 37. The connection is intuitive but the article is narrower than that: perfidy under Article 37 requires the intent to kill, injure, or capture an adversary by inviting their confidence in a protection under international law, and Article 37(2) expressly permits ruses of war and names misinformation among them. False information directed at a civilian population to endanger it engages other rules, including the general protection of civilians and the prohibition on acts whose primary purpose is to spread terror, more directly than it engages Article 37. The mechanic is valuable and the concern is real. The legal framing would benefit from revision.

10

Limitations and Caveats

Nobody is harmed. The model has no concept of injury or death. Being trapped is the worst outcome available, and being trapped is recorded as a status rather than as a consequence. Corridor violations, attacks on convoys, and secondary threats during transit are all identified in the repository's own planning documents as desirable additions and none are implemented.

The timings are not calibrated. This bears repeating because it constrains every quantitative statement the tool makes. The preparation ranges, movement speeds, and confirmation thresholds are plausible relative values chosen by the author. They are not derived from evacuation data.

The population is tiny and the results are noisy. With six households, a single unlucky random draw can change a headline figure substantially. No number from a single run should be quoted.

The vulnerability categories cannot overlap. Because a person is tested for each category in a fixed order and assigned to the first that matches, the model cannot represent a pregnant woman with a disability, an elderly person who is also chronically ill, or a child who is both under five and unaccompanied. Compounding vulnerability, which is exactly what determines outcomes in real displacement, is outside the model.

Several important categories are absent. The wounded and sick, persons with disabilities, and people lacking identity documents are all discussed at length in the repository's extensions document and none are implemented. The first of these is the most significant omission, since the wounded and sick hold the strongest claim to priority evacuation in the entire legal framework.

The threat is spatially uniform. Everyone on the map faces the same threat level regardless of where they stand. There is no front line, no direction from which danger approaches, and no reason why a household near the fighting should be alerted before one further away. The repository identifies directional threat as an important missing feature and it remains missing.

Coercion is rare in practice. Because of the very small coefficient in the formula, even the maximum coercion setting produces a low per-tick probability. Users who set the slider to a high value and observe few coercion events are not misreading the display.

Misinformation has no effect when only one gate is open. The mechanic works by misrouting people to a non-nearest gate, so it requires at least two open gates to do anything at all beyond arriving faster. In a single-corridor scenario, misinformation makes the population move sooner and no worse.

There is no test suite. The repository contains no automated tests, so the behaviour of the model rests on visual inspection.

Above all, this is a teaching model. It is designed to make a small number of relationships vivid, not to forecast what a real population will do. Its outputs should never inform an operational decision.

11

Practical Nature and Intended Audience

The tool is a web page. It runs in an ordinary browser with nothing to install and is published as a public demonstration. It opens on a guide page rather than on the simulation itself, which is a deliberate and correct choice: the guide walks the reader through the five stages, the information channels, the population categories, the corridor system, the armed conflict mechanics, and the legal framework before they touch a single control. The guide is longer than the simulation code that it explains.

The intended users are students and humanitarian practitioners. The tool suits a classroom particularly well, because the event log converts every statistical outcome back into a named person with a household and a reason for their delay. A discussion of why the Okafor family was still milling when the North corridor closed is a more productive discussion than a discussion of a percentage.

The most instructive exercise the tool supports is comparative. Run the same community on foot and by car with an identical corridor window and identical demographics, and the difference in who gets out is entirely attributable to access to transport. That single comparison carries the project's central argument more effectively than any of its individual settings.

12

Conclusion

EvacSim makes one argument well. Evacuation outcomes in armed conflict are determined less by the severity of the danger than by the interaction between how quickly information becomes credible and how long particular people take to act on it. The tool holds those two quantities apart and lets a user vary each independently, and its most valuable outputs are the ones that show where they collide: the elder still preparing when the corridor closes, the unaccompanied child awaiting an escort that arrives after the window has passed.

Its treatment of vulnerability deserves particular note. By expressing protected status as time rather than as a risk score, the model produces a result that is easy to state and hard to argue with. Vulnerability is harmless until it meets a constraint, and every constraint that International Humanitarian Law regulates, the closed corridor, the checkpoint, the destroyed transmitter, the false direction, works by shortening the time available to the people who need the most of it.

The prototype is modest about its status and should be. Its timings are uncalibrated, its population is small, nobody in it can be hurt, and its vulnerability categories cannot overlap in the ways that matter most in real displacement. Its legal citations require correction before it is used to teach law. What it offers is not prediction but a way of seeing, and for that purpose the choice to keep every simulated person individually visible, named, and traceable through the event log is worth more than any additional mechanic would be.

References

Sources

  1. 01Quarantelli, E.L. Disaster Research Center, University of Delaware. Work establishing that people rarely panic in disasters and instead seek confirmation from multiple sources before accepting that a threat is real.
  2. 02Drabek, T.E. University of Denver. Work documenting that families evacuate as units rather than as individuals, waiting until all members are present before departing.
  3. 03International Committee of the Red Cross. Article 17, Protocol Additional to the Geneva Conventions of 12 August 1949, and relating to the Protection of Victims of Non-International Armed Conflicts (Protocol II), 1977. Prohibition of forced movement of civilians.
  4. 04International Committee of the Red Cross. Article 17, Geneva Convention (IV) relative to the Protection of Civilian Persons in Time of War, 1949. Removal from besieged or encircled areas.
  5. 05International Committee of the Red Cross. Article 16, Geneva Convention (IV), 1949. Particular protection and respect for the wounded and sick, the infirm, and expectant mothers.
  6. 06International Committee of the Red Cross. Article 26, Geneva Convention (IV), 1949. Facilitating enquiries by dispersed families.
  7. 07International Committee of the Red Cross. Article 37, Protocol Additional to the Geneva Conventions of 12 August 1949, and relating to the Protection of Victims of International Armed Conflicts (Protocol I), 1977. Prohibition of perfidy, and permitted ruses of war.
  8. 08International Committee of the Red Cross. Article 52, Additional Protocol I, 1977. General protection of civilian objects.
  9. 09International Committee of the Red Cross. Article 54, Additional Protocol I, 1977. Protection of objects indispensable to the survival of the civilian population.
  10. 10International Committee of the Red Cross. Article 57, Additional Protocol I, 1977. Precautions in attack, including effective advance warning.
  11. 11International Committee of the Red Cross. Article 58, Additional Protocol I, 1977. Precautions against the effects of attacks.
  12. 12International Committee of the Red Cross. Article 70, Additional Protocol I, 1977. Relief actions, and priority in the distribution of relief.
  13. 13International Committee of the Red Cross. Article 77, Additional Protocol I, 1977. Protection of children.
  14. 14International Committee of the Red Cross. Article 78, Additional Protocol I, 1977. Evacuation of children to a foreign country.
  15. 15International Committee of the Red Cross. Customary International Humanitarian Law Study, Rule 9. Definition of civilian objects.
  16. 16International Committee of the Red Cross. Customary International Humanitarian Law Study, Rule 10. Civilian objects lose their protection only when they are military objectives.
  17. 17International Committee of the Red Cross. Customary International Humanitarian Law Study, Rule 20. Advance warning.
  18. 18International Committee of the Red Cross. Customary International Humanitarian Law Study, Rule 53. Starvation as a method of warfare.
  19. 19International Committee of the Red Cross. Customary International Humanitarian Law Study, Rule 55. Access for humanitarian relief to civilians in need.
  20. 20International Committee of the Red Cross. Customary International Humanitarian Law Study, Rule 99. Deprivation of liberty.
  21. 21International Committee of the Red Cross. Customary International Humanitarian Law Study, Rule 129. The act of displacement.
  22. 22International Committee of the Red Cross. Customary International Humanitarian Law Study, Rule 131. Treatment of displaced persons.
  23. 23International Committee of the Red Cross. Customary International Humanitarian Law Study, Rule 138. The elderly, disabled and infirm.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file +

Publications · Academic report

The Evacuation Simulator

An Agent-Based Model of How Civilians Decide to Leave During Armed Conflict

Ethical Tech CoLab

NYU Center for Global Affairs

July 2026

Melanie MacKew. Developed as part of masters research at the NYU Center for Global Affairs, under the Ethical Tech CoLab.

Warnings are issued more often than they are believed. Most people do not run when an alert goes out; they telephone a relative, they step outside to see whether the family next door is loading a car, they wait for a second source to say the same thing. That interval between the warning and the departure is where humanitarian evacuations are won or lost, and EvacSim is a research prototype that makes it visible: a small simulated community of six households, each person moving through hearing, believing, preparing, and leaving, with the timing driven by who they are and what information reaches them.

Key figures

6

households in the simulated community, kept small so every person stays individually visible and traceable

4

information channels supplying confirmations: official broadcast, humanitarian actor, neighbours, and hostile misinformation

5.5

the identical travel speed every category reaches by car, where on foot a child moves at 1.5 and an adult at 2.6

10

limitations the report states about the prototype, beginning with the fact that nobody in the model can be harmed

01

Executive Summary

EvacSim is an interactive simulation, a piece of software that plays out an imagined situation on screen so that a user can watch it unfold and change the conditions. It models a community of six households receiving an evacuation warning during an armed conflict, and it shows who leaves, who is delayed, and who does not get out at all.

The tool is built on a simple idea drawn from disaster sociology: people do not act on a single warning. Each simulated person needs a certain number of independent confirmations before they will start preparing to leave. Everything else in the model follows from how quickly those confirmations arrive and how long preparation and travel then take.

Confirmations can arrive through four channels: an official broadcast, a humanitarian aid organisation, neighbours who can be seen already moving, and a hostile misinformation channel that supplies confirmations which are convincing but false. Which channel actually drove each household to move is recorded and reported at the end of every run.

The model treats vulnerability as a matter of timing rather than a score. Elders, children under five, pregnant women, and unaccompanied minors are each given specific delays in preparation, movement speed, and the number of confirmations required. The consequences are then allowed to play out. When a corridor closes on a schedule, it is these categories that are most often left behind, and the tool shows exactly that.

Escape is routed through corridors: four named gates on the edges of the map that can be opened, closed, or timed to open and close mid-run. This models a negotiated humanitarian passage with a fixed validity window. When all gates are shut, people who are ready to move are marked as trapped, which is the simulation's representation of siege.

The prototype is delivered as a web page that runs in an ordinary browser and is published as a public demonstration. It carries an extensive built-in guide explaining its own mechanics and their legal grounding.

The tool is honest, in its own documentation, that its timings are qualitative approximations rather than measured real-world durations. This report endorses that caution and adds one further caveat: several of the International Humanitarian Law citations in the repository are inaccurate and should be corrected before the tool is used in teaching. Section 09 sets out which ones and what the correct provisions are.

02

Background and Rationale

The problem. Evacuation planning in armed conflict is usually discussed in terms of capacity: how many buses, how many kilometres of road, how many hours of ceasefire. These are the easy quantities to count. They are also the least predictive. A corridor with ample capacity fails if the population does not believe the corridor exists, does not trust the party announcing it, or cannot physically reach it in the time allowed.

The gap. The behavioural side of evacuation is well documented in disaster sociology, but that literature grew out of hurricanes, floods, and industrial accidents. It assumes a benign movement environment: the warning is issued in good faith, the roads are open, and nobody is trying to deceive you. Armed conflict inverts all three assumptions. The threat is deliberate, the information is contested, and movement itself may be obstructed or compelled.

The response. EvacSim takes the established behavioural model, the confirmation-seeking household that moves as a unit, and places it inside a conflict environment. It adds the things that make conflict evacuation different: corridors that can close, checkpoints that impose delay, telecommunications that degrade under bombardment, deliberate misinformation about routes, and coercion that removes the household's choice entirely.

The repository is explicit that the behavioural foundations come from two named bodies of work. Enrico Quarantelli, of the Disaster Research Center at the University of Delaware, established that people rarely panic in disasters and instead seek confirmation from multiple sources before accepting that a threat is real. Thomas Drabek, of the University of Denver, documented that families evacuate as units rather than as individuals, waiting until all members are present before departing. These two findings are the direct source of the model's two central mechanics: the confirmation counter and the household hub that waits for its slowest member.

03

Objectives

The tool is designed to:

  1. Show how the quality and source of information, rather than the severity of the threat alone, determines how quickly a population moves.
  2. Make the cost of vulnerability legible in units of time, so that the delay imposed by an elder or an unaccompanied child can be compared directly against the length of a corridor window.
  3. Demonstrate the asymmetry between modes of transport, showing that the same obstruction which is a minor inconvenience to a household with a car may be fatal to a household on foot.
  4. Represent the operational conditions that International Humanitarian Law regulates, including corridor access, checkpoint obstruction, attacks on communications infrastructure, perfidious misinformation, and forced displacement, so that legal rules can be seen as behavioural consequences rather than abstract prohibitions.
  5. Support comparison between runs, so that a user can change one condition, run the scenario again, and see what that single change cost the population.
  6. Serve as a teaching instrument for students and humanitarian practitioners rather than as an operational planning tool.

04

How the Simulation Works

The community. The simulation places six households on a rectangular map, given the family names Rivera, Kim, Okafor, Hassan, Novak, and Tanaka. Each household has one member designated as its hub, drawn at the centre of the cluster, with the remaining members arranged around it. The hub represents the household as a coordinating unit. It does not depart until the slowest person in the household is ready, which is how the model implements Drabek's finding that families move together.

Households are linked to one another in a social network. Each family is connected to the family on either side of it in a ring, and also to the family two positions away. The result is that every household can see four of the other five. Only these linked households can influence one another.

Every individual passes through five stages in order:

  1. Unaware. The person has not yet heard anything.
  2. Seeking. The person has heard an alert and is now looking for corroboration.
  3. Milling. The person believes the alert and is preparing to leave: gathering family members, packing, securing the house.
  4. Evacuating. The person is physically moving toward an exit.
  5. Evacuated. The person has reached safety and is out of the simulation.

Each transition is probabilistic, meaning it is decided by a weighted chance each time the clock advances rather than by a fixed rule. Two people in identical circumstances will not necessarily move at the same moment.

The clock. Time in the simulation is measured in ticks. A tick is a deliberately abstract unit. When the simulation is running at normal speed one tick elapses every 200 milliseconds of real time, roughly five ticks per second, but a tick is not claimed to correspond to any particular number of real-world minutes. The documentation states the reason plainly: the timing values were chosen to produce realistic relative behaviour, not calibrated against measured durations, and labelling ticks as minutes would imply a precision the model does not have. This is an unusually candid design choice and it should be respected when interpreting any result.

A person in the Seeking stage accumulates confirmations, and the model records which source supplied the final one that tipped the person into Milling. Confirmations arrive through four channels.

  • The official broadcast. A single information node at the centre of the map, representing a government or military announcement, with its reliability set directly by the information clarity setting.
  • The humanitarian actor. A separate node in the upper left of the map, representing an aid organisation such as the ICRC. Its reliability is fixed at 0.75, higher than a government broadcast at typical clarity settings, but it reaches only a fraction of households, determined by the humanitarian access setting. This encodes the operational reality that a neutral organisation is more trusted but must negotiate for physical access.
  • Neighbours. If any linked household has a member who is already milling or evacuating, a person in the Seeking stage may take that as a confirmation. This is social contagion, and it is the mechanism by which an evacuation cascades through a community.
  • Misinformation. A hostile channel that supplies confirmations which count toward the threshold exactly as genuine ones do, but which then send the person to the wrong exit.

05

The Variables Explained

Every setting a user can change is listed here with what it represents, why it takes the range it does, and how it reaches the outcome. Where a number is quoted, it is the number in the source code.

Threat level. A whole number from 1 to 10, set to 6 by default. This represents how severe and how visible the danger is. It works by setting the chance, each tick, that any given unaware person hears an alert for the first time. The formula is the threat level divided by ten, multiplied by 0.35, with a small additional term that rises with elapsed time, and the result held between 2 and 55 per cent. At the default setting of 6 an unaware person has roughly a one in five chance of hearing something in any given tick at the start of a run, rising slowly as the run continues. The important point is that threat level in this model governs how fast news travels, not how much damage is done. Nobody is injured or killed in this simulation.

Threat rise rate. A whole number from 0 to 20, set to 0 by default. At zero the threat level stays fixed for the entire run. Above zero the threat climbs as the run proceeds, adding one level of threat every hundred ticks divided by the setting, and the climb stops at the maximum of 10. This models the compressing decision window that characterises conflict displacement: a household that hesitates finds that the situation has deteriorated around it while it was deciding.

Information clarity. A whole number from 1 to 10, set to 5 by default. This is the single most consequential setting in the model, because it acts in two places at once. First, it sets the reliability of the official broadcast node, which is simply clarity divided by ten, held between 0.1 and 0.95. That reliability then drives the per-tick chance that a seeking person receives a confirmation. Second, and more sharply, if clarity is below 4 then every person in the community is given one or two extra confirmations to obtain before they will move. Poor information is therefore doubly punishing: each confirmation is slower to arrive, and more of them are needed. Populations that have been given vague, contradictory, or previously false warnings become more sceptical, and their scepticism is rational.

Humanitarian access. A percentage from 0 to 100, set to 40 by default. This represents how far an aid organisation has been permitted to operate. At the moment the simulation is built, each individual is independently assigned as reachable or not reachable by the humanitarian node, with the probability equal to the access setting. Someone who is not reachable never receives anything from that channel, no matter how long the run lasts. Those who are reachable receive confirmations from it at a fixed chance of 37.5 per cent per tick. Setting this to zero removes the humanitarian actor entirely and is the simulation's model of denied access.

The four armed conflict settings are all set to zero by default, meaning the simulation begins in a relatively benign state which the user then degrades deliberately.

Checkpoint delay. A whole number of ticks from 0 to 15. When above zero, each person has a 55 per cent chance of being stopped at the moment they finish preparing and begin to move, with a random delay added to their travel time. This represents document checks and security screening. The design choice worth noting is that the delay is applied to everyone equally, which means its humanitarian effect is not equal: a household that was already close to the edge of the corridor window is pushed over it, while a fast household absorbs the same delay without consequence.

Misinformation. A percentage from 0 to 100. When above zero, a person in the Seeking stage has a per-tick chance equal to the setting multiplied by 0.35 of receiving a false confirmation, so a setting of 60 per cent produces a 21 per cent chance per tick. The false confirmation counts toward the threshold exactly as a genuine one does, which is the point: the household cannot tell the difference. The consequence appears later. If the confirmation that finally tipped a person into preparing came from the misinformation channel, then when that person begins to move they are routed not to their nearest open gate but to a randomly chosen different open gate. If only one gate is open there is no wrong way to send them, and the misdirection has no effect, which is a small but real artefact of the implementation.

Infrastructure damage. A whole number from 0 to 20, modelling the destruction of telecommunications. When above zero, the effective information clarity falls every tick, reduced by the elapsed ticks multiplied by the setting and divided by fifteen, with a floor of 1. Both the broadcast node's reliability and the per-tick confirmation chance fall with it. The households punished by this setting are precisely the slowest ones, because they are still seeking confirmation at the point when the information environment collapses.

Two observations about that setting. First, it is far more aggressive than the guide describes. The built-in guide states that at a setting of 10 a starting clarity of 7 will have fallen to around 4 by tick 45. The formula in the source code does not produce that result. A setting of 10 reduces clarity by two thirds of a point every tick, so a starting clarity of 7 reaches the floor of 1 in around nine ticks, not forty-five. Even the setting labelled light damage in the interface collapses clarity from 7 to the floor within about thirteen ticks. Users should treat any non-zero value here as modelling a rapid communications blackout rather than a gradual decline. Second, the extra confirmations imposed by low clarity are assigned once when the run is built and are not increased retrospectively as clarity degrades, so degradation makes each confirmation harder to obtain but does not raise the threshold that must be reached.

Coercion risk. A percentage from 0 to 100, representing forced displacement. When above zero, an unaware person who has not received any alert in a given tick has a chance of being pushed directly into the Milling stage without ever seeking or receiving confirmation. Readers should note how small this is in practice: even at maximum threat and maximum coercion the per-tick chance is 0.8 per cent, well below the cap of 6 per cent, so coercion accumulates slowly over a long run rather than sweeping through the community. Coerced individuals are flagged, drawn with a red ring, logged individually, and counted separately in the end-of-run summary. They may well reach safety faster than their neighbours, because they skipped the entire confirmation process, and the tool is right to insist that this speed is not a good outcome. They left without preparation, without verifying the route, and without choosing to go.

Average family size. A whole number from 1 to 7, set to 3 by default. The actual size of each household is drawn at random between one below and one above this figure, so households vary. Larger households are slower for a structural reason: the hub adopts the longest preparation time and the longest travel time of any member, so every additional person is another chance of drawing a slow one.

Elder ratio. A percentage from 0 to 60, set to 20 by default. Each non-hub member of each household is independently tested against this percentage to determine whether they are an elder. An elder receives one additional required confirmation, extra preparation time, and a reduced movement speed. The extra confirmation is the mechanically interesting part. It encodes the finding that older people are more likely to want a second or third source before accepting a warning, particularly where there is attachment to a home and scepticism toward the announcing authority.

Children under five. A percentage from 0 to 50, set to 20 by default, applied to those not already assigned as elders. Young children receive the longest preparation penalty of the two original categories, reflecting the work of gathering and packing for a small child, and the slowest movement speed of any category on foot. They do not receive extra required confirmations, because a small child is not the one deciding.

Pregnant women. A percentage from 0 to 30, set to 0 by default. Preparation delay is comparable to an elder and movement speed sits between an elder and an adult. No extra confirmations are required. This category was added specifically to represent a group that International Humanitarian Law singles out for protection.

Unaccompanied minors. A percentage from 0 to 20, set to 0 by default. This is the most severely penalised category in the model and deliberately so. An unaccompanied minor requires two additional confirmations, more than any other category, and receives by far the largest preparation delay: six to twelve extra ticks on foot, against two to five for an elder. They then move at the slow speed of a child. A separated child cannot simply leave. Somebody must be found who is authorised to take responsibility for them, and family tracing and reunification take time that no other category incurs. In a run with a timed corridor window, this is the category that gets left behind, and demonstrating that is the point of including it.

One important structural detail. These four categories are assigned in strict priority order and are mutually exclusive: a person is tested for elder first, then child, then pregnant, then unaccompanied minor, and once assigned cannot also be something else. A pregnant elder or a child who is also unaccompanied cannot exist in this model. Household hubs are never assigned any vulnerability category at all. The practical effect is that raising the elder percentage quietly suppresses the number of people available to be assigned to the later categories, so the sliders are not independent of one another.

Neighbour influence. A percentage from 0 to 100, set to 55 by default. Each tick, if any of a household's four linked neighbours has a member visibly milling or evacuating, every seeking member of that household has a chance equal to this setting of gaining a confirmation from that observation alone. At high settings the community behaves as a cascade, where one household departing pulls the rest out behind it. At zero, social observation counts for nothing and only official and humanitarian channels matter.

Three quantities are not exposed as settings but are generated for each individual when the run is built, and they are what the settings above actually act upon. Confirmations needed is a random whole number from 1 to 3, plus one or two more if information clarity is below 4, plus a category penalty. It is the single number that most determines who leaves early and who leaves late. Preparation time is drawn at random from a range that depends on the scenario, with an additional draw added for the person's vulnerability category. Travel time is drawn the same way, and speed is a fixed value per category per scenario, measured in screen pixels per tick.

CategoryExtra confirmationsExtra preparationSpeed on foot
AdultNoneNone2.6
Elder12 to 5 ticks1.8
Child under fiveNone3 to 6 ticks1.5
Pregnant womanNone2 to 5 ticks2.0
Unaccompanied minor26 to 12 ticks1.5
How each category is penalised in the pedestrian scenario. The base preparation range is 2 to 4 ticks for everyone, and the category delay is added on top.

The scenario setting. The user chooses one of three modes of evacuation, and this choice rewrites all the preparation and travel figures at once. The three modes are not simply faster or slower versions of each other. They differ in shape.

ModeBase preparationTravel speed
Pedestrian2 to 4 ticks1.5 for a child up to 2.6 for an adult
Car1 to 3 ticks5.5 for every category
Train4 to 8 ticks8.0 for every category

Pedestrian is the slowest and the most unequal, and vulnerability matters more there than in any other mode. Car equalises mobility almost completely, and this is the model's clearest single finding: access to transport does not merely make evacuation faster, it makes vulnerability stop mattering. Train is the reverse profile, where the long preparation represents the wait for a departure but the journey itself is the fastest in the model. The train scenario has no corridor gates at all; passengers leave by the rail line.

06

Corridors: How People Actually Get Out

In the pedestrian and car scenarios, escape is not in all directions. Four gates are placed at the midpoints of the four edges of the map and named North, South, East, and West. When a person finishes preparing, they identify the nearest gate that is currently open and move toward it. The train scenario has no gates.

Each gate has three settings.

  • It can be open or closed at the start.
  • It can be given a closing tick, at which it shuts.
  • It can be given an opening tick, at which it opens, in which case it begins the run closed.

Combining an opening tick with a closing tick produces a humanitarian window: a gate that opens at tick 15 and closes at tick 35 gives the population a twenty-tick passage. This is the mechanic that makes every preparation delay described above consequential rather than merely descriptive. An eight-tick elder delay is invisible in a run with no time limit and decisive in a run with a twenty-tick window.

When a gate closes mid-run, anyone already travelling toward it recalculates and heads for the nearest remaining open gate, and the reroute is logged individually. Rerouting costs time, and the cost is much higher on foot than by car, which is the asymmetry the documentation identifies as a research finding in its own right. The same closure that mildly inconveniences a household with a vehicle can leave a household on foot unable to reach any alternative.

If no gate is open when a person is ready to move, they are marked as trapped. Trapped individuals stop, are logged by name, and are counted in the final summary. A run ends when every person is either evacuated or trapped. This is the simulation's representation of encirclement and siege, and it is the condition the tool is most clearly designed to make visible: the visual and numerical gap between the households that got out and the households that were still preparing when the gate shut.

07

Reading the Results

While a run is in progress the map is the primary output. Individuals change colour as they pass through the five stages, from grey when unaware, through blue when seeking, amber when milling, red when evacuating, to green when evacuated. Each population category has its own shape, so an elder, a child, a pregnant woman, and an unaccompanied minor can be distinguished at a glance. Lines drawn between the information nodes and individuals show confirmations arriving, colour-coded by which channel supplied them, with the hostile misinformation channel drawn in red.

Below the map, three rows of figures update every tick: how many people are in each of the five stages, how many belong to each vulnerability category, and a single prominent percentage showing how much of the population has completed evacuation.

An event log records everything in plain sentences as it happens: who received an alert, how many confirmations they now hold, who was held at a checkpoint and for how long, which corridor closed and who is rerouting, who was coerced, and who was trapped. The log is capped at the most recent eighty entries. It is the most useful output for teaching, because it converts the statistical outcome back into individual stories.

When a run ends, a summary panel reports the total number of ticks, the average time the population spent in each of the seeking, milling, and evacuating stages, which household finished last, and what the model believes caused that household to be slowest, expressed as its composition. It also reports the channel split, along with counts of those coerced, those held at checkpoints, and those trapped.

The channel split is the analytically richest output. It answers a question that matters operationally: was this evacuation driven by the authorities or by the community? A run in which social confirmation dominates is a run in which official communication failed and the population compensated for it. A run in which the humanitarian channel dominates is a run in which access mattered more than broadcasting. A run in which misinformation dominates is a run in which the population moved decisively in the wrong direction.

The last five runs are retained, and one can be pinned for direct comparison. This is the intended way to use the tool. A single run tells the user very little, because so much is decided by random draws. The signal appears in the difference between two runs that were identical except for the one thing the user changed.

08

Methodological Choices

Why time is abstract. The decision to measure time in ticks rather than minutes is the most defensible choice in the project. The preparation and travel figures were selected to produce plausible relative behaviour, and there is no empirical calibration behind them. Converting them to minutes would manufacture a precision that does not exist. The documentation says so directly and points readers toward calibrated evacuation timing studies for anyone who needs real durations.

Why vulnerability is expressed as delay rather than as a risk score. The model never assigns anyone a number representing how much danger they are in. Instead it gives them extra ticks. This is a meaningful choice. It means vulnerability only produces a bad outcome when it interacts with a constraint, such as a closing corridor or a degrading information environment. In a run with no time pressure, a household full of elders arrives late and arrives safely. That is a more honest representation of how vulnerability actually operates than a standing risk score would be.

Why the household waits. The hub takes the maximum preparation and travel time of anyone in the household, so the whole family moves at the pace of its slowest member. This directly implements Drabek's finding and it is the mechanism by which one unaccompanied minor or one elder slows an entire household rather than only themselves.

Why everything is probabilistic. Almost every transition in the model is decided by a random draw. Two runs with identical settings will produce different results. This is deliberate and correct, since it reflects genuine uncertainty in warning response, but it has a consequence for how the tool should be used. Conclusions should never be drawn from a single run.

Why the population is small. Six households and typically fifteen to twenty individuals is a very small population. This keeps every individual visible and traceable on screen, which serves the teaching purpose, but it means the results are statistically noisy and no percentage produced by a single run should be treated as an estimate of anything.

09

Grounding in International Humanitarian Law

The repository connects each of its mechanics to a provision of International Humanitarian Law, the body of law that regulates the conduct of hostilities and protects those not taking part in the fighting. This is a valuable design feature: it turns legal rules into observable behaviour.

The following provisions are cited in the repository and are cited correctly.

  • Additional Protocol II, Article 17(1), which provides that the displacement of the civilian population shall not be ordered for reasons related to the conflict unless the security of the civilians involved or imperative military reasons so demand, and that where displacement does occur all possible measures shall be taken so that the population is received under satisfactory conditions. Article 17(2) provides that civilians shall not be compelled to leave their own territory. This is the correct basis for the coercion mechanic.
  • Customary International Humanitarian Law Rule 129, which prohibits parties from deporting or forcibly transferring the civilian population of an occupied territory, and from ordering the displacement of the civilian population in non-international armed conflict, subject to the same two exceptions. Rule 131 requires that displaced civilians be received in satisfactory conditions and that families not be separated.
  • Additional Protocol I, Article 37, which prohibits killing, injuring, or capturing an adversary by perfidy, and Article 38, which prohibits improper use of the red cross and red crescent emblems, the flag of truce, and the emblem of the United Nations. These are the bases the repository gives for the misinformation mechanic, subject to the framing point noted at the end of this section.
  • Additional Protocol I, Article 52, which provides that civilian objects shall not be the object of attack, that attacks shall be limited strictly to military objectives, and that where there is doubt whether a normally civilian object is being used militarily it shall be presumed not to be. This is the correct basis for the infrastructure damage mechanic.
  • Additional Protocol I, Article 58, which requires parties to endeavour to remove civilians from the vicinity of military objectives and to take other necessary precautions to protect civilians under their control. Its opening words expressly preserve Article 49 of the Fourth Geneva Convention, which matters: the duty to move civilians away from danger cannot be used to justify displacement that Article 49 forbids.
  • Customary Rule 53 and Additional Protocol I, Article 54, which prohibit starvation of the civilian population as a method of warfare. Article 54(2) is particularly apposite to the trapped mechanic, because it prohibits destroying or rendering useless objects indispensable to civilian survival whatever the motive, expressly including the motive of causing civilians to move away.
  • Additional Protocol I, Article 70 and Customary Rule 55, which require parties to allow and facilitate rapid and unimpeded passage of impartial humanitarian relief for civilians in need, subject to a right of control. This is the correct basis for the humanitarian access setting.
  • Additional Protocol I, Article 57(2)(c) and Customary Rule 20, which require each party to give effective advance warning of attacks which may affect the civilian population unless circumstances do not permit. A high-threat, low-clarity run is a fair representation of a warning that was issued but was too vague to act on.
  • Customary Rule 138, which provides that the elderly, disabled, and infirm affected by armed conflict are entitled to special respect and protection. This supports the elder mechanic.

Four citations in the repository documentation are inaccurate and should be corrected. This is not a criticism of the project's substance, which is sound. It is a matter of the accuracy that a teaching tool on legal subjects requires.

Customary Rule 99. It is cited repeatedly, in the corridor documentation, the checkpoint documentation, and the built-in guide, as the rule requiring parties to allow civilians to leave and prohibiting disproportionate impediment to civilian movement. Rule 99 does not say this. Rule 99 provides that arbitrary deprivation of liberty is prohibited. It is a detention rule, situated in the chapter of the ICRC Customary Law Study on fundamental guarantees. There is in fact no single customary rule creating a general civilian right to leave a besieged area. The nearest provision is Article 17 of the Fourth Geneva Convention, which requires parties to endeavour to conclude local agreements for the removal from besieged or encircled areas of the wounded, sick, infirm, and aged, of children, and of maternity cases. Two limits of that article are instructive for this project: it is only an obligation to endeavour, and it covers specific vulnerable categories rather than the whole civilian population.

Article 16 of Additional Protocol I. It is cited in the population settings as the provision giving special protection to pregnant women. It is not. Article 16 of Additional Protocol I is titled General protection of medical duties and concerns the protection of medical personnel from being punished or compelled to act against medical ethics. The intended citation is almost certainly Article 16 of the Fourth Geneva Convention, which does provide that the wounded and sick, the infirm, and expectant mothers shall be the object of particular protection and respect. The stronger provision for this purpose is Article 8(a) of Additional Protocol I, which brings maternity cases, newborn babies, and expectant mothers within the definition of the wounded and sick and therefore within that entire protective regime.

Article 78 of Additional Protocol I. It is cited as the basis for the unaccompanied minors mechanic. The article is real and its content is relevant, but its scope is narrower than the repository implies. Article 78 governs the evacuation of children who are not a party's own nationals to a foreign country, and requires that such evacuation be temporary, justified by compelling reasons, consented to in writing by parents or legal guardians, and supervised by a Protecting Power. It does not regulate the internal evacuation of a state's own children, which is what the simulation depicts. For the protection of children generally, Article 77 of Additional Protocol I and Articles 24 and 50 of the Fourth Geneva Convention are the appropriate provisions. The reunification delay the model applies is nonetheless well founded: it reflects the tracing and reunification work that Article 26 of the Fourth Geneva Convention and Article 74 of Additional Protocol I require states to facilitate.

Customary Rule 9. It is cited alongside Article 52 of Additional Protocol I for the prohibition on attacking civilian objects. Rule 9 is a definition rather than a prohibition: it states that civilian objects are all objects that are not military objectives. The protective proposition is Rule 10, which provides that civilian objects are protected against attack unless and for such time as they are military objectives. Citing Rules 9 and 10 together would be correct.

One further point is a matter of framing rather than a citation error. The repository presents misinformation about evacuation routes as falling squarely within the prohibition on perfidy in Article 37. The connection is intuitive but the article is narrower than that: perfidy under Article 37 requires the intent to kill, injure, or capture an adversary by inviting their confidence in a protection under international law, and Article 37(2) expressly permits ruses of war and names misinformation among them. False information directed at a civilian population to endanger it engages other rules, including the general protection of civilians and the prohibition on acts whose primary purpose is to spread terror, more directly than it engages Article 37. The mechanic is valuable and the concern is real. The legal framing would benefit from revision.

10

Limitations and Caveats

Nobody is harmed. The model has no concept of injury or death. Being trapped is the worst outcome available, and being trapped is recorded as a status rather than as a consequence. Corridor violations, attacks on convoys, and secondary threats during transit are all identified in the repository's own planning documents as desirable additions and none are implemented.

The timings are not calibrated. This bears repeating because it constrains every quantitative statement the tool makes. The preparation ranges, movement speeds, and confirmation thresholds are plausible relative values chosen by the author. They are not derived from evacuation data.

The population is tiny and the results are noisy. With six households, a single unlucky random draw can change a headline figure substantially. No number from a single run should be quoted.

The vulnerability categories cannot overlap. Because a person is tested for each category in a fixed order and assigned to the first that matches, the model cannot represent a pregnant woman with a disability, an elderly person who is also chronically ill, or a child who is both under five and unaccompanied. Compounding vulnerability, which is exactly what determines outcomes in real displacement, is outside the model.

Several important categories are absent. The wounded and sick, persons with disabilities, and people lacking identity documents are all discussed at length in the repository's extensions document and none are implemented. The first of these is the most significant omission, since the wounded and sick hold the strongest claim to priority evacuation in the entire legal framework.

The threat is spatially uniform. Everyone on the map faces the same threat level regardless of where they stand. There is no front line, no direction from which danger approaches, and no reason why a household near the fighting should be alerted before one further away. The repository identifies directional threat as an important missing feature and it remains missing.

Coercion is rare in practice. Because of the very small coefficient in the formula, even the maximum coercion setting produces a low per-tick probability. Users who set the slider to a high value and observe few coercion events are not misreading the display.

Misinformation has no effect when only one gate is open. The mechanic works by misrouting people to a non-nearest gate, so it requires at least two open gates to do anything at all beyond arriving faster. In a single-corridor scenario, misinformation makes the population move sooner and no worse.

There is no test suite. The repository contains no automated tests, so the behaviour of the model rests on visual inspection.

Above all, this is a teaching model. It is designed to make a small number of relationships vivid, not to forecast what a real population will do. Its outputs should never inform an operational decision.

11

Practical Nature and Intended Audience

The tool is a web page. It runs in an ordinary browser with nothing to install and is published as a public demonstration. It opens on a guide page rather than on the simulation itself, which is a deliberate and correct choice: the guide walks the reader through the five stages, the information channels, the population categories, the corridor system, the armed conflict mechanics, and the legal framework before they touch a single control. The guide is longer than the simulation code that it explains.

The intended users are students and humanitarian practitioners. The tool suits a classroom particularly well, because the event log converts every statistical outcome back into a named person with a household and a reason for their delay. A discussion of why the Okafor family was still milling when the North corridor closed is a more productive discussion than a discussion of a percentage.

The most instructive exercise the tool supports is comparative. Run the same community on foot and by car with an identical corridor window and identical demographics, and the difference in who gets out is entirely attributable to access to transport. That single comparison carries the project's central argument more effectively than any of its individual settings.

12

Conclusion

EvacSim makes one argument well. Evacuation outcomes in armed conflict are determined less by the severity of the danger than by the interaction between how quickly information becomes credible and how long particular people take to act on it. The tool holds those two quantities apart and lets a user vary each independently, and its most valuable outputs are the ones that show where they collide: the elder still preparing when the corridor closes, the unaccompanied child awaiting an escort that arrives after the window has passed.

Its treatment of vulnerability deserves particular note. By expressing protected status as time rather than as a risk score, the model produces a result that is easy to state and hard to argue with. Vulnerability is harmless until it meets a constraint, and every constraint that International Humanitarian Law regulates, the closed corridor, the checkpoint, the destroyed transmitter, the false direction, works by shortening the time available to the people who need the most of it.

The prototype is modest about its status and should be. Its timings are uncalibrated, its population is small, nobody in it can be hurt, and its vulnerability categories cannot overlap in the ways that matter most in real displacement. Its legal citations require correction before it is used to teach law. What it offers is not prediction but a way of seeing, and for that purpose the choice to keep every simulated person individually visible, named, and traceable through the event log is worth more than any additional mechanic would be.

References

Sources

  1. 01Quarantelli, E.L. Disaster Research Center, University of Delaware. Work establishing that people rarely panic in disasters and instead seek confirmation from multiple sources before accepting that a threat is real.
  2. 02Drabek, T.E. University of Denver. Work documenting that families evacuate as units rather than as individuals, waiting until all members are present before departing.
  3. 03International Committee of the Red Cross. Article 17, Protocol Additional to the Geneva Conventions of 12 August 1949, and relating to the Protection of Victims of Non-International Armed Conflicts (Protocol II), 1977. Prohibition of forced movement of civilians.
  4. 04International Committee of the Red Cross. Article 17, Geneva Convention (IV) relative to the Protection of Civilian Persons in Time of War, 1949. Removal from besieged or encircled areas.
  5. 05International Committee of the Red Cross. Article 16, Geneva Convention (IV), 1949. Particular protection and respect for the wounded and sick, the infirm, and expectant mothers.
  6. 06International Committee of the Red Cross. Article 26, Geneva Convention (IV), 1949. Facilitating enquiries by dispersed families.
  7. 07International Committee of the Red Cross. Article 37, Protocol Additional to the Geneva Conventions of 12 August 1949, and relating to the Protection of Victims of International Armed Conflicts (Protocol I), 1977. Prohibition of perfidy, and permitted ruses of war.
  8. 08International Committee of the Red Cross. Article 52, Additional Protocol I, 1977. General protection of civilian objects.
  9. 09International Committee of the Red Cross. Article 54, Additional Protocol I, 1977. Protection of objects indispensable to the survival of the civilian population.
  10. 10International Committee of the Red Cross. Article 57, Additional Protocol I, 1977. Precautions in attack, including effective advance warning.
  11. 11International Committee of the Red Cross. Article 58, Additional Protocol I, 1977. Precautions against the effects of attacks.
  12. 12International Committee of the Red Cross. Article 70, Additional Protocol I, 1977. Relief actions, and priority in the distribution of relief.
  13. 13International Committee of the Red Cross. Article 77, Additional Protocol I, 1977. Protection of children.
  14. 14International Committee of the Red Cross. Article 78, Additional Protocol I, 1977. Evacuation of children to a foreign country.
  15. 15International Committee of the Red Cross. Customary International Humanitarian Law Study, Rule 9. Definition of civilian objects.
  16. 16International Committee of the Red Cross. Customary International Humanitarian Law Study, Rule 10. Civilian objects lose their protection only when they are military objectives.
  17. 17International Committee of the Red Cross. Customary International Humanitarian Law Study, Rule 20. Advance warning.
  18. 18International Committee of the Red Cross. Customary International Humanitarian Law Study, Rule 53. Starvation as a method of warfare.
  19. 19International Committee of the Red Cross. Customary International Humanitarian Law Study, Rule 55. Access for humanitarian relief to civilians in need.
  20. 20International Committee of the Red Cross. Customary International Humanitarian Law Study, Rule 99. Deprivation of liberty.
  21. 21International Committee of the Red Cross. Customary International Humanitarian Law Study, Rule 129. The act of displacement.
  22. 22International Committee of the Red Cross. Customary International Humanitarian Law Study, Rule 131. Treatment of displaced persons.
  23. 23International Committee of the Red Cross. Customary International Humanitarian Law Study, Rule 138. The elderly, disabled and infirm.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file diff --git a/static-site/print/evacuation-simulation/index.txt b/static-site/print/evacuation-simulation/index.txt index 1959d59b8..8ef4b2eef 100644 --- a/static-site/print/evacuation-simulation/index.txt +++ b/static-site/print/evacuation-simulation/index.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -12:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","print","evacuation-simulation",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","evacuation-simulation","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -14:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -16:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","print","evacuation-simulation",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","evacuation-simulation","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:"$Sreact.suspense" -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1c:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -29,7 +29,7 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 19:[] 10:"$W19" 11:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1e:Ta14, @page { size: 210mm 297mm; margin: 20mm 18mm; } @@ -171,6 +171,6 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 57:["$","li","21",{"children":[["$","span",null,{"className":"print-ref-num","children":"22"}],["$","span",null,{"children":"International Committee of the Red Cross. Customary International Humanitarian Law Study, Rule 131. Treatment of displaced persons."}]]}] 58:["$","li","22",{"children":[["$","span",null,{"className":"print-ref-num","children":"23"}],["$","span",null,{"children":"International Committee of the Red Cross. Customary International Humanitarian Law Study, Rule 138. The elderly, disabled and infirm."}]]}] 1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -59:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +59:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 18:null 1d:[["$","title","0",{"children":"The Evacuation Simulator (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L59","6",{}]] diff --git a/static-site/print/forced-labor-structural-risk-index/__next._full.txt b/static-site/print/forced-labor-structural-risk-index/__next._full.txt index 6873207ad..a97993870 100644 --- a/static-site/print/forced-labor-structural-risk-index/__next._full.txt +++ b/static-site/print/forced-labor-structural-risk-index/__next._full.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -12:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","print","forced-labor-structural-risk-index",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","forced-labor-structural-risk-index","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -14:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -16:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","print","forced-labor-structural-risk-index",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","forced-labor-structural-risk-index","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:"$Sreact.suspense" -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1c:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -29,7 +29,7 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 19:[] 10:"$W19" 11:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1e:Ta14, @page { size: 210mm 297mm; margin: 20mm 18mm; } @@ -180,6 +180,6 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 61:["$","li","23",{"children":[["$","span",null,{"className":"print-ref-num","children":"24"}],["$","span",null,{"children":"IPUMS-International. Harmonised census microdata behind the sub-national layer, with the Geocoded Disasters dataset for sub-national disaster exposure."}]]}] 62:["$","li","24",{"children":[["$","span",null,{"className":"print-ref-num","children":"25"}],["$","span",null,{"children":"Walk Free. Global Slavery Index, used as the external prevalence benchmark and for vulnerability dimension alignment."}]]}] 1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -67:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +67:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 18:null 1d:[["$","title","0",{"children":"The Forced Labor Structural Risk Index (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L67","6",{}]] diff --git a/static-site/print/forced-labor-structural-risk-index/__next._head.txt b/static-site/print/forced-labor-structural-risk-index/__next._head.txt index 07a45c572..c242faad0 100644 --- a/static-site/print/forced-labor-structural-risk-index/__next._head.txt +++ b/static-site/print/forced-labor-structural-risk-index/__next._head.txt @@ -1,6 +1,6 @@ 1:"$Sreact.fragment" -2:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -3:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +2:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +3:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 4:"$Sreact.suspense" -5:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +5:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 0:{"rsc":["$","$1","h",{"children":[null,["$","$L2",null,{"children":[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]]}],["$","div",null,{"hidden":true,"children":["$","$L3",null,{"children":["$","$4",null,{"name":"Next.Metadata","children":[["$","title","0",{"children":"The Forced Labor Structural Risk Index (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L5","6",{}]]}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/print/forced-labor-structural-risk-index/__next._index.txt b/static-site/print/forced-labor-structural-risk-index/__next._index.txt index f98a4fd90..18b9d82b6 100644 --- a/static-site/print/forced-labor-structural-risk-index/__next._index.txt +++ b/static-site/print/forced-labor-structural-risk-index/__next._index.txt @@ -1,13 +1,13 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -d:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +d:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} 8:["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$Ld",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$Ld",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$Ld",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$Ld",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] diff --git a/static-site/print/forced-labor-structural-risk-index/__next._tree.txt b/static-site/print/forced-labor-structural-risk-index/__next._tree.txt index e4bfb1877..b1ed02605 100644 --- a/static-site/print/forced-labor-structural-risk-index/__next._tree.txt +++ b/static-site/print/forced-labor-structural-risk-index/__next._tree.txt @@ -1,7 +1,7 @@ -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] 0:{"tree":{"name":"","param":null,"prefetchHints":16,"slots":{"children":{"name":"print","param":null,"prefetchHints":0,"slots":{"children":{"name":"slug","param":{"type":"d","key":"forced-labor-structural-risk-index","siblings":null},"prefetchHints":0,"slots":{"children":{"name":"__PAGE__","param":null,"prefetchHints":0,"slots":null}}}}}}},"staleTime":300,"buildId":"static-snapshot"} diff --git a/static-site/print/forced-labor-structural-risk-index/__next.print.txt b/static-site/print/forced-labor-structural-risk-index/__next.print.txt index 460afe242..fa617e997 100644 --- a/static-site/print/forced-labor-structural-risk-index/__next.print.txt +++ b/static-site/print/forced-labor-structural-risk-index/__next.print.txt @@ -1,4 +1,4 @@ 1:"$Sreact.fragment" -2:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -3:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +2:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +3:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] 0:{"rsc":["$","$1","c",{"children":[null,[["$","style",null,{"children":"\n /* The site is dark-first, and a dark color-scheme makes Chrome paint\n the printed sheet's margin area dark too. Paper is light. */\n html { color-scheme: light !important; }\n html, body { background: #ffffff !important; }\n body::after { content: none !important; }\n header, footer, .site-bg, .aura { display: none !important; }\n "}],["$","$L2",null,{"parallelRouterKey":"children","template":["$","$L3",null,{}]}]]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/print/forced-labor-structural-risk-index/index.html b/static-site/print/forced-labor-structural-risk-index/index.html index 2d582d18f..a831fb51e 100644 --- a/static-site/print/forced-labor-structural-risk-index/index.html +++ b/static-site/print/forced-labor-structural-risk-index/index.html @@ -1,4 +1,4 @@ -The Forced Labor Structural Risk Index (print edition) · NYU Ethical Tech CoLab

Publications · Academic report

The Forced Labor Structural Risk Index

A Country-Level Measure of the Conditions Under Which Forced Labour Becomes More Likely

Ethical Tech CoLab

NYU Center for Global Affairs

December 2025

Prepared as masters research at the NYU Center for Global Affairs.

Forced labour is one of the few large-scale human rights violations that almost nobody can count, and the countries where the risk is gravest are frequently the countries where the evidence is thinnest, because the same institutional weakness that permits exploitation also prevents it from being recorded. FLSRI does not try to estimate how many people are in forced labour. It measures something different and, for prevention purposes, more tractable: how strongly the structural conditions that make forced labour more likely are present, and how those conditions combine.

Key figures

184

countries of roughly 195 receive a composite score; the rest are reported as not scored rather than given an invented value

43

standardised indicators across eleven domains, each drawn from a named cross-national dataset

10,000

simulated re-scorings behind the uncertainty band published for every scored country

45

ranks wide is the median uncertainty band, which is why the index asks to be read in tiers rather than as a league table

01

Executive Summary

FLSRI is a research prototype: a country-level index that scores the structural conditions under which forced labour becomes more likely. It scores 184 of roughly 195 countries on a scale running from 0 to 1, where 0 is the lowest modelled structural risk and 1 the highest. The remaining countries are reported as not scored rather than being given an invented value.

The index is deliberately not a prevalence estimate. It does not say how many people are exploited. A high score means the enabling conditions hold together strongly; a low score does not certify a country as free of forced labour. This distinction is stated repeatedly in the project's own documentation and it governs every legitimate use of the results.

The framework rests on a simple structural claim borrowed from criminology: exploitation becomes likely when a population exposed to coercive recruitment coincides with an environment in which exploiting that population goes unchecked. The index therefore has two scored halves. Recruitment, written as R, asks who is structurally exposed, through poverty, debt, blocked mobility, exclusion, weak legal identity, gendered labour structures, childhood exposure, and shocks such as conflict and disaster. Exploitation, written as E, asks whether exploiting them can run without consequence, through blocked exit, demand in high-risk sectors, and the state's own production of unfreedom.

The two halves are combined by a geometric mean, that is, the square root of R multiplied by E. This is a deliberate methodological choice. It means a country can only score high when both halves are high. A very exposed population in a country with strong labour enforcement, or a permissive enforcement environment in a country with little structural vulnerability, is pulled down rather than averaged out.

A third phase, Monetization, covering the financial conditions under which the proceeds of forced labour can be moved, hidden, and kept, is computed and published but deliberately excluded from the headline score. It answers a different question, namely where intervention against the money would bite, and it scores high for wealthy financially opaque economies in a way that would distort a structural risk reading.

Eleven domains and forty-three standardised indicators feed the published score, each drawn from an established cross-national dataset such as the World Bank World Development Indicators, ILOSTAT, the V-Dem democracy dataset, the UNHCR population statistics, the UCDP conflict event dataset, and the EM-DAT disaster database.

The project is unusually disciplined about two failure modes that afflict composite indices of this kind. The first is circularity, that is, predicting forced labour using a measure of forced labour. The second is the risk that the index becomes a governance ranking wearing a new label. Both are addressed explicitly, tested, and reported rather than concealed.

Every scored country carries a published uncertainty band derived from 10,000 simulated re-scorings, and the documentation insists that the results be read in broad tiers rather than as an exact league table. The middle of the table is openly described as unstable. The index is delivered as an interactive public website together with the complete pipeline that produces it, so that any figure shown can be traced back to the data and the code that generated it.

02

Background and Rationale

The problem. Forced labour is defined in international law and prohibited almost universally, yet it is measured very unevenly. Global estimates exist, but they are produced from surveys and administrative sources that are not available for every country, are refreshed slowly, and were never designed to support a fine-grained country-by-country comparison. Detection statistics, such as counts of identified trafficking victims, are worse still for this purpose: they largely record how much capacity a state has to detect and register cases, so a well-resourced state can appear worse than a state where nothing is investigated at all.

The gap. Prevention and triage need to know where conditions are dangerous before cases surface. Existing vulnerability measures go some way toward this, but several of them are built partly from the same governance material that dominates most cross-national comparison, which makes it hard to tell whether they are measuring forced-labour risk or simply restating institutional weakness.

The response. FLSRI proposes a purpose-built structural measure. Its guiding commitments, visible throughout the code and documentation, are that every figure should be traceable to a named public dataset, that missing information should be declared rather than filled in, that the reasoning behind each methodological choice should be written down, and that the limits of the result should be published as prominently as the result itself.

The project describes itself as publishing a framework, a pipeline, and code, with the country scores presented as estimates of structural conditions carrying explicit uncertainty. It does not present the country rankings as findings about the world.

03

Objectives

The index is designed to do the following.

  • Measure the structural conditions associated with forced labour across almost all countries, on a single comparable scale, without estimating prevalence.
  • Make the underlying theory explicit, by separating exposure to recruitment from the conditions for unchecked exploitation and requiring both to be present before a country scores high.
  • Ground every indicator in a real, citable, cross-national dataset, with the source, vintage, coverage, licence, and required citation recorded for each one.
  • Refuse to fabricate. Missing data is never silently converted to zero, and countries whose evidence base is too thin are left unscored.
  • Guard against circularity by excluding inputs that are themselves measures or detections of forced labour and trafficking, wherever using them would mean predicting forced labour with a measure of forced labour.
  • Report, rather than engineer away, the index's association with weak governance, while demonstrating that the index is not reducible to a governance ranking.
  • Publish uncertainty alongside every score, and direct readers toward tiers rather than exact ranks.
  • Support prevention and triage by identifying where structural risk concentrates, including below the national level, and where in the structure a government, a buyer, or a non-governmental organisation might act.

04

The Legal and Normative Frame

Forced labour has a settled definition in international law. The International Labour Organization's Forced Labour Convention, 1930 (No. 29), Article 2(1), defines forced or compulsory labour as all work or service which is exacted from any person under the menace of any penalty and for which the said person has not offered himself voluntarily. The Abolition of Forced Labour Convention, 1957 (No. 105), prohibits the use of forced labour for specified purposes including political coercion, labour discipline, and discrimination. The 2014 Protocol to Convention No. 29 obliges ratifying states to take effective measures on prevention, on the protection of victims, and on access to remedies including compensation. It entered into force in November 2016.

To help identify forced labour in practice, the ILO publishes a set of eleven indicators of forced labour, first issued in 2012 and reissued in a revised edition in 2025. The eleven are these.

  • Abuse of vulnerability.
  • Deception.
  • Restriction of movement.
  • Isolation.
  • Physical and sexual violence.
  • Intimidation and threats.
  • Retention of identity documents.
  • Withholding of wages.
  • Debt bondage.
  • Abusive working and living conditions.
  • Excessive overtime.

The ILO describes their purpose as assisting law enforcement officials, labour inspectors, trade union officers, and non-governmental organisation staff to identify people who may be trapped in forced labour and may need urgent help.

The ILO is careful about how much weight any one indicator carries. Its guidance states that the presence of a single indicator may in some cases imply the existence of forced labour, while in other cases several indicators must be read together before a case can be said to exist. The eleven are a screening aid for practitioners, not a legal test, and they are frequently over-read as a checklist. That caution matters for an index built on the same vocabulary.

These are operational indicators for recognising a case in front of you. FLSRI works at a different level: it measures the standing country conditions that make those case-level indicators more likely to arise, and its documentation states that its domain set is broadly aligned with these indicators and with the vulnerability dimensions used in the Walk Free Global Slavery Index.

The obligation is not only a state obligation. The United Nations Guiding Principles on Business and Human Rights, endorsed by the UN Human Rights Council in 2011, rest on three pillars: the state duty to protect human rights, the corporate responsibility to respect them, and access to remedy for those harmed. The Guiding Principles on human rights due diligence require enterprises to identify, prevent, mitigate, and account for how they address their human rights impacts, including impacts in their supply chains and business relationships. A structural risk measure is directly relevant to that first step: an enterprise or a public procurement body cannot prioritise due diligence without some defensible view of where risk is concentrated.

Target 8.7 of the Sustainable Development Goals commits states to take immediate and effective measures to eradicate forced labour, end modern slavery and human trafficking, and secure the prohibition and elimination of the worst forms of child labour. Two of the index's own indicators are drawn directly from the SDG monitoring framework: informal employment as a share of total employment, which is SDG indicator 8.3.1, and the completeness of birth registration, which is SDG indicator 16.9.1.

The index does not claim to establish a violation of any of these instruments, and it should not be cited as evidence that a particular state is in breach of its obligations. It is a prioritisation instrument for prevention, not a compliance finding.

05

How the Index Is Built

The index has a fixed shape with four levels: indicator, domain, phase, and composite. Each level sits on the same 0 to 1 scale.

The indicator. An indicator is a single measured quantity taken from a real dataset, for example the share of the population living below the 6.85 dollar a day poverty line, or the number of labour inspectors per ten thousand employed people. Each one arrives in its own units, so the pipeline rescales it onto the common 0 to 1 risk scale.

The domain. A domain is a group of indicators measuring one mechanism, for example economic precarity or constrained mobility. The domain score is the plain average of the indicators that are present for that country. Indicators that are missing are dropped from the average and are never entered as zero.

The phase. A phase is one side of the structural claim. Recruitment contains eight domains; Exploitation contains three. A phase score is the plain average of the domains scored inside it. The eleven are set out below and described one by one in the next section.

PhaseDomainMechanism it is meant to capture
RecruitmentEconomic precarityNeeds that the available work cannot meet
RecruitmentDebt and financialised dependencyObligation that turns a voluntary arrangement into an inescapable one
RecruitmentConstrained mobilityInability to move legally or affordably
RecruitmentAscriptive exclusionExclusion by group membership rather than by conduct
RecruitmentLegal non-recognitionInability to prove who you are and so to claim protection
RecruitmentGendered labourConcentration of women in exploitation-exposed work
RecruitmentAge and childhood structuringDirect exploitability of children and household labour exhaustion
RecruitmentStructural disruptionConflict, displacement, and disaster shocks
ExploitationForeclosed exitThe cost of walking away from a bad situation
ExploitationEconomic structure and demandHow much of the economy sits in high-risk work
ExploitationState production of unfreedomUnfreedom generated by legal architecture and impunity
The eleven scored domains and the phase each belongs to. The Monetization lens is computed separately and excluded from the headline score.

The composite. The published score is the geometric mean of the two phases, that is, the square root of Recruitment multiplied by Exploitation. If either phase cannot be scored for a country, the composite is not scored. There is no substitution of zero for a missing phase.

Indicators are rescaled by what the project calls absolute anchoring. For each indicator, a floor value and a ceiling value are fixed in advance. The floor is the raw value that maps to 0 and the ceiling is the raw value that maps to 1; anything beyond either end is clamped.

IndicatorFloor, mapping to 0Ceiling, mapping to 1
Share of employment in agriculture0 per cent80 per cent
Gender gap in labour force participation0 percentage points50 percentage points
Conflict deaths0 per 100,000 people100 per 100,000 people
Three examples of the anchor ranges the pipeline uses.

Anchoring to fixed reference points rather than to the best and worst countries observed is a deliberate choice. It means a country's score does not move simply because a different set of countries happened to have data this year, and it makes scores comparable across refreshes of the underlying data. Where a quantity depends on the size of a country, it is first converted to a rate, a share, or a gap before scaling, so that a large country is not scored as riskier merely for being large.

Each indicator carries an explicit direction. Most point the same way as risk: more poverty means more risk. Some are protective and are inverted, so that a high raw value becomes a low risk score. Trade union density, collective bargaining coverage, labour inspector density, freedom of movement, and visa-free passport access are all protective indicators entered in inverted form. By the time any indicator reaches the scoring stage, every value points in the same direction, with higher meaning more risk.

The documentation is candid that several anchors were set from the observed distribution, for instance a ceiling near the 95th percentile, which reintroduces a degree of dependence on the particular sample of countries into a scale that is presented as absolute. Locking these anchors and testing how much the results move if they shift is listed as outstanding work.

Every level uses equal weights. Indicators inside a domain are averaged equally, domains inside a phase are averaged equally, and the two phases enter the geometric mean equally. The documentation is unusually direct that this is a value judgment rather than a neutral default. Equal weighting asserts that each component is equally relevant to the construct, which is a substantive claim and not an absence of one.

Equal weighting at each level produces unequal weight per indicator once the whole structure is taken into account, because the boxes are not the same size. A domain containing one indicator gives that indicator the full weight of a domain, while a domain containing five splits it. More consequentially, the Exploitation phase contains three domains against Recruitment's eight, so a single Exploitation domain carries roughly 2.7 times the marginal weight of a single Recruitment domain. The project's own interactive material states this plainly rather than leaving it to be discovered.

The project reports that shifting the balance between the two phases as far as 60 to 40 in either direction barely moves the ranking. It also states that it has not yet published a similar test for perturbing weights at the indicator and domain level, and records this as a known gap.

The governing principle on missing data is that it is never silently treated as zero. A missing indicator is dropped from its domain average rather than counted as an absence of risk. Dropping cannot go on indefinitely, so a coverage floor applies. A domain is scored only if at least half of its mapped indicators are present and never fewer than two. The same floor applies to the count of scored domains within a phase. Below the floor, the domain or phase is marked not scored. A domain that is designed around a single indicator cannot mechanically meet the requirement for two; these are handled as a named exception, scored if the one indicator is present but always carried as low confidence.

Under these rules, 184 of roughly 195 countries receive a composite score. The eleven unscored countries are Andorra, Dominica, the Federated States of Micronesia, Saint Kitts and Nevis, Liechtenstein, Monaco, the Marshall Islands, Nauru, San Marino, Tuvalu, and the Holy See. The documentation notes that this missingness is not random: these are micro-states and small island states for which the labour and governance series are simply not collected. It also warns that within the scored set, thin coverage tends to push a score down rather than up, so a very low score for a data-sparse state deserves extra caution.

06

The Variables, Explained in Plain Terms

This section is the heart of the report. It sets out, domain by domain, what each indicator represents, why it was chosen, and how it enters the score. All indicators are averaged equally within their domain. Phase R, Recruitment, asks who is structurally exposed, and holds eight domains.

Economic precarity. The proposition is that people who cannot meet their needs through the work available to them accept terms they would otherwise refuse. Five indicators carry it. Poverty headcount is the share of the population living on less than 6.85 dollars a day, from the World Bank World Development Indicators, anchored from 0 to 100 per cent. Informal employment share is the proportion of total employment that is informal, taken from ILOSTAT as SDG indicator 8.3.1; informal work sits outside labour inspection, contract enforcement, and social protection, which is precisely the exposure the domain is trying to capture. Agrarian employment share is the proportion of employment in agriculture, anchored from 0 to 80 per cent, because agriculture concentrates seasonal, isolated, and poorly regulated work. Income volatility is the standard deviation of real GDP growth over roughly fifteen years, anchored from 0 to 8 percentage points, on the reasoning that instability and not only low income is what pushes households into distress decisions. Income inequality is the Gini index, anchored from 20 to 65.

Debt and financialised dependency. Debt is the classic mechanism by which a voluntary arrangement becomes an inescapable one, and debt bondage is one of the ILO's eleven case-level indicators. All three measures come from the World Bank Global Findex Database, a large survey of how adults around the world save, borrow, and make payments: the share of adults with no financial account, the share who borrowed from informal sources in the past year, and the share who borrowed any money in the past year. The domain is carried as low confidence, and the reason is stated openly. The mechanism the designers actually wanted, namely recruitment-fee and migration debt, has no cross-national source, so the domain rests on financial exclusion and borrowing patterns instead. Borrowing prevalence in particular measures how widely credit is used, not whether households are in distress. The borrowing questions cover only about 59 per cent of countries.

Constrained mobility. The idea is that people who cannot move legally or affordably are easier to recruit on bad terms and harder to extract from a bad situation. Freedom of movement comes from V-Dem, an expert-coded dataset on the qualities of political regimes, entered inverted. Passport access is the number of destinations reachable without a visa, from the Henley Passport Index, also inverted, on the reasoning that a weak passport means legal migration routes are scarce and irregular ones become the alternative. Refugees originating from the country, per 100,000 people, come from the UNHCR population statistics. A tied-status coding is derived from the DEMSCORE legal-status data, covering 29 countries, alongside a hand-coded kafala tied-status signal covering eight sponsorship-system states.

The kafala signal. Under sponsorship systems used across parts of the Gulf and the wider region, a migrant worker's legal status is tied to a single employer, so leaving that employer can mean losing the right to remain. The index codes three features for the eight adopting states it covers: whether status is tied to one employer, whether the employer's consent is required to leave, and whether leaving is treated as an offence. This is a narrow, hand-built input covering a handful of countries, not a global series, and the project is explicit that the sponsorship mechanism remains unsourced at country scale. The domain is carried as low confidence.

Ascriptive exclusion. Exclusion by group membership, meaning by who a person is rather than what they have done, is captured by a single indicator: the depth-weighted share of the population belonging to politically excluded ethnic groups, from the Ethnic Power Relations dataset maintained at ETH Zurich, which codes ethnic groups' access to executive power. Two limitations are recorded. The indicator captures political exclusion, not exclusion within the labour market, which is the channel most relevant to forced labour. And roughly 41 countries lie outside the dataset's universe and are left missing rather than scored as zero. The domain is single-indicator and therefore always carried as low confidence.

Legal non-recognition. A person who cannot prove who they are cannot easily enforce a contract, claim a wage, access protection, or cross a border lawfully. Birth-registration incompleteness is derived from the completeness of birth registration reported through the World Bank from UNICEF and UN Statistics Division civil registration data, corresponding to SDG indicator 16.9.1, entered inverted. Statelessness prevalence, in persons per 100,000, comes from the UNHCR population statistics. Where statelessness figures are suppressed or absent, the domain rests on birth registration alone; 26 countries are in that position and are flagged as low confidence.

Gendered labour. Forced labour is strongly gender-patterned, particularly in domestic work, care work, and sectors where women are concentrated. Four indicators carry the domain: the gap between male and female labour force participation, in percentage points, anchored from 0 to 50; the UNDP Gender Inequality Index, covering reproductive health, empowerment, and labour market participation, running from 0 for equality to 1 for maximum inequality; legal constraints on women's mobility, taken from the mobility component of the World Bank Women, Business and the Law project, entered inverted; and sex-by-sector channelling from ILOSTAT, the extent to which one sex is concentrated in exploitation-exposed sectors of employment. The documentation is careful to record that a global prevalence estimate is used only to justify why this mechanism matters, and is never scored. The value entered is always a structural employment share, never a victim count.

Age and childhood structuring. Children are both directly exploitable and an indicator that households are already using every available source of labour. Four indicators, all from the World Bank's mirrors of ILO and UNICEF SDG series: child labour prevalence, being the share of children aged 7 to 14 in employment; the out-of-school rate for lower-secondary-age adolescents; the share of the population aged 0 to 14; and the child marriage rate, the share of women aged 20 to 24 first married before 18. This domain carries the project's most awkward data problem. Child labour prevalence, the indicator that most directly measures the mechanism, covers only about 47 per cent of countries, from survey vintages spanning 2005 to 2016. That is below the coverage floor. Consistent with the no-imputation rule, the gap is disclosed rather than filled. The documentation also records that the causal link from child marriage to forced labour is under review rather than settled.

Structural disruption. Shocks displace people, break household economies, and hand recruiters a supply of people with nothing to fall back on. Disaster-affected intensity comes from EM-DAT, maintained by CRED at UCLouvain, measured over the five years 2020 to 2024, with heat-wave entries excluded on the documented ground that heat mortality is recorded only where a country runs attribution studies, so it tracks measurement capacity rather than shock. Climate vulnerability comes from the ND-GAIN Country Index, using only the vulnerability axis; ND-GAIN's separate readiness axis is left out to avoid re-importing governance. Conflict intensity, in deaths per 100,000 over the five years 2019 to 2023, comes from the Uppsala Conflict Data Program's Georeferenced Event Dataset. Internally displaced persons per 100,000 and refugees originating from the country per 100,000 both come from UNHCR. The last of these also appears in constrained mobility, and the documentation flags this shared use as a possible double-count awaiting a correlation check.

Phase E, Exploitation, asks whether exploitation can run unchecked, and holds three domains.

Foreclosed exit. This domain is meant to capture the cost of walking away: whether a worker in a bad situation has any realistic alternative. Its three indicators are all from ILOSTAT and all protective, entered inverted: labour inspectors per ten thousand employed persons, the share of employees covered by collective agreements, and trade union density. The project is candid that this domain does not measure what it was designed to measure. The intended mechanism, the cost of exit and the degree to which a single employer dominates a local labour market, cannot be sourced across nearly 200 countries. What is available are protective factors whose absence is being read as risk, which is not the same thing. Coverage runs from about 43 to 69 per cent. The domain is carried as insufficient data and is explicitly labelled a stand-in; whether it should be scored at all, folded into a neighbouring domain, or held aside as a diagnostic, remains an open question. Because Exploitation has only three domains, this weakness matters more than it would in Recruitment.

Economic structure and demand. Some economies simply have more of the work in which forced labour occurs. Three indicators: hazardous-sector share, using employment in agriculture, anchored from 0 to 60 per cent; informal employment share, from ILOSTAT; and export concentration, from UNCTADSTAT, a measure of how concentrated a country's merchandise exports are in a small number of products, used here as a proxy for concentrated buyer power. Two caveats are recorded. The export concentration measure is a low-confidence proxy for the buyer concentration the designers actually wanted. And the criminal-market side of demand has no sourced indicator at all, because the obvious candidate, organised-crime outcome measures, was excluded as circular. Agriculture and informality also appear in Recruitment, a cross-phase overlap that the documentation names as an unresolved issue.

State production of unfreedom. The proposition is that states can generate unfreedom directly, through the legal architecture they impose and the impunity they permit. The published build scores it from clientelism, from V-Dem, capturing the exchange of goods and favours for political support; bribery risk, from the TRACE International Bribery Risk Matrix; firm bribery incidence, from the World Bank, being the share of firms reporting at least one request for a bribe; the forced-labour share of detected trafficking victims, derived from the UNODC Global Report on Trafficking in Persons, covering 142 countries; and the DEMSCORE tied-status coding together with the eight-country kafala coding. Two things must be said about this domain. First, only two of five designed drivers are sourced, which is why it is carried as insufficient data; the tied-status, immigration-architecture, and protective-floor legal codings that the design calls for do not exist as global series. Second, the domain is where the index comes closest to being a corruption measure, which is exactly why it receives the special treatment described in the next section.

The Monetization lens, computed but excluded. A third phase covers the financial conditions under which the proceeds of exploitation can be moved, hidden, and retained. It has two domains. Transnational concealment uses the FATF mutual-evaluation effectiveness score and grey-list or black-list standing, both taken from the Basel AML Index published by the Basel Institute on Governance, together with the Tax Justice Network Financial Secrecy Index. Cash and informal retention uses the World Bank Informal Economy Database estimate of the shadow economy as a share of official GDP, and the share of adults without a financial account. This phase is computed, published, and deliberately kept out of the headline score. The reason given is that it answers an intervention question, namely where the money could be disrupted, rather than a structural risk question, and that it scores high for wealthy, financially opaque economies in a way that would distort the risk reading. The project reports having tested whether any governance-independent part of the lens deserves inclusion, and found that the remainder carried no forced-labour-specific signal and diluted the composite. The lens therefore remains display-only. The project frames this distinction as one between drivers and disruptors: Recruitment and Exploitation generate risk, while Monetization is where the cycle could be broken without first removing the underlying vulnerability.

07

Guarding Against Circularity and the Governance Problem

Circularity. An index that predicts forced labour using a measure of forced labour tells you nothing. The project applies this discipline in several places. The United States Trafficking in Persons report, organised-crime outcome measures, and the Basel AML composite were excluded as circular. Detection counts of trafficking victims were refused as a validation benchmark on the ground that detection reflects state capacity rather than prevalence, so matching it would reward strong states. In the published build, the V-Dem indicator measuring freedom from forced labour was dropped entirely from the scoring, on the ground that using a freedom-from-forced-labour measure to predict forced-labour risk is circular by construction.

The discipline is not applied perfectly, and a careful reader should notice the tension. Detection data is refused as a benchmark for checking the index, yet one derived detection measure, the forced-labour share of detected trafficking victims from the UNODC reporting, does enter the scoring of the state production of unfreedom domain. The share of detected cases is not the same quantity as the count of detected cases, and it is less directly a function of state capacity, so the two positions can be reconciled. But the reconciliation is not spelled out in the documentation, and it is the one place where the project's own circularity rule is applied less strictly to an input than to a benchmark.

The governance problem. Most cross-national indicators correlate with the quality of a country's institutions. An index assembled from such indicators can easily become a rule-of-law ranking with a new name, which would make it uninformative for the specific question at hand.

The published response is called de-biasing. Each indicator was tested for how much of its variation is explained by a standard rule-of-law measure. Any indicator explained at or above 55 per cent was treated as too entangled with governance and entered at half weight, applied within the state production of unfreedom domain, which is where such indicators concentrate. This is a targeted reduction rather than a removal.

The result is reported honestly rather than presented as solved. After de-biasing, rule of law still explains about 63 per cent of the variation in the composite, corresponding to a correlation of about 0.79. The project's position is that this residual is correct rather than an artefact: weak governance is a genuine structural driver of forced-labour risk, so a substantial but bounded association is what one should expect. Roughly a third of the variation is not governance.

That position was stress-tested. The de-biasing was re-run varying both its scope, applying it to the one domain or to all domains, and the reference measure used to flag entangled indicators, using the World Bank rule-of-law measure, the V-Dem rule-of-law measure, or an equal blend of the two. Across every combination, rank agreement with the published index stayed at or above 0.96 on Kendall's measure of rank correlation, the governance share stayed near 63 per cent, the top five and bottom five countries were identical, and the result for the Gulf states was unchanged. The check can be re-run by anyone with the repository. This is the strongest available answer to the objection that the index is governance with a new label, and it is a more thorough answer than most composite indices provide.

One nuance is worth noting for readers who look at the code. An older and simpler mechanism, in which a domain score was reduced in proportion to a governance dial, survives in a secondary reproduction build that the repository also ships. It is not the mechanism behind the published figures. The documentation distinguishes the two carefully, and also notes that this secondary build is more governance-dominated than the published one.

08

Reading the Results

The scale. Scores run from 0 to 1. In the published build the scored range is narrower than the theoretical one, running from about 0.11 to about 0.67. The highest-scoring countries are Yemen, Chad, Afghanistan, South Sudan, and Sudan; the lowest are the Nordic countries, Iceland, and several Western European states. This ordering is a face-validity check, not a discovery.

Read tiers, not ranks. The index is banded into three tiers using cuts at 0.281 and 0.402, giving 59 countries in the higher tier, 60 in the middle, and 65 in the lower tier. The cut points were fixed at the registration stage and deliberately not re-derived once the results were known, on the principle that banding thresholds must not be tuned after seeing the output. The project notes that the current build's own thirds would fall at 0.274 and 0.397, and that nine boundary countries would change tier if the cuts were re-derived.

Uncertainty bands. Every scored country carries a rank band produced by running the whole scoring 10,000 times with small random disturbances added to its two phase scores and with the balance between the phases redrawn each time between 40 and 60 per cent. The band reported is the range within which the country's rank fell in 90 per cent of those runs.

These bands are the single most useful discipline the index publishes. The median band is 45 ranks wide, and only about 2 per cent of countries have a band narrower than 10 ranks. In plain terms, a mid-table country's exact position is close to meaningless, while the extremes are stable: the top decile retains about 79 per cent of its members across the simulations, and a country stays in its published tier in about 84 per cent of runs on average.

The low-confidence badge. Forty scored countries have two or more of the eleven domains unscored. These carry a visible lower-confidence marker on the site, and the same rule widens their uncertainty band in the simulation. One rule drives both statements, which prevents the presentation and the arithmetic from drifting apart. The uncertainty code notes that an earlier version of this rule was applied incorrectly, silently widening the band for every country, and records the date the error was corrected.

What a score is not. A high score does not mean a specified number of people are exploited. A low score does not certify a country as free of forced labour. A difference of a few places in the middle of the table is not a finding.

09

Beyond the National Average

A national score averages an entire country into one number, and that average hides the regions where risk actually concentrates. The project therefore publishes a sub-national layer covering first-level administrative regions, built from census microdata from IPUMS-International, an archive of harmonised census samples from around the world, with disaster exposure taken from the Geocoded Disasters dataset.

The published surface covers 1,735 first-level administrative units, of which 1,454 pass reliability filtering and carry a risk score. Roughly 17 per cent of the variation in the underlying precarity measure sits within countries rather than between them, which is the quantitative form of the argument that national averages conceal a great deal.

The project also publishes a corridor view, on the reasoning that conditions enabling forced labour frequently span neighbouring states and that a response confined to one country addresses only part of a cross-border problem.

These layers are best read as illustrative rather than authoritative. They rest on a restricted-access data source that cannot be redistributed with the repository, and they cover fewer countries than the national index.

10

Validation

The project registered its validation criteria in advance, including the thresholds at which each test would be judged to have failed, and then re-ran them on the build that the site actually displays. Pre-registration matters here: it prevents the criteria from being adjusted after the results are known.

The reported outcomes on the displayed build are as follows. Rule of law explains 0.628 of the composite's variation, passing the pre-set requirement that it stay at or below 0.80. The two phases correlate at 0.659, which is within the pre-set window of 0.30 to 0.90 and supports the claim that they are related but not redundant. A forced-labour-specific signal, child labour prevalence drawn from an independent source, remains significantly associated with the composite once governance is held constant on both sides. The top decile retains 86.5 per cent of its members under the suite's mild-noise test, against a requirement of 70 per cent. The published uncertainty model, which disturbs the inputs more aggressively, gives a lower figure of about 79 per cent for the same quantity; both are reported.

One criterion was not met, and the project reports it rather than dropping it. Tested against an external prevalence estimate, the Walk Free Global Slavery Index country prevalence figures, with governance netted out on both sides, no significant association was demonstrated. The project's own reading is that this null result is uninformative rather than disconfirming, because the benchmark is itself heavily entangled with governance. That is a defensible reading, but it is also an admission that the index has no clean external check.

The absence of a governance-independent measure of forced labour prevalence is precisely the gap that motivates a structural index, and it is also the reason the index cannot be fully validated. Readers should hold both facts at once.

A note on what was validated. The tests examine the structure: whether the index is distinguishable from governance, whether its two halves are distinct, whether it carries forced-labour-specific information, and whether its extremes are stable. They do not validate any individual country's score.

11

Limitations and Caveats

The project's credibility rests substantially on the candour of this list, which is reproduced here in the terms the repository itself uses.

It measures conditions, not cases. The index does not estimate prevalence and cannot be read as a count of victims.

It reads origin-side risk and under-reads destination systems. This is the most consequential limitation. The mechanisms that drive forced labour in destination economies, namely sponsorship systems that tie a worker's legal status to one employer, recruitment-fee debt, and migration brokerage, are named in the framework but are not sourced at country scale, beyond the narrow hand-coded signal covering eight states. The available indicators describe resident citizens rather than the migrant workforce most exposed. As a direct result, the United Arab Emirates ranks 146th of 184 scored countries and the other Gulf states sit near it, despite well-documented risk in their labour systems. A low score for a known destination country means that this index does not yet capture that pathway, not that the country is clear. The project mitigates this by attaching a caveat banner to the eight sponsorship-system country profiles, which makes the blind spot visible where a reader would otherwise be misled.

It is correlated with weak governance by design. About two-thirds of the variation is shared with rule of law. The relationship is disclosed and stress-tested rather than engineered away, but a reader who wants a measure independent of institutional quality will not find one here.

The Exploitation phase is thin. One of its three domains does not measure its intended mechanism at all and is carried as insufficient data. Another leans partly on corruption proxies. Because the phase has only three domains, each carries roughly 2.7 times the marginal weight of a Recruitment domain, so these weaknesses propagate further than they would elsewhere.

Some indicators do double duty. Informality and agricultural sector share enter both phases, and refugee outflow enters two Recruitment domains. The collinearity screen that would resolve this is flagged in the rules but has not been run and reported.

Equal weighting is a substantive claim. It has not been tested by perturbation at the indicator and domain level, only at the phase level.

Several anchors were derived from the observed distribution. They were not derived from theory, which weakens the claim that the scale is absolute. An anchor-shift sensitivity test is outstanding.

Coverage is uneven and the gaps are not random. The child labour indicator, which measures the mechanism most directly, covers under half the countries and draws on survey vintages up to two decades old. The unscored countries are systematically small states. Thin coverage tends to depress rather than inflate a score.

Some inputs are licence-restricted. Several sources cannot be redistributed with the repository, and several others carry re-publication flags that the documentation itself marks as unconfirmed, including the Henley Passport Index, the TRACE Bribery Risk Matrix, the UNDP Gender Inequality Index republication, and the Tax Justice Network Financial Secrecy Index. EM-DAT and IPUMS-International are recorded as redistribution-restricted and are not bundled. Anyone republishing this material should resolve those questions first.

The documentation does not describe the published scoring quite completely. The generated codebook lists the indicators drawn from the main data pipeline and states that it cannot drift from that pipeline, but the published scorer additionally enters three inputs held outside it: the UNODC detected-victim composition measure, the DEMSCORE tied-status coding for 29 countries, and the kafala tied-status coding for eight countries. These are listed on the site's own indicator data and are therefore visible to a reader who looks, but they are absent from the codebook, and the narrower coverage of two of them is not reflected in the domain confidence flags. A future revision should bring the codebook and the published scorer back into alignment.

The mid-table is not readable as a ranking. With a median uncertainty band of 45 ranks, ordinal comparisons between similarly placed countries are not supportable.

There is no clean external benchmark. The index cannot be checked against a governance-independent measure of forced labour prevalence, because none exists.

It is a research prototype. It was produced as masters research, it is published as a framework and a pipeline rather than as authoritative country figures, and it should not be used as the sole basis for any consequential decision about a country, a supplier, or a person.

12

What Is Published, and How It Can Be Checked

The deliverable is an interactive website that runs in an ordinary browser and can be served from any static host. It carries a world map, a sortable ranking of all 184 scored countries with the unscored cases shown openly, individual country profiles with a phase and domain breakdown, the full indicator and source list, a limitations page, and a simulation page.

The simulation page lets a reader change the balance between the two phases, switch the combining rule from a geometric mean to a plain average, and move a country's domain scores, watching the map and rankings recompute. Its own description is careful: this shows how sensitive the index is to the choices made in building it, and is not a forecast of anything, nor a prediction of what any intervention would achieve.

Every figure shown on the site is read from the published build output rather than entered by hand, and the whole build can be regenerated with a single command from the inputs stored in the repository. The rebuild verifies its own output against the published baseline and stops if anything has drifted unexpectedly. Roughly two-thirds of the data sources can be re-pulled automatically from public interfaces; the remainder require a human to obtain a registration-gated or licence-gated file first, and each of these is listed with its provider, its address, and its licence terms.

The per-indicator source, vintage, coverage, licence, and required citation are recorded for every signal. The codebook that documents which indicators sit in which domain is generated from the code itself, so it cannot drift away from the pipeline it describes. These are small pieces of discipline, and they are the reason the rest of this report could be written from the repository alone.

13

Intended Audience and Use

The index is addressed to those working on prevention rather than prosecution: labour ministries and inspectorates, humanitarian and development programmers, procurement and due-diligence teams carrying responsibilities under the UN Guiding Principles, and researchers.

Its stated value is as a triage instrument. It points to where vulnerability and unchecked exploitation hold together, consistently across countries that are otherwise difficult to compare, so that attention can be directed before harm becomes visible. It is a starting point for inquiry, not a verdict on any country.

The project is explicit about the boundary of the intervention material it publishes. Identifying where the structure is sensitive is not the same as predicting what an intervention would achieve. The step from noticing a lever to knowing that pulling it works in a particular place belongs to evaluation on the ground.

Three disciplines govern responsible use: read the structure rather than a body count; read inside the country rather than only the national figure, where the sub-national layer allows it; and read tiers rather than ranks.

14

Conclusion

FLSRI addresses a real and well-recognised problem: the places where forced labour is most likely are frequently the places where it is least well recorded, which leaves prevention work waiting for evidence that arrives late or not at all. By measuring conditions rather than cases, the index offers a way to prioritise attention without pretending to count what cannot yet be counted.

Its most substantial contribution is conceptual and procedural rather than numerical. The insistence that a country scores high only where both an exposed population and an unchecked environment are present is a genuine structural claim, encoded in the arithmetic rather than asserted in prose. The refusal to convert missing data into zero, the exclusion of inputs that would predict forced labour from a measure of forced labour, and the decision to publish uncertainty bands wide enough to embarrass the ranking are all choices that make the index harder to over-read.

The honest summary of its current state is that its structure is defensible and its individual country figures are not yet findings. Its Exploitation half rests on one domain that does not measure what it was designed to measure. It under-reads the sponsorship systems that drive some of the most thoroughly documented forced labour in the world, and it says so in the same breath as it publishes the ranks that reflect that gap. It shares a great deal of its variation with measures of institutional quality, and it demonstrates the bound rather than denying the overlap.

What makes the work useful is that these limits are not concessions extracted from it but statements it makes about itself, in its documentation, in its code comments, and on the face of the published site. An index that tells its reader where not to trust it is more usable than one that does not, and that is the standard by which this prototype should be judged and, in time, improved.

References

Sources

  1. 01International Labour Organization. Forced Labour Convention, 1930 (No. 29), Article 2(1); Abolition of Forced Labour Convention, 1957 (No. 105); and the 2014 Protocol to Convention No. 29.
  2. 02International Labour Organization. ILO Indicators of Forced Labour, first issued 2012, revised edition 2025.
  3. 03United Nations. Guiding Principles on Business and Human Rights, endorsed by the UN Human Rights Council in 2011.
  4. 04United Nations. Sustainable Development Goal Target 8.7, with SDG indicators 8.3.1 and 16.9.1 used directly by the index.
  5. 05World Bank. World Development Indicators, including poverty headcount, agricultural employment, GDP growth volatility, the Gini index, birth registration completeness, and firm bribery incidence.
  6. 06International Labour Organization. ILOSTAT, including informal employment share, labour inspector density, collective bargaining coverage, trade union density, and sex-by-sector employment.
  7. 07Varieties of Democracy project. V-Dem dataset, freedom of movement and clientelism indicators.
  8. 08UNHCR. Refugee population statistics, covering refugees originating from a country, internally displaced persons, and statelessness.
  9. 09Uppsala Conflict Data Program. Georeferenced Event Dataset, conflict deaths 2019 to 2023.
  10. 10CRED, UCLouvain. EM-DAT, the international disaster database, disaster-affected intensity 2020 to 2024.
  11. 11World Bank. Global Findex Database, financial account access and borrowing behaviour.
  12. 12Henley and Partners. Henley Passport Index, visa-free destination access.
  13. 13ETH Zurich. Ethnic Power Relations dataset, politically excluded ethnic groups.
  14. 14United Nations Development Programme. Gender Inequality Index, Human Development Report.
  15. 15World Bank. Women, Business and the Law, mobility component.
  16. 16Notre Dame Global Adaptation Initiative. ND-GAIN Country Index, vulnerability axis.
  17. 17UNCTAD. UNCTADSTAT, merchandise export concentration.
  18. 18TRACE International. Bribery Risk Matrix.
  19. 19UNODC. Global Report on Trafficking in Persons, forced-labour share of detected victims, 142 countries.
  20. 20DEMSCORE. Legal-status data used for the tied-status coding covering 29 countries.
  21. 21Basel Institute on Governance. Basel AML Index, FATF effectiveness scores and list standing, used in the Monetization lens only.
  22. 22Tax Justice Network. Financial Secrecy Index, used in the Monetization lens only.
  23. 23World Bank. Informal Economy Database, shadow economy as a share of official GDP.
  24. 24IPUMS-International. Harmonised census microdata behind the sub-national layer, with the Geocoded Disasters dataset for sub-national disaster exposure.
  25. 25Walk Free. Global Slavery Index, used as the external prevalence benchmark and for vulnerability dimension alignment.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file +

Publications · Academic report

The Forced Labor Structural Risk Index

A Country-Level Measure of the Conditions Under Which Forced Labour Becomes More Likely

Ethical Tech CoLab

NYU Center for Global Affairs

December 2025

Prepared as masters research at the NYU Center for Global Affairs.

Forced labour is one of the few large-scale human rights violations that almost nobody can count, and the countries where the risk is gravest are frequently the countries where the evidence is thinnest, because the same institutional weakness that permits exploitation also prevents it from being recorded. FLSRI does not try to estimate how many people are in forced labour. It measures something different and, for prevention purposes, more tractable: how strongly the structural conditions that make forced labour more likely are present, and how those conditions combine.

Key figures

184

countries of roughly 195 receive a composite score; the rest are reported as not scored rather than given an invented value

43

standardised indicators across eleven domains, each drawn from a named cross-national dataset

10,000

simulated re-scorings behind the uncertainty band published for every scored country

45

ranks wide is the median uncertainty band, which is why the index asks to be read in tiers rather than as a league table

01

Executive Summary

FLSRI is a research prototype: a country-level index that scores the structural conditions under which forced labour becomes more likely. It scores 184 of roughly 195 countries on a scale running from 0 to 1, where 0 is the lowest modelled structural risk and 1 the highest. The remaining countries are reported as not scored rather than being given an invented value.

The index is deliberately not a prevalence estimate. It does not say how many people are exploited. A high score means the enabling conditions hold together strongly; a low score does not certify a country as free of forced labour. This distinction is stated repeatedly in the project's own documentation and it governs every legitimate use of the results.

The framework rests on a simple structural claim borrowed from criminology: exploitation becomes likely when a population exposed to coercive recruitment coincides with an environment in which exploiting that population goes unchecked. The index therefore has two scored halves. Recruitment, written as R, asks who is structurally exposed, through poverty, debt, blocked mobility, exclusion, weak legal identity, gendered labour structures, childhood exposure, and shocks such as conflict and disaster. Exploitation, written as E, asks whether exploiting them can run without consequence, through blocked exit, demand in high-risk sectors, and the state's own production of unfreedom.

The two halves are combined by a geometric mean, that is, the square root of R multiplied by E. This is a deliberate methodological choice. It means a country can only score high when both halves are high. A very exposed population in a country with strong labour enforcement, or a permissive enforcement environment in a country with little structural vulnerability, is pulled down rather than averaged out.

A third phase, Monetization, covering the financial conditions under which the proceeds of forced labour can be moved, hidden, and kept, is computed and published but deliberately excluded from the headline score. It answers a different question, namely where intervention against the money would bite, and it scores high for wealthy financially opaque economies in a way that would distort a structural risk reading.

Eleven domains and forty-three standardised indicators feed the published score, each drawn from an established cross-national dataset such as the World Bank World Development Indicators, ILOSTAT, the V-Dem democracy dataset, the UNHCR population statistics, the UCDP conflict event dataset, and the EM-DAT disaster database.

The project is unusually disciplined about two failure modes that afflict composite indices of this kind. The first is circularity, that is, predicting forced labour using a measure of forced labour. The second is the risk that the index becomes a governance ranking wearing a new label. Both are addressed explicitly, tested, and reported rather than concealed.

Every scored country carries a published uncertainty band derived from 10,000 simulated re-scorings, and the documentation insists that the results be read in broad tiers rather than as an exact league table. The middle of the table is openly described as unstable. The index is delivered as an interactive public website together with the complete pipeline that produces it, so that any figure shown can be traced back to the data and the code that generated it.

02

Background and Rationale

The problem. Forced labour is defined in international law and prohibited almost universally, yet it is measured very unevenly. Global estimates exist, but they are produced from surveys and administrative sources that are not available for every country, are refreshed slowly, and were never designed to support a fine-grained country-by-country comparison. Detection statistics, such as counts of identified trafficking victims, are worse still for this purpose: they largely record how much capacity a state has to detect and register cases, so a well-resourced state can appear worse than a state where nothing is investigated at all.

The gap. Prevention and triage need to know where conditions are dangerous before cases surface. Existing vulnerability measures go some way toward this, but several of them are built partly from the same governance material that dominates most cross-national comparison, which makes it hard to tell whether they are measuring forced-labour risk or simply restating institutional weakness.

The response. FLSRI proposes a purpose-built structural measure. Its guiding commitments, visible throughout the code and documentation, are that every figure should be traceable to a named public dataset, that missing information should be declared rather than filled in, that the reasoning behind each methodological choice should be written down, and that the limits of the result should be published as prominently as the result itself.

The project describes itself as publishing a framework, a pipeline, and code, with the country scores presented as estimates of structural conditions carrying explicit uncertainty. It does not present the country rankings as findings about the world.

03

Objectives

The index is designed to do the following.

  • Measure the structural conditions associated with forced labour across almost all countries, on a single comparable scale, without estimating prevalence.
  • Make the underlying theory explicit, by separating exposure to recruitment from the conditions for unchecked exploitation and requiring both to be present before a country scores high.
  • Ground every indicator in a real, citable, cross-national dataset, with the source, vintage, coverage, licence, and required citation recorded for each one.
  • Refuse to fabricate. Missing data is never silently converted to zero, and countries whose evidence base is too thin are left unscored.
  • Guard against circularity by excluding inputs that are themselves measures or detections of forced labour and trafficking, wherever using them would mean predicting forced labour with a measure of forced labour.
  • Report, rather than engineer away, the index's association with weak governance, while demonstrating that the index is not reducible to a governance ranking.
  • Publish uncertainty alongside every score, and direct readers toward tiers rather than exact ranks.
  • Support prevention and triage by identifying where structural risk concentrates, including below the national level, and where in the structure a government, a buyer, or a non-governmental organisation might act.

04

The Legal and Normative Frame

Forced labour has a settled definition in international law. The International Labour Organization's Forced Labour Convention, 1930 (No. 29), Article 2(1), defines forced or compulsory labour as all work or service which is exacted from any person under the menace of any penalty and for which the said person has not offered himself voluntarily. The Abolition of Forced Labour Convention, 1957 (No. 105), prohibits the use of forced labour for specified purposes including political coercion, labour discipline, and discrimination. The 2014 Protocol to Convention No. 29 obliges ratifying states to take effective measures on prevention, on the protection of victims, and on access to remedies including compensation. It entered into force in November 2016.

To help identify forced labour in practice, the ILO publishes a set of eleven indicators of forced labour, first issued in 2012 and reissued in a revised edition in 2025. The eleven are these.

  • Abuse of vulnerability.
  • Deception.
  • Restriction of movement.
  • Isolation.
  • Physical and sexual violence.
  • Intimidation and threats.
  • Retention of identity documents.
  • Withholding of wages.
  • Debt bondage.
  • Abusive working and living conditions.
  • Excessive overtime.

The ILO describes their purpose as assisting law enforcement officials, labour inspectors, trade union officers, and non-governmental organisation staff to identify people who may be trapped in forced labour and may need urgent help.

The ILO is careful about how much weight any one indicator carries. Its guidance states that the presence of a single indicator may in some cases imply the existence of forced labour, while in other cases several indicators must be read together before a case can be said to exist. The eleven are a screening aid for practitioners, not a legal test, and they are frequently over-read as a checklist. That caution matters for an index built on the same vocabulary.

These are operational indicators for recognising a case in front of you. FLSRI works at a different level: it measures the standing country conditions that make those case-level indicators more likely to arise, and its documentation states that its domain set is broadly aligned with these indicators and with the vulnerability dimensions used in the Walk Free Global Slavery Index.

The obligation is not only a state obligation. The United Nations Guiding Principles on Business and Human Rights, endorsed by the UN Human Rights Council in 2011, rest on three pillars: the state duty to protect human rights, the corporate responsibility to respect them, and access to remedy for those harmed. The Guiding Principles on human rights due diligence require enterprises to identify, prevent, mitigate, and account for how they address their human rights impacts, including impacts in their supply chains and business relationships. A structural risk measure is directly relevant to that first step: an enterprise or a public procurement body cannot prioritise due diligence without some defensible view of where risk is concentrated.

Target 8.7 of the Sustainable Development Goals commits states to take immediate and effective measures to eradicate forced labour, end modern slavery and human trafficking, and secure the prohibition and elimination of the worst forms of child labour. Two of the index's own indicators are drawn directly from the SDG monitoring framework: informal employment as a share of total employment, which is SDG indicator 8.3.1, and the completeness of birth registration, which is SDG indicator 16.9.1.

The index does not claim to establish a violation of any of these instruments, and it should not be cited as evidence that a particular state is in breach of its obligations. It is a prioritisation instrument for prevention, not a compliance finding.

05

How the Index Is Built

The index has a fixed shape with four levels: indicator, domain, phase, and composite. Each level sits on the same 0 to 1 scale.

The indicator. An indicator is a single measured quantity taken from a real dataset, for example the share of the population living below the 6.85 dollar a day poverty line, or the number of labour inspectors per ten thousand employed people. Each one arrives in its own units, so the pipeline rescales it onto the common 0 to 1 risk scale.

The domain. A domain is a group of indicators measuring one mechanism, for example economic precarity or constrained mobility. The domain score is the plain average of the indicators that are present for that country. Indicators that are missing are dropped from the average and are never entered as zero.

The phase. A phase is one side of the structural claim. Recruitment contains eight domains; Exploitation contains three. A phase score is the plain average of the domains scored inside it. The eleven are set out below and described one by one in the next section.

PhaseDomainMechanism it is meant to capture
RecruitmentEconomic precarityNeeds that the available work cannot meet
RecruitmentDebt and financialised dependencyObligation that turns a voluntary arrangement into an inescapable one
RecruitmentConstrained mobilityInability to move legally or affordably
RecruitmentAscriptive exclusionExclusion by group membership rather than by conduct
RecruitmentLegal non-recognitionInability to prove who you are and so to claim protection
RecruitmentGendered labourConcentration of women in exploitation-exposed work
RecruitmentAge and childhood structuringDirect exploitability of children and household labour exhaustion
RecruitmentStructural disruptionConflict, displacement, and disaster shocks
ExploitationForeclosed exitThe cost of walking away from a bad situation
ExploitationEconomic structure and demandHow much of the economy sits in high-risk work
ExploitationState production of unfreedomUnfreedom generated by legal architecture and impunity
The eleven scored domains and the phase each belongs to. The Monetization lens is computed separately and excluded from the headline score.

The composite. The published score is the geometric mean of the two phases, that is, the square root of Recruitment multiplied by Exploitation. If either phase cannot be scored for a country, the composite is not scored. There is no substitution of zero for a missing phase.

Indicators are rescaled by what the project calls absolute anchoring. For each indicator, a floor value and a ceiling value are fixed in advance. The floor is the raw value that maps to 0 and the ceiling is the raw value that maps to 1; anything beyond either end is clamped.

IndicatorFloor, mapping to 0Ceiling, mapping to 1
Share of employment in agriculture0 per cent80 per cent
Gender gap in labour force participation0 percentage points50 percentage points
Conflict deaths0 per 100,000 people100 per 100,000 people
Three examples of the anchor ranges the pipeline uses.

Anchoring to fixed reference points rather than to the best and worst countries observed is a deliberate choice. It means a country's score does not move simply because a different set of countries happened to have data this year, and it makes scores comparable across refreshes of the underlying data. Where a quantity depends on the size of a country, it is first converted to a rate, a share, or a gap before scaling, so that a large country is not scored as riskier merely for being large.

Each indicator carries an explicit direction. Most point the same way as risk: more poverty means more risk. Some are protective and are inverted, so that a high raw value becomes a low risk score. Trade union density, collective bargaining coverage, labour inspector density, freedom of movement, and visa-free passport access are all protective indicators entered in inverted form. By the time any indicator reaches the scoring stage, every value points in the same direction, with higher meaning more risk.

The documentation is candid that several anchors were set from the observed distribution, for instance a ceiling near the 95th percentile, which reintroduces a degree of dependence on the particular sample of countries into a scale that is presented as absolute. Locking these anchors and testing how much the results move if they shift is listed as outstanding work.

Every level uses equal weights. Indicators inside a domain are averaged equally, domains inside a phase are averaged equally, and the two phases enter the geometric mean equally. The documentation is unusually direct that this is a value judgment rather than a neutral default. Equal weighting asserts that each component is equally relevant to the construct, which is a substantive claim and not an absence of one.

Equal weighting at each level produces unequal weight per indicator once the whole structure is taken into account, because the boxes are not the same size. A domain containing one indicator gives that indicator the full weight of a domain, while a domain containing five splits it. More consequentially, the Exploitation phase contains three domains against Recruitment's eight, so a single Exploitation domain carries roughly 2.7 times the marginal weight of a single Recruitment domain. The project's own interactive material states this plainly rather than leaving it to be discovered.

The project reports that shifting the balance between the two phases as far as 60 to 40 in either direction barely moves the ranking. It also states that it has not yet published a similar test for perturbing weights at the indicator and domain level, and records this as a known gap.

The governing principle on missing data is that it is never silently treated as zero. A missing indicator is dropped from its domain average rather than counted as an absence of risk. Dropping cannot go on indefinitely, so a coverage floor applies. A domain is scored only if at least half of its mapped indicators are present and never fewer than two. The same floor applies to the count of scored domains within a phase. Below the floor, the domain or phase is marked not scored. A domain that is designed around a single indicator cannot mechanically meet the requirement for two; these are handled as a named exception, scored if the one indicator is present but always carried as low confidence.

Under these rules, 184 of roughly 195 countries receive a composite score. The eleven unscored countries are Andorra, Dominica, the Federated States of Micronesia, Saint Kitts and Nevis, Liechtenstein, Monaco, the Marshall Islands, Nauru, San Marino, Tuvalu, and the Holy See. The documentation notes that this missingness is not random: these are micro-states and small island states for which the labour and governance series are simply not collected. It also warns that within the scored set, thin coverage tends to push a score down rather than up, so a very low score for a data-sparse state deserves extra caution.

06

The Variables, Explained in Plain Terms

This section is the heart of the report. It sets out, domain by domain, what each indicator represents, why it was chosen, and how it enters the score. All indicators are averaged equally within their domain. Phase R, Recruitment, asks who is structurally exposed, and holds eight domains.

Economic precarity. The proposition is that people who cannot meet their needs through the work available to them accept terms they would otherwise refuse. Five indicators carry it. Poverty headcount is the share of the population living on less than 6.85 dollars a day, from the World Bank World Development Indicators, anchored from 0 to 100 per cent. Informal employment share is the proportion of total employment that is informal, taken from ILOSTAT as SDG indicator 8.3.1; informal work sits outside labour inspection, contract enforcement, and social protection, which is precisely the exposure the domain is trying to capture. Agrarian employment share is the proportion of employment in agriculture, anchored from 0 to 80 per cent, because agriculture concentrates seasonal, isolated, and poorly regulated work. Income volatility is the standard deviation of real GDP growth over roughly fifteen years, anchored from 0 to 8 percentage points, on the reasoning that instability and not only low income is what pushes households into distress decisions. Income inequality is the Gini index, anchored from 20 to 65.

Debt and financialised dependency. Debt is the classic mechanism by which a voluntary arrangement becomes an inescapable one, and debt bondage is one of the ILO's eleven case-level indicators. All three measures come from the World Bank Global Findex Database, a large survey of how adults around the world save, borrow, and make payments: the share of adults with no financial account, the share who borrowed from informal sources in the past year, and the share who borrowed any money in the past year. The domain is carried as low confidence, and the reason is stated openly. The mechanism the designers actually wanted, namely recruitment-fee and migration debt, has no cross-national source, so the domain rests on financial exclusion and borrowing patterns instead. Borrowing prevalence in particular measures how widely credit is used, not whether households are in distress. The borrowing questions cover only about 59 per cent of countries.

Constrained mobility. The idea is that people who cannot move legally or affordably are easier to recruit on bad terms and harder to extract from a bad situation. Freedom of movement comes from V-Dem, an expert-coded dataset on the qualities of political regimes, entered inverted. Passport access is the number of destinations reachable without a visa, from the Henley Passport Index, also inverted, on the reasoning that a weak passport means legal migration routes are scarce and irregular ones become the alternative. Refugees originating from the country, per 100,000 people, come from the UNHCR population statistics. A tied-status coding is derived from the DEMSCORE legal-status data, covering 29 countries, alongside a hand-coded kafala tied-status signal covering eight sponsorship-system states.

The kafala signal. Under sponsorship systems used across parts of the Gulf and the wider region, a migrant worker's legal status is tied to a single employer, so leaving that employer can mean losing the right to remain. The index codes three features for the eight adopting states it covers: whether status is tied to one employer, whether the employer's consent is required to leave, and whether leaving is treated as an offence. This is a narrow, hand-built input covering a handful of countries, not a global series, and the project is explicit that the sponsorship mechanism remains unsourced at country scale. The domain is carried as low confidence.

Ascriptive exclusion. Exclusion by group membership, meaning by who a person is rather than what they have done, is captured by a single indicator: the depth-weighted share of the population belonging to politically excluded ethnic groups, from the Ethnic Power Relations dataset maintained at ETH Zurich, which codes ethnic groups' access to executive power. Two limitations are recorded. The indicator captures political exclusion, not exclusion within the labour market, which is the channel most relevant to forced labour. And roughly 41 countries lie outside the dataset's universe and are left missing rather than scored as zero. The domain is single-indicator and therefore always carried as low confidence.

Legal non-recognition. A person who cannot prove who they are cannot easily enforce a contract, claim a wage, access protection, or cross a border lawfully. Birth-registration incompleteness is derived from the completeness of birth registration reported through the World Bank from UNICEF and UN Statistics Division civil registration data, corresponding to SDG indicator 16.9.1, entered inverted. Statelessness prevalence, in persons per 100,000, comes from the UNHCR population statistics. Where statelessness figures are suppressed or absent, the domain rests on birth registration alone; 26 countries are in that position and are flagged as low confidence.

Gendered labour. Forced labour is strongly gender-patterned, particularly in domestic work, care work, and sectors where women are concentrated. Four indicators carry the domain: the gap between male and female labour force participation, in percentage points, anchored from 0 to 50; the UNDP Gender Inequality Index, covering reproductive health, empowerment, and labour market participation, running from 0 for equality to 1 for maximum inequality; legal constraints on women's mobility, taken from the mobility component of the World Bank Women, Business and the Law project, entered inverted; and sex-by-sector channelling from ILOSTAT, the extent to which one sex is concentrated in exploitation-exposed sectors of employment. The documentation is careful to record that a global prevalence estimate is used only to justify why this mechanism matters, and is never scored. The value entered is always a structural employment share, never a victim count.

Age and childhood structuring. Children are both directly exploitable and an indicator that households are already using every available source of labour. Four indicators, all from the World Bank's mirrors of ILO and UNICEF SDG series: child labour prevalence, being the share of children aged 7 to 14 in employment; the out-of-school rate for lower-secondary-age adolescents; the share of the population aged 0 to 14; and the child marriage rate, the share of women aged 20 to 24 first married before 18. This domain carries the project's most awkward data problem. Child labour prevalence, the indicator that most directly measures the mechanism, covers only about 47 per cent of countries, from survey vintages spanning 2005 to 2016. That is below the coverage floor. Consistent with the no-imputation rule, the gap is disclosed rather than filled. The documentation also records that the causal link from child marriage to forced labour is under review rather than settled.

Structural disruption. Shocks displace people, break household economies, and hand recruiters a supply of people with nothing to fall back on. Disaster-affected intensity comes from EM-DAT, maintained by CRED at UCLouvain, measured over the five years 2020 to 2024, with heat-wave entries excluded on the documented ground that heat mortality is recorded only where a country runs attribution studies, so it tracks measurement capacity rather than shock. Climate vulnerability comes from the ND-GAIN Country Index, using only the vulnerability axis; ND-GAIN's separate readiness axis is left out to avoid re-importing governance. Conflict intensity, in deaths per 100,000 over the five years 2019 to 2023, comes from the Uppsala Conflict Data Program's Georeferenced Event Dataset. Internally displaced persons per 100,000 and refugees originating from the country per 100,000 both come from UNHCR. The last of these also appears in constrained mobility, and the documentation flags this shared use as a possible double-count awaiting a correlation check.

Phase E, Exploitation, asks whether exploitation can run unchecked, and holds three domains.

Foreclosed exit. This domain is meant to capture the cost of walking away: whether a worker in a bad situation has any realistic alternative. Its three indicators are all from ILOSTAT and all protective, entered inverted: labour inspectors per ten thousand employed persons, the share of employees covered by collective agreements, and trade union density. The project is candid that this domain does not measure what it was designed to measure. The intended mechanism, the cost of exit and the degree to which a single employer dominates a local labour market, cannot be sourced across nearly 200 countries. What is available are protective factors whose absence is being read as risk, which is not the same thing. Coverage runs from about 43 to 69 per cent. The domain is carried as insufficient data and is explicitly labelled a stand-in; whether it should be scored at all, folded into a neighbouring domain, or held aside as a diagnostic, remains an open question. Because Exploitation has only three domains, this weakness matters more than it would in Recruitment.

Economic structure and demand. Some economies simply have more of the work in which forced labour occurs. Three indicators: hazardous-sector share, using employment in agriculture, anchored from 0 to 60 per cent; informal employment share, from ILOSTAT; and export concentration, from UNCTADSTAT, a measure of how concentrated a country's merchandise exports are in a small number of products, used here as a proxy for concentrated buyer power. Two caveats are recorded. The export concentration measure is a low-confidence proxy for the buyer concentration the designers actually wanted. And the criminal-market side of demand has no sourced indicator at all, because the obvious candidate, organised-crime outcome measures, was excluded as circular. Agriculture and informality also appear in Recruitment, a cross-phase overlap that the documentation names as an unresolved issue.

State production of unfreedom. The proposition is that states can generate unfreedom directly, through the legal architecture they impose and the impunity they permit. The published build scores it from clientelism, from V-Dem, capturing the exchange of goods and favours for political support; bribery risk, from the TRACE International Bribery Risk Matrix; firm bribery incidence, from the World Bank, being the share of firms reporting at least one request for a bribe; the forced-labour share of detected trafficking victims, derived from the UNODC Global Report on Trafficking in Persons, covering 142 countries; and the DEMSCORE tied-status coding together with the eight-country kafala coding. Two things must be said about this domain. First, only two of five designed drivers are sourced, which is why it is carried as insufficient data; the tied-status, immigration-architecture, and protective-floor legal codings that the design calls for do not exist as global series. Second, the domain is where the index comes closest to being a corruption measure, which is exactly why it receives the special treatment described in the next section.

The Monetization lens, computed but excluded. A third phase covers the financial conditions under which the proceeds of exploitation can be moved, hidden, and retained. It has two domains. Transnational concealment uses the FATF mutual-evaluation effectiveness score and grey-list or black-list standing, both taken from the Basel AML Index published by the Basel Institute on Governance, together with the Tax Justice Network Financial Secrecy Index. Cash and informal retention uses the World Bank Informal Economy Database estimate of the shadow economy as a share of official GDP, and the share of adults without a financial account. This phase is computed, published, and deliberately kept out of the headline score. The reason given is that it answers an intervention question, namely where the money could be disrupted, rather than a structural risk question, and that it scores high for wealthy, financially opaque economies in a way that would distort the risk reading. The project reports having tested whether any governance-independent part of the lens deserves inclusion, and found that the remainder carried no forced-labour-specific signal and diluted the composite. The lens therefore remains display-only. The project frames this distinction as one between drivers and disruptors: Recruitment and Exploitation generate risk, while Monetization is where the cycle could be broken without first removing the underlying vulnerability.

07

Guarding Against Circularity and the Governance Problem

Circularity. An index that predicts forced labour using a measure of forced labour tells you nothing. The project applies this discipline in several places. The United States Trafficking in Persons report, organised-crime outcome measures, and the Basel AML composite were excluded as circular. Detection counts of trafficking victims were refused as a validation benchmark on the ground that detection reflects state capacity rather than prevalence, so matching it would reward strong states. In the published build, the V-Dem indicator measuring freedom from forced labour was dropped entirely from the scoring, on the ground that using a freedom-from-forced-labour measure to predict forced-labour risk is circular by construction.

The discipline is not applied perfectly, and a careful reader should notice the tension. Detection data is refused as a benchmark for checking the index, yet one derived detection measure, the forced-labour share of detected trafficking victims from the UNODC reporting, does enter the scoring of the state production of unfreedom domain. The share of detected cases is not the same quantity as the count of detected cases, and it is less directly a function of state capacity, so the two positions can be reconciled. But the reconciliation is not spelled out in the documentation, and it is the one place where the project's own circularity rule is applied less strictly to an input than to a benchmark.

The governance problem. Most cross-national indicators correlate with the quality of a country's institutions. An index assembled from such indicators can easily become a rule-of-law ranking with a new name, which would make it uninformative for the specific question at hand.

The published response is called de-biasing. Each indicator was tested for how much of its variation is explained by a standard rule-of-law measure. Any indicator explained at or above 55 per cent was treated as too entangled with governance and entered at half weight, applied within the state production of unfreedom domain, which is where such indicators concentrate. This is a targeted reduction rather than a removal.

The result is reported honestly rather than presented as solved. After de-biasing, rule of law still explains about 63 per cent of the variation in the composite, corresponding to a correlation of about 0.79. The project's position is that this residual is correct rather than an artefact: weak governance is a genuine structural driver of forced-labour risk, so a substantial but bounded association is what one should expect. Roughly a third of the variation is not governance.

That position was stress-tested. The de-biasing was re-run varying both its scope, applying it to the one domain or to all domains, and the reference measure used to flag entangled indicators, using the World Bank rule-of-law measure, the V-Dem rule-of-law measure, or an equal blend of the two. Across every combination, rank agreement with the published index stayed at or above 0.96 on Kendall's measure of rank correlation, the governance share stayed near 63 per cent, the top five and bottom five countries were identical, and the result for the Gulf states was unchanged. The check can be re-run by anyone with the repository. This is the strongest available answer to the objection that the index is governance with a new label, and it is a more thorough answer than most composite indices provide.

One nuance is worth noting for readers who look at the code. An older and simpler mechanism, in which a domain score was reduced in proportion to a governance dial, survives in a secondary reproduction build that the repository also ships. It is not the mechanism behind the published figures. The documentation distinguishes the two carefully, and also notes that this secondary build is more governance-dominated than the published one.

08

Reading the Results

The scale. Scores run from 0 to 1. In the published build the scored range is narrower than the theoretical one, running from about 0.11 to about 0.67. The highest-scoring countries are Yemen, Chad, Afghanistan, South Sudan, and Sudan; the lowest are the Nordic countries, Iceland, and several Western European states. This ordering is a face-validity check, not a discovery.

Read tiers, not ranks. The index is banded into three tiers using cuts at 0.281 and 0.402, giving 59 countries in the higher tier, 60 in the middle, and 65 in the lower tier. The cut points were fixed at the registration stage and deliberately not re-derived once the results were known, on the principle that banding thresholds must not be tuned after seeing the output. The project notes that the current build's own thirds would fall at 0.274 and 0.397, and that nine boundary countries would change tier if the cuts were re-derived.

Uncertainty bands. Every scored country carries a rank band produced by running the whole scoring 10,000 times with small random disturbances added to its two phase scores and with the balance between the phases redrawn each time between 40 and 60 per cent. The band reported is the range within which the country's rank fell in 90 per cent of those runs.

These bands are the single most useful discipline the index publishes. The median band is 45 ranks wide, and only about 2 per cent of countries have a band narrower than 10 ranks. In plain terms, a mid-table country's exact position is close to meaningless, while the extremes are stable: the top decile retains about 79 per cent of its members across the simulations, and a country stays in its published tier in about 84 per cent of runs on average.

The low-confidence badge. Forty scored countries have two or more of the eleven domains unscored. These carry a visible lower-confidence marker on the site, and the same rule widens their uncertainty band in the simulation. One rule drives both statements, which prevents the presentation and the arithmetic from drifting apart. The uncertainty code notes that an earlier version of this rule was applied incorrectly, silently widening the band for every country, and records the date the error was corrected.

What a score is not. A high score does not mean a specified number of people are exploited. A low score does not certify a country as free of forced labour. A difference of a few places in the middle of the table is not a finding.

09

Beyond the National Average

A national score averages an entire country into one number, and that average hides the regions where risk actually concentrates. The project therefore publishes a sub-national layer covering first-level administrative regions, built from census microdata from IPUMS-International, an archive of harmonised census samples from around the world, with disaster exposure taken from the Geocoded Disasters dataset.

The published surface covers 1,735 first-level administrative units, of which 1,454 pass reliability filtering and carry a risk score. Roughly 17 per cent of the variation in the underlying precarity measure sits within countries rather than between them, which is the quantitative form of the argument that national averages conceal a great deal.

The project also publishes a corridor view, on the reasoning that conditions enabling forced labour frequently span neighbouring states and that a response confined to one country addresses only part of a cross-border problem.

These layers are best read as illustrative rather than authoritative. They rest on a restricted-access data source that cannot be redistributed with the repository, and they cover fewer countries than the national index.

10

Validation

The project registered its validation criteria in advance, including the thresholds at which each test would be judged to have failed, and then re-ran them on the build that the site actually displays. Pre-registration matters here: it prevents the criteria from being adjusted after the results are known.

The reported outcomes on the displayed build are as follows. Rule of law explains 0.628 of the composite's variation, passing the pre-set requirement that it stay at or below 0.80. The two phases correlate at 0.659, which is within the pre-set window of 0.30 to 0.90 and supports the claim that they are related but not redundant. A forced-labour-specific signal, child labour prevalence drawn from an independent source, remains significantly associated with the composite once governance is held constant on both sides. The top decile retains 86.5 per cent of its members under the suite's mild-noise test, against a requirement of 70 per cent. The published uncertainty model, which disturbs the inputs more aggressively, gives a lower figure of about 79 per cent for the same quantity; both are reported.

One criterion was not met, and the project reports it rather than dropping it. Tested against an external prevalence estimate, the Walk Free Global Slavery Index country prevalence figures, with governance netted out on both sides, no significant association was demonstrated. The project's own reading is that this null result is uninformative rather than disconfirming, because the benchmark is itself heavily entangled with governance. That is a defensible reading, but it is also an admission that the index has no clean external check.

The absence of a governance-independent measure of forced labour prevalence is precisely the gap that motivates a structural index, and it is also the reason the index cannot be fully validated. Readers should hold both facts at once.

A note on what was validated. The tests examine the structure: whether the index is distinguishable from governance, whether its two halves are distinct, whether it carries forced-labour-specific information, and whether its extremes are stable. They do not validate any individual country's score.

11

Limitations and Caveats

The project's credibility rests substantially on the candour of this list, which is reproduced here in the terms the repository itself uses.

It measures conditions, not cases. The index does not estimate prevalence and cannot be read as a count of victims.

It reads origin-side risk and under-reads destination systems. This is the most consequential limitation. The mechanisms that drive forced labour in destination economies, namely sponsorship systems that tie a worker's legal status to one employer, recruitment-fee debt, and migration brokerage, are named in the framework but are not sourced at country scale, beyond the narrow hand-coded signal covering eight states. The available indicators describe resident citizens rather than the migrant workforce most exposed. As a direct result, the United Arab Emirates ranks 146th of 184 scored countries and the other Gulf states sit near it, despite well-documented risk in their labour systems. A low score for a known destination country means that this index does not yet capture that pathway, not that the country is clear. The project mitigates this by attaching a caveat banner to the eight sponsorship-system country profiles, which makes the blind spot visible where a reader would otherwise be misled.

It is correlated with weak governance by design. About two-thirds of the variation is shared with rule of law. The relationship is disclosed and stress-tested rather than engineered away, but a reader who wants a measure independent of institutional quality will not find one here.

The Exploitation phase is thin. One of its three domains does not measure its intended mechanism at all and is carried as insufficient data. Another leans partly on corruption proxies. Because the phase has only three domains, each carries roughly 2.7 times the marginal weight of a Recruitment domain, so these weaknesses propagate further than they would elsewhere.

Some indicators do double duty. Informality and agricultural sector share enter both phases, and refugee outflow enters two Recruitment domains. The collinearity screen that would resolve this is flagged in the rules but has not been run and reported.

Equal weighting is a substantive claim. It has not been tested by perturbation at the indicator and domain level, only at the phase level.

Several anchors were derived from the observed distribution. They were not derived from theory, which weakens the claim that the scale is absolute. An anchor-shift sensitivity test is outstanding.

Coverage is uneven and the gaps are not random. The child labour indicator, which measures the mechanism most directly, covers under half the countries and draws on survey vintages up to two decades old. The unscored countries are systematically small states. Thin coverage tends to depress rather than inflate a score.

Some inputs are licence-restricted. Several sources cannot be redistributed with the repository, and several others carry re-publication flags that the documentation itself marks as unconfirmed, including the Henley Passport Index, the TRACE Bribery Risk Matrix, the UNDP Gender Inequality Index republication, and the Tax Justice Network Financial Secrecy Index. EM-DAT and IPUMS-International are recorded as redistribution-restricted and are not bundled. Anyone republishing this material should resolve those questions first.

The documentation does not describe the published scoring quite completely. The generated codebook lists the indicators drawn from the main data pipeline and states that it cannot drift from that pipeline, but the published scorer additionally enters three inputs held outside it: the UNODC detected-victim composition measure, the DEMSCORE tied-status coding for 29 countries, and the kafala tied-status coding for eight countries. These are listed on the site's own indicator data and are therefore visible to a reader who looks, but they are absent from the codebook, and the narrower coverage of two of them is not reflected in the domain confidence flags. A future revision should bring the codebook and the published scorer back into alignment.

The mid-table is not readable as a ranking. With a median uncertainty band of 45 ranks, ordinal comparisons between similarly placed countries are not supportable.

There is no clean external benchmark. The index cannot be checked against a governance-independent measure of forced labour prevalence, because none exists.

It is a research prototype. It was produced as masters research, it is published as a framework and a pipeline rather than as authoritative country figures, and it should not be used as the sole basis for any consequential decision about a country, a supplier, or a person.

12

What Is Published, and How It Can Be Checked

The deliverable is an interactive website that runs in an ordinary browser and can be served from any static host. It carries a world map, a sortable ranking of all 184 scored countries with the unscored cases shown openly, individual country profiles with a phase and domain breakdown, the full indicator and source list, a limitations page, and a simulation page.

The simulation page lets a reader change the balance between the two phases, switch the combining rule from a geometric mean to a plain average, and move a country's domain scores, watching the map and rankings recompute. Its own description is careful: this shows how sensitive the index is to the choices made in building it, and is not a forecast of anything, nor a prediction of what any intervention would achieve.

Every figure shown on the site is read from the published build output rather than entered by hand, and the whole build can be regenerated with a single command from the inputs stored in the repository. The rebuild verifies its own output against the published baseline and stops if anything has drifted unexpectedly. Roughly two-thirds of the data sources can be re-pulled automatically from public interfaces; the remainder require a human to obtain a registration-gated or licence-gated file first, and each of these is listed with its provider, its address, and its licence terms.

The per-indicator source, vintage, coverage, licence, and required citation are recorded for every signal. The codebook that documents which indicators sit in which domain is generated from the code itself, so it cannot drift away from the pipeline it describes. These are small pieces of discipline, and they are the reason the rest of this report could be written from the repository alone.

13

Intended Audience and Use

The index is addressed to those working on prevention rather than prosecution: labour ministries and inspectorates, humanitarian and development programmers, procurement and due-diligence teams carrying responsibilities under the UN Guiding Principles, and researchers.

Its stated value is as a triage instrument. It points to where vulnerability and unchecked exploitation hold together, consistently across countries that are otherwise difficult to compare, so that attention can be directed before harm becomes visible. It is a starting point for inquiry, not a verdict on any country.

The project is explicit about the boundary of the intervention material it publishes. Identifying where the structure is sensitive is not the same as predicting what an intervention would achieve. The step from noticing a lever to knowing that pulling it works in a particular place belongs to evaluation on the ground.

Three disciplines govern responsible use: read the structure rather than a body count; read inside the country rather than only the national figure, where the sub-national layer allows it; and read tiers rather than ranks.

14

Conclusion

FLSRI addresses a real and well-recognised problem: the places where forced labour is most likely are frequently the places where it is least well recorded, which leaves prevention work waiting for evidence that arrives late or not at all. By measuring conditions rather than cases, the index offers a way to prioritise attention without pretending to count what cannot yet be counted.

Its most substantial contribution is conceptual and procedural rather than numerical. The insistence that a country scores high only where both an exposed population and an unchecked environment are present is a genuine structural claim, encoded in the arithmetic rather than asserted in prose. The refusal to convert missing data into zero, the exclusion of inputs that would predict forced labour from a measure of forced labour, and the decision to publish uncertainty bands wide enough to embarrass the ranking are all choices that make the index harder to over-read.

The honest summary of its current state is that its structure is defensible and its individual country figures are not yet findings. Its Exploitation half rests on one domain that does not measure what it was designed to measure. It under-reads the sponsorship systems that drive some of the most thoroughly documented forced labour in the world, and it says so in the same breath as it publishes the ranks that reflect that gap. It shares a great deal of its variation with measures of institutional quality, and it demonstrates the bound rather than denying the overlap.

What makes the work useful is that these limits are not concessions extracted from it but statements it makes about itself, in its documentation, in its code comments, and on the face of the published site. An index that tells its reader where not to trust it is more usable than one that does not, and that is the standard by which this prototype should be judged and, in time, improved.

References

Sources

  1. 01International Labour Organization. Forced Labour Convention, 1930 (No. 29), Article 2(1); Abolition of Forced Labour Convention, 1957 (No. 105); and the 2014 Protocol to Convention No. 29.
  2. 02International Labour Organization. ILO Indicators of Forced Labour, first issued 2012, revised edition 2025.
  3. 03United Nations. Guiding Principles on Business and Human Rights, endorsed by the UN Human Rights Council in 2011.
  4. 04United Nations. Sustainable Development Goal Target 8.7, with SDG indicators 8.3.1 and 16.9.1 used directly by the index.
  5. 05World Bank. World Development Indicators, including poverty headcount, agricultural employment, GDP growth volatility, the Gini index, birth registration completeness, and firm bribery incidence.
  6. 06International Labour Organization. ILOSTAT, including informal employment share, labour inspector density, collective bargaining coverage, trade union density, and sex-by-sector employment.
  7. 07Varieties of Democracy project. V-Dem dataset, freedom of movement and clientelism indicators.
  8. 08UNHCR. Refugee population statistics, covering refugees originating from a country, internally displaced persons, and statelessness.
  9. 09Uppsala Conflict Data Program. Georeferenced Event Dataset, conflict deaths 2019 to 2023.
  10. 10CRED, UCLouvain. EM-DAT, the international disaster database, disaster-affected intensity 2020 to 2024.
  11. 11World Bank. Global Findex Database, financial account access and borrowing behaviour.
  12. 12Henley and Partners. Henley Passport Index, visa-free destination access.
  13. 13ETH Zurich. Ethnic Power Relations dataset, politically excluded ethnic groups.
  14. 14United Nations Development Programme. Gender Inequality Index, Human Development Report.
  15. 15World Bank. Women, Business and the Law, mobility component.
  16. 16Notre Dame Global Adaptation Initiative. ND-GAIN Country Index, vulnerability axis.
  17. 17UNCTAD. UNCTADSTAT, merchandise export concentration.
  18. 18TRACE International. Bribery Risk Matrix.
  19. 19UNODC. Global Report on Trafficking in Persons, forced-labour share of detected victims, 142 countries.
  20. 20DEMSCORE. Legal-status data used for the tied-status coding covering 29 countries.
  21. 21Basel Institute on Governance. Basel AML Index, FATF effectiveness scores and list standing, used in the Monetization lens only.
  22. 22Tax Justice Network. Financial Secrecy Index, used in the Monetization lens only.
  23. 23World Bank. Informal Economy Database, shadow economy as a share of official GDP.
  24. 24IPUMS-International. Harmonised census microdata behind the sub-national layer, with the Geocoded Disasters dataset for sub-national disaster exposure.
  25. 25Walk Free. Global Slavery Index, used as the external prevalence benchmark and for vulnerability dimension alignment.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file diff --git a/static-site/print/forced-labor-structural-risk-index/index.txt b/static-site/print/forced-labor-structural-risk-index/index.txt index 6873207ad..a97993870 100644 --- a/static-site/print/forced-labor-structural-risk-index/index.txt +++ b/static-site/print/forced-labor-structural-risk-index/index.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -12:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","print","forced-labor-structural-risk-index",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","forced-labor-structural-risk-index","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -14:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -16:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","print","forced-labor-structural-risk-index",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","forced-labor-structural-risk-index","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:"$Sreact.suspense" -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1c:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -29,7 +29,7 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 19:[] 10:"$W19" 11:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1e:Ta14, @page { size: 210mm 297mm; margin: 20mm 18mm; } @@ -180,6 +180,6 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 61:["$","li","23",{"children":[["$","span",null,{"className":"print-ref-num","children":"24"}],["$","span",null,{"children":"IPUMS-International. Harmonised census microdata behind the sub-national layer, with the Geocoded Disasters dataset for sub-national disaster exposure."}]]}] 62:["$","li","24",{"children":[["$","span",null,{"className":"print-ref-num","children":"25"}],["$","span",null,{"children":"Walk Free. Global Slavery Index, used as the external prevalence benchmark and for vulnerability dimension alignment."}]]}] 1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -67:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +67:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 18:null 1d:[["$","title","0",{"children":"The Forced Labor Structural Risk Index (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L67","6",{}]] diff --git a/static-site/print/haste/__next._full.txt b/static-site/print/haste/__next._full.txt index 9bc15be3c..8b63e58a9 100644 --- a/static-site/print/haste/__next._full.txt +++ b/static-site/print/haste/__next._full.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -12:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","print","haste",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","haste","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -14:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -16:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","print","haste",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","haste","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:"$Sreact.suspense" -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1c:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -29,7 +29,7 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 19:[] 10:"$W19" 11:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1e:Ta14, @page { size: 210mm 297mm; margin: 20mm 18mm; } @@ -185,6 +185,6 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 65:["$","div","6",{"children":["$undefined",["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"Archival imagery. HASTE assesses whatever imagery it is given, and it holds none. A retrospective study needs cloud-free 2022 coverage of Mariupol on the specific dates that would fill the gaps, which is a sourcing problem before it is a modelling one."}],["$","li","1",{"className":"pl-1","children":"Building outlines. The platform can only attach a prediction to a building it has an outline for, and section 10 records that outline coverage is weakest in exactly the places under greatest pressure. Outlines derived before the siege will also disagree with imagery of a city whose ground and structures have moved."}],["$","li","2",{"className":"pl-1","children":"The event type. HASTE is benchmarked on xBD, a dataset of disaster damage. Shelling and airstrike damage is not what the reported figures were measured on, and the platform's own design assumption is that a model fitted to one kind of event is not expected to transfer to another."}],["$","li","3",{"className":"pl-1","children":"Human validation. Every HASTE result depends on an analyst marking training examples and judging an independent validation sample. For a retrospective conflict case that means someone competent to read 2022 imagery of this city, and their judgment, not the model's, sets the ceiling on what the number is worth."}],["$","li","4",{"className":"pl-1","children":"Status of the output. HASTE describes its outputs as preliminary signals requiring expert validation rather than authoritative damage assessments. The Mariupol model already declares its damage component a lower bound. A HASTE-derived series would have to carry both caveats, not shed them by being newer."}]]}]]}] 66:["$","p","7",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"The risk the pairing creates."}]," ","A denser series looks better evidenced whether or not it is. Replacing five anchors with, say, twenty assessed dates would produce a damage curve that appears to respond to events day by day, and a reader would reasonably infer that the model now knows more about 18 March than it did. It would know more only if each new point were validated to the standard the five UNOSAT anchors carry. Both projects state their uncertainty explicitly and neither should be allowed to launder the other's: HASTE's confidence interval accounts for sampling error and nothing else, and the severity model's daily resolution is a property of its output rather than of its evidence. If HASTE assessments do enter the model, they will be labelled as such, dated, and reported alongside the UNOSAT figures rather than blended into them."]}] 1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -67:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +67:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 18:null 1d:[["$","title","0",{"children":"HASTE: High-speed Assessment and Satellite Tracking for Emergencies (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L67","6",{}]] diff --git a/static-site/print/haste/__next._head.txt b/static-site/print/haste/__next._head.txt index 24d8fb3ab..414cd05d4 100644 --- a/static-site/print/haste/__next._head.txt +++ b/static-site/print/haste/__next._head.txt @@ -1,6 +1,6 @@ 1:"$Sreact.fragment" -2:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -3:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +2:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +3:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 4:"$Sreact.suspense" -5:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +5:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 0:{"rsc":["$","$1","h",{"children":[null,["$","$L2",null,{"children":[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]]}],["$","div",null,{"hidden":true,"children":["$","$L3",null,{"children":["$","$4",null,{"name":"Next.Metadata","children":[["$","title","0",{"children":"HASTE: High-speed Assessment and Satellite Tracking for Emergencies (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L5","6",{}]]}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/print/haste/__next._index.txt b/static-site/print/haste/__next._index.txt index f98a4fd90..18b9d82b6 100644 --- a/static-site/print/haste/__next._index.txt +++ b/static-site/print/haste/__next._index.txt @@ -1,13 +1,13 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -d:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +d:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} 8:["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$Ld",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$Ld",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$Ld",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$Ld",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] diff --git a/static-site/print/haste/__next._tree.txt b/static-site/print/haste/__next._tree.txt index b7524d296..b855a4855 100644 --- a/static-site/print/haste/__next._tree.txt +++ b/static-site/print/haste/__next._tree.txt @@ -1,7 +1,7 @@ -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] 0:{"tree":{"name":"","param":null,"prefetchHints":16,"slots":{"children":{"name":"print","param":null,"prefetchHints":0,"slots":{"children":{"name":"slug","param":{"type":"d","key":"haste","siblings":null},"prefetchHints":0,"slots":{"children":{"name":"__PAGE__","param":null,"prefetchHints":0,"slots":null}}}}}}},"staleTime":300,"buildId":"static-snapshot"} diff --git a/static-site/print/haste/__next.print.txt b/static-site/print/haste/__next.print.txt index 460afe242..fa617e997 100644 --- a/static-site/print/haste/__next.print.txt +++ b/static-site/print/haste/__next.print.txt @@ -1,4 +1,4 @@ 1:"$Sreact.fragment" -2:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -3:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +2:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +3:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] 0:{"rsc":["$","$1","c",{"children":[null,[["$","style",null,{"children":"\n /* The site is dark-first, and a dark color-scheme makes Chrome paint\n the printed sheet's margin area dark too. Paper is light. */\n html { color-scheme: light !important; }\n html, body { background: #ffffff !important; }\n body::after { content: none !important; }\n header, footer, .site-bg, .aura { display: none !important; }\n "}],["$","$L2",null,{"parallelRouterKey":"children","template":["$","$L3",null,{}]}]]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/print/haste/index.html b/static-site/print/haste/index.html index 1dfec4d41..25b1a4128 100644 --- a/static-site/print/haste/index.html +++ b/static-site/print/haste/index.html @@ -1,4 +1,4 @@ -HASTE: High-speed Assessment and Satellite Tracking for Emergencies (print edition) · NYU Ethical Tech CoLab

Publications · Academic report

HASTE: High-speed Assessment and Satellite Tracking for Emergencies

Rapid Post-Disaster Building Damage Assessment

Software by the Microsoft AI for Good Lab · Report by Ethical Tech CoLab

NYU Center for Global Affairs

July 2026

HASTE was developed by the Microsoft AI for Good Lab and released as open-source software under the MIT Licence. The contributors recorded in the repository's commit history are Meygha Machado, Caleb Robinson, Joaquín Rivero, Cameron Birge, Marcelo Duarte, and Anthony Cintron Roman. The accompanying research paper additionally credits Anthony Ortiz, Simone Fobi Nsutezo, Kevin White, Inbal Becker-Reshef, and Juan M. Lavista Ferres. This plain-language report was prepared under the Ethical Tech CoLab at the NYU Center for Global Affairs as part of masters research (2026), on the fork of the project held at Ethical-Tech-CoLab/haste.

HASTE is an open-source platform that lets a trained analyst, working without code, fit a disposable machine-learning model to a single disaster and produce a building-by-building damage estimate within hours. This report reads the platform's own source code alongside its documentation to set out what every adjustable and hardcoded setting means, what the published performance figures do and do not establish, and where the answer it produces is constrained by things the software does not control.

Key figures

31

field deployments since early 2023, developer-reported (§08)

3

labelled buildings, across at least two categories, before the in-browser classifier will train

0.1

damage threshold, one tenth of a building's visible area, the most consequential single number in the platform (§05)

0.84

area under the ROC curve at one per cent of labels, against 0.88 fully supervised (§08)

00

Foreword

In the hours after an earthquake, a hurricane, or a wildfire, the single most useful thing a relief coordinator can hold is a map of which buildings are still standing. Where should search teams go first? Which neighbourhoods need shelter, and how many people are likely to need it? Which roads lead to places that no longer exist? Those questions have to be answered before anyone on the ground has been able to walk the affected area, and often before the affected area is safe to walk at all.

The traditional answers are slow. Teams fly helicopters over the damage, photograph it, and interpret the photographs by hand. Official mapping services task a satellite, wait for a cloud-free pass, and publish a product days later. Both are careful and both are valuable, but neither reliably delivers inside the first days, which is the window in which decisions about people, supplies, and attention are actually being made.

HASTE, which stands for High-speed Assessment and Satellite Tracking for Emergencies, is an open-source research platform built to shorten that gap. It lets a trained analyst who is not a machine-learning engineer take fresh satellite or aerial imagery of a disaster zone, mark a small number of examples by hand, and have a computer extend those examples across the whole affected area to produce a building-by-building damage estimate. This report explains, in non-technical language, what HASTE does, how it does it, what each of its settings means, and what it cannot be trusted to do.

01

Executive Summary

HASTE is a web-based platform that turns post-disaster satellite imagery into an estimate of which individual buildings have been damaged. It was developed by the Microsoft AI for Good Lab and released as open-source software under the MIT licence. The version reviewed here is the copy held in the Ethical Tech CoLab repository, which is a fork of the original project at microsoft/haste.

The platform's central design choice is to train a fresh model for each disaster rather than maintain one global model that tries to recognise damage everywhere. A hurricane in Jamaica and an earthquake in Türkiye leave visually different traces on visually different building stock, and a model tuned to one is not expected to work on the other. HASTE accepts that limitation deliberately in exchange for speed: a model fitted to one event, from one analyst's labels, can be ready in minutes.

A human being is required at every stage, and there is no automatic mode. The project documentation states repeatedly that outputs are preliminary signals requiring expert validation rather than authoritative damage assessments. Section 09 sets out where the human sits in the workflow.

HASTE offers two routes from imagery to an answer. The faster route, Rapid Building Assessment, computes a numerical fingerprint for every building in the area, asks the analyst to label a handful of them, and trains a very small classifier inside the web browser, which scores all the rest in seconds. The slower route, Damage Mapping, asks the analyst to draw damaged and undamaged areas by hand and trains a full image-segmentation model on a graphics processor, producing a continuous, pixel-level damage map.

Both routes end at the same place: a per-building damage figure, a set of accuracy measures computed against a human-labelled validation sample, and an estimate of the total number of damaged buildings with a stated margin of error.

The platform depends on outside data it does not produce. Imagery comes from commercial and public providers such as Planet, Maxar, Airbus, the European Union's Copernicus programme, and the United States National Oceanic and Atmospheric Administration. Building outlines come from the Overture Maps Foundation. Where those outlines are missing or wrong, HASTE has nothing to attach its predictions to.

According to the research paper published alongside the platform, HASTE has been used in thirty-one field deployments since early 2023, and its outputs have been released openly through the Humanitarian Data Exchange.

02

Background and Rationale

The problem. Rapid damage assessment is a timing problem before it is a technical one. Imagery of a disaster zone often becomes available within a day. An interpretation of that imagery detailed enough to direct a response team to a particular neighbourhood usually does not. The interval between the two is where humanitarian decisions are made with the least information and the greatest consequences.

The gap. Established products have real strengths and known constraints. Copernicus Emergency Management Service Rapid Mapping, the European Union's free on-demand crisis mapping service, delivers standardised map products within hours to days of an activation, but must be formally activated by an authorised user and covers only large-scale emergencies. Manual aerial surveys are accurate but limited in geographic reach and slow to process. Neither easily absorbs the specific situational context that a particular responding organisation cares about, such as one parish, one road corridor, or one category of structure.

The nearer comparator is the automated damage classification literature that grew out of the xView2 challenge, on whose xBD dataset HASTE is itself benchmarked. Against that work the claim is not that nothing existed. It is that existing machine learning approaches assume a globally pretrained model applied to a new disaster, and HASTE trades that for a disposable per event model an analyst fits by hand, in a browser, without writing code.

The response. The proposition HASTE was built around, carried over from earlier in-browser damage-assessment research at the same laboratory, is that human oversight should be structural rather than advisory: the operator is not reviewing a machine's conclusion after the fact, the operator is the source of everything the machine knows about this event.

The design also reflects a practical constraint on who does this work. The people who understand what damage looks like in a given country are rarely the people who can write machine-learning code. HASTE is presented as a no-code platform so that the person supplying the expert judgment and the person operating the model can be the same person.

03

Objectives

The platform is designed to do the following.

  • Produce a building-level damage estimate from post-disaster imagery quickly enough to be useful inside the first days of a response.
  • Allow a non-programmer to fit a model to a specific event using only a map interface, a mouse, and their own visual judgment.
  • Keep a human in control of every consequential step, including the decision to release a result at all.
  • Express uncertainty honestly, by measuring the model against a separate human-labelled sample and reporting a margin of error on the headline damage figure rather than a single confident number.
  • Produce outputs in standard geographic file formats so that they can be opened in the mapping software humanitarian organisations already use, and published openly where appropriate.
  • Remain deployable by others. The platform can be run on a laptop for evaluation or installed on an organisation's own cloud infrastructure, in which case that organisation, and not Microsoft, controls the imagery and the outputs.

04

How HASTE Works

The workflow has a shared beginning, then splits into two routes, then rejoins for validation and reporting.

The shared beginning. An analyst first creates a project, which is simply a container for one disaster event. It records a name, a description, the date of the event, and the affected countries. The date and the countries are stored for reference and do not affect any calculation.

Into that project the analyst adds an image layer, which is the imagery being assessed. It arrives as GeoTIFF, a standard image format that carries the geographic coordinates of every pixel alongside the picture itself. The analyst can upload several files covering the same area, and HASTE merges them into a single mosaic. Imagery from before the event is optional and is used for visual comparison, not for any calculation.

HASTE then obtains the outlines of every building in the area covered by the imagery. By default it downloads them automatically from Overture Maps, an open dataset maintained by the Overture Maps Foundation under the Linux Foundation, which combines OpenStreetMap with machine-derived building datasets from Microsoft and Google and, on the foundation's own account, covers roughly 2.3 billion buildings worldwide. An analyst who has better local data can upload their own outlines instead, as a GeoPackage file of up to 500 megabytes; HASTE converts it to the standard global coordinate system and trims it to the imagery area. When the layer is created the analyst chooses a workflow, Building or Standard, which determines which of the two routes is available.

Route A, Rapid Building Assessment. This route is available when building outlines exist and produces an answer in minutes with no separate training job. First, HASTE computes an embedding for every building. An embedding is a compact list of numbers that describes what the imagery around that building looks like: its texture, its colour, its edges. Two buildings that look alike receive similar lists of numbers. Nothing in the embedding knows anything about damage; it is simply a numerical description of appearance.

The analyst then opens a map, clicks buildings, and assigns each one to one of three categories: Intact, Damaged, or Cloudy, the last meaning that cloud or shadow makes the building impossible to judge. A left click labels a building, a right click removes the label, and holding the control key while dragging labels many at once.

Once at least three buildings have been labelled across at least two categories, a very small statistical model called a logistic regression trains automatically in the browser and immediately predicts a category for every other building in view. The analyst can flip between what they labelled and what the model predicted, see where it is wrong, label a few more examples, and watch the prediction improve. This loop takes seconds, which is what makes the route fast. When satisfied, the analyst runs a final pass that scores every building in the layer and saves the result as a geographic data file.

Route B, Damage Mapping with a trained model. This route does not require building outlines and produces a continuous damage picture across every pixel of the imagery rather than a verdict per building. The analyst draws polygons, rectangles, or circles over a small part of the image and assigns each shape a class. The default classes are Background, Building, and Damaged Building; additional classes of No Damage and Flood Extent are offered for particular event types. The project documentation advises drawing at least five to ten shapes per class, considers seventy to one hundred a good training set, and states that more than roughly one hundred and fifty is unnecessary.

Those shapes train an image-segmentation model, which is a model that assigns a class to every individual pixel rather than to a whole picture. The architecture used is a U-Net with a ResNeXt-50 encoder, a standard and well-understood design in satellite image analysis, started from weights pre-trained on ordinary photographs and then fitted to the analyst's labels. Training runs on a graphics processor, either locally in a container or on cloud computing capacity. The trained model is then run across the entire image layer, producing a per-pixel damage prediction that can be viewed alongside the imagery and downloaded.

Turning pixels into buildings. A per-pixel damage map is not directly useful to a responder, who needs to know about buildings. HASTE therefore overlays the building outlines onto the pixel predictions and, for each building, counts what proportion of the classified pixels inside that outline were called damaged. That proportion, between zero and one, is the building's damage fraction. It is the number that drives everything downstream. The same step also records what proportion of each building was obscured by cloud.

Validation. Both routes end with the same discipline. The analyst opens a validation tool that presents a random sample of roughly two hundred building outlines over the post-event imagery and asks them to judge each one independently as Damaged, Not Damaged, or Unknown. These human judgments are treated as the ground truth against which the model's predictions are scored. Buildings marked Unknown are excluded from the scoring entirely rather than being guessed at.

05

The Variables, Explained Simply

Every number an analyst can adjust, and every number fixed inside the code, is described below in ordinary terms: what it represents, why it is set where it is, and what it changes about the answer.

The label classes. These are the categories an analyst draws with. Each is stored internally as a class value, and later steps identify a class by that position rather than by its name.

ClassStored valueWhat it marks
Background1Ground and anything that is not a building.
Building2An intact structure.
Damaged Building3A structure showing damage. This is the class the damage figures are computed from.
Cloud4Areas where cloud or shadow makes the imagery impossible to read.
The four positional label classes and the values they are stored as.

Two of these choices carry more weight than they appear to. The reason Background exists as a class in its own right is that the model is only taught by the pixels the analyst actually marked, so if a damaged roof is labelled without the ground around it, the model is never told what the ground is, and can produce a blurry, spreading prediction without ever being penalised for it. Labelling a building together with its surroundings is what forces the model to learn a boundary. And cloudy buildings are excluded from the damage statistics rather than counted as intact, which would otherwise bias the result downward in exactly the wet, storm-affected conditions where damage assessment matters most.

No Damage and Flood Extent. Two additional classes offered for earthquake, fire, and flood event types respectively. Flood Extent lets the analyst mark standing water, which HASTE can intersect with building outlines to say which buildings are in water.

Normalisation means and standard deviations. Image pixels are stored as whole numbers, but models learn more stably from values between 0 and 1, so each channel is rescaled by subtracting a mean and dividing by a standard deviation. In the worked example shipped with the platform these are set to 0 and 255, which is a plain rescaling of ordinary 8-bit imagery. When an image layer is prepared automatically, HASTE instead sets every mean to zero and sets each channel's divisor to the brightness value at that channel's 98th percentile, so that the brightest two per cent of pixels are treated as glare and clipped rather than being allowed to compress everything else into a narrow band. This matters because imagery from different sensors arrives on different numeric scales, and a wrong divisor makes the picture unrecognisable to the model.

Ground resolution. HASTE never checks, records, or standardises how many metres of ground each pixel covers. It works on whatever grid the supplied imagery has. Several settings are expressed in pixels rather than metres as a result, and the physical area covered by a training tile therefore varies with the imagery source.

Learning rate, default 0.0001. How large a correction the model makes each time it discovers it was wrong. Too large and it lurches past the right answer; too small and it never gets there within the time available. The default is a conservative value appropriate to fine-tuning a model that already carries useful pre-trained weights.

Batch size, default 32 for training. How many image tiles the model examines before making one correction. Larger batches give a steadier, less noisy signal about which direction to move in, but require proportionally more memory on the graphics processor.

Maximum epochs. One epoch is one complete pass through the training data. A small number is intended here, for a clear reason: the model is being fitted to a few dozen hand-drawn shapes from a single event, and training it longer would mainly teach it to memorise those particular shapes rather than the general appearance of damage. The repository is inconsistent about the actual figure. The worked example specifies 10, the form in the web interface offers 3, and the server falls back to 1 when nothing is supplied, while the training script separately imposes a hardcoded minimum of 10. An analyst who selects 3 in the interface is therefore asking for a maximum below the enforced minimum, which anyone reproducing a result should know.

Training chip size, 256 pixels. The model does not see the whole scene at once. It is shown small square cut-outs 256 pixels on a side, drawn at random from the labelled area, and the training run feeds it 1,024 batches of them per epoch. The model's whole view of the disaster is assembled from these fragments.

Label buffer, nominally 3 metres. Applied to the Building class using the Background class. When a person traces a building, the traced line is never exactly on the wall. This setting draws a narrow band around each building polygon and treats it as Background, which stops the imprecision of human tracing from teaching the model that the pavement is part of the structure. The setting is written in metres but is applied by counting pixels, so it means three metres only when a pixel happens to represent one metre of ground. The script's own documentation notes this.

Patch size, default 2048, and padding, default 64. Used when running the finished model across the full image. Satellite scenes are far too large to process in one piece, so they are cut into square tiles. Predictions are least reliable at the very edge of a tile, where the model can see no context, so each tile is processed with 64 pixels of overlap that are then discarded. This is what stops a visible grid pattern from appearing across the finished damage map.

Initial weights, default none. Meaning the model starts from weights learned on ordinary photographs. An analyst can instead start from a previously trained HASTE model held in the model catalogue, which is offered only where the event type and imagery source match, since a model fitted to wildfire imagery is not a sensible starting point for flood imagery.

The cloud penalty. An optional setting, switched off by default, changes how the model is corrected in areas the analyst marked as Cloud. Rather than being told what class those pixels really are, which nobody knows, the model is penalised whenever it predicts damage beneath a cloud. In plain terms it is taught not to claim to see through weather.

Embedding backbone. The method used to convert the appearance of a building into numbers. The lightweight default is MOSAIKS, an approach published in Nature Communications in 2021 that passes the image through a large set of randomly generated filters rather than learned ones. The counter-intuitive finding of that work is that random filters, applied at sufficient scale, describe an image well enough for a simple linear model to make accurate predictions from them, at a tiny fraction of the computational cost. The alternative offered is DINOv2, a family of vision models released by Meta AI in 2023 that learned general visual features from 142 million unlabelled images, available in small and base sizes.

Output dimensions, default 1024. Applicable to MOSAIKS only. How many numbers describe each building. More numbers capture more nuance and cost more time and memory. When DINOv2 is used this setting is ignored, because the size of the description is fixed by the model itself: 384 numbers for the small version, 768 for the base version, 1024 for the large one.

Kernel size, default 7. The width in pixels of each random filter. Small filters respond to fine texture such as roof material and rubble; larger ones respond to coarser structure. Seven pixels is a middle setting suited to detecting the change in surface texture that collapse produces.

Resize factor, default 4. How much the small image crop around each building is enlarged before being described. In moderate-resolution satellite imagery a single house may span only a handful of pixels, too few for the filters to find anything. Enlarging the crop fourfold gives them something to work with. This is a real trade-off and not a free gain: enlarging an image does not create detail that the sensor never recorded, it only makes the recorded detail large enough to be measured. Choosing DINOv2 forces this setting back to 1, because that model brings its own fixed way of dividing an image into pieces.

Context padding, 8 pixels. Each building crop reaches a little beyond the traced outline, so that the description includes some of the ground immediately around the structure. Debris and scorch marks sit there rather than on the roof.

What is actually described. The crop is divided into a grid of small tiles, each tile is described separately, and only those tiles falling inside the building's outline are averaged together to produce the building's final list of numbers. Buildings that fall outside the imagery keep their place in the file with an empty entry rather than being dropped, because everything downstream matches buildings to predictions by position in the file rather than by name. That design decision is efficient and fragile in equal measure.

The crop cap, 192 source pixels. A limit that stops a single very large building, such as a warehouse or a stadium, from generating an enormous crop and exhausting the available memory. Oversized footprints are cropped from the centre instead.

The in-browser classifier. The model that turns a handful of labelled buildings into predictions for all of them is a logistic regression, one of the oldest and simplest classifiers in statistics. It fits a straight-line boundary through the space of building descriptions and reports, for each building, how far onto the damaged side of that boundary it falls. Its learning rate is 0.1 and it takes 500 steps, values fixed in the code rather than exposed to the analyst, chosen to complete in well under a second so that the label-and-see loop stays interactive.

Regularisation strength, 0.01. A deliberate penalty that discourages the model from placing heavy reliance on any single one of the thousand-odd numbers describing each building. Without it, a model trained on three examples would seize on whatever coincidental feature happened to separate those three, and would fail on the fourth. This single small number is the main defence against overfitting in a workflow explicitly built around very few labels.

Holdout fraction, 0.2 in practice. One fifth of the analyst's labels are held back from training and used only to score the model, which is what produces the live precision, recall, and F1 figures shown for the Damaged class in the side panel. Because those numbers come from labels the model was never shown, they are an honest indication of quality rather than a report of how well the model memorised its own training data. Two weaknesses are worth naming. The sample is tiny: with twenty labels the holdout is four buildings, and four buildings cannot tell you much. And the split is drawn afresh every time a label is added, so the displayed figures jump around from click to click, as the code comments acknowledge.

Measurement rings, 0, 10, and 20 metres. Not to be confused with the label buffer above. HASTE calculates each building's damage fraction three times: once inside the traced outline, once inside a ring extending ten metres beyond it, and once at twenty metres. There are two reasons. Building outlines and satellite imagery are frequently misaligned by several metres, particularly in dense cities and where the ground itself has moved, so the strict outline may sit over the neighbouring plot. And some kinds of damage, notably debris fields and burn scars, appear around a structure rather than on it. The wider measurements are recorded so that this can be examined rather than assumed.

Damage fraction. For a given buffer, the number of pixels classified as Damaged Building divided by the number of classified pixels of any kind inside that shape. Unclassified pixels, recorded as zero, are excluded from both the top and the bottom of that division, so a building half outside the imagery is judged on the half that was actually seen. The result is capped at 1.

Damage threshold, default 0.1. The proportion of damaged pixels above which a building is declared damaged. This is the most consequential single number in the entire platform, and it is set low on purpose. A tenth of a building's visible area showing damage indicators is treated as enough, reflecting a judgment that in the first days of a response, missing a damaged building is a worse error than flagging one that turns out to be intact. It also compensates for partial visibility: a structure photographed from overhead shows mainly its roof, and serious structural damage may present as only a modest patch of altered texture there. Because the threshold is a setting rather than a fixed rule, and because the platform also publishes a full precision-and-recall curve across all possible thresholds, an analyst who disagrees with this trade-off can see exactly what a different choice would cost.

A wrinkle a careful reader should know about. The exported data file also carries a simple yes-or-no damaged column, and that column is set whenever a building contains even one damaged pixel, with no threshold at all. The validation report reads that column; the assessment report applies the tenth-part threshold. The same model and the same human labels will therefore yield slightly different accuracy figures in the two reports. Nothing in the documentation flags this, and it is the kind of discrepancy that could confuse a reader comparing the two.

Minimum footprint area, default 50 square metres. Buildings smaller than this are excluded from the population used for the final headline estimate. The stated reasoning is that these are structures a human reviewer could not realistically have judged in the validation step, so including them in an extrapolation built from human judgments would be unsound. The consequence, which the platform does not hide, is that small informal structures are absent from the headline count, and those structures are common in precisely the settlements most exposed to disaster.

Cloud exclusion. Any building with a cloud fraction above zero is excluded from the damage count entirely. This is a strict rule, not a proportional one; a building need only be slightly obscured to be set aside.

The final estimate. HASTE does not publish a raw count of the buildings its model called damaged. It uses the human-validated sample to estimate the true damage rate and scales that rate up to the full population of buildings. The arithmetic is a standard finite-population survey estimate. The sample proportion, written as p-hat, is the number of buildings a human confirmed as damaged divided by the number of buildings the human judged either way, with Unknown buildings excluded from both figures. The population, written as N, is the count of building outlines larger than the minimum area, and the estimated number of damaged buildings is simply N multiplied by p-hat.

The sampling fraction, written as f. The size of the validated sample divided by the population. It enters the calculation through what statisticians call a finite-population correction, the term (1 minus f) in the variance. Its effect is intuitive: if you have checked half of all the buildings by hand, your uncertainty about the other half should be smaller than if you had checked one per cent of them. Standard survey formulas assume an infinite population and would overstate the uncertainty here.

The critical value z, fixed at 1.959963984540054. This is the constant that turns a standard error into a 95 per cent confidence interval, and it is written out as a literal number in the code specifically so that the platform does not have to load a heavy statistical library to obtain it. The reported interval runs from N times (p-hat minus z times the standard error) to N times (p-hat plus z times the standard error).

What the interval means in plain terms. If the same sampling exercise were repeated many times, an interval calculated this way would contain the true number of damaged buildings in about nineteen cases out of twenty. It accounts for the fact that a sample was taken. It does not account for a mislabelled sample, misaligned building outlines, or a model that is wrong in a consistent direction, and it should not be read as expressing total uncertainty about the figure.

06

Reading the Results

The visualiser. The pre-event and post-event imagery are shown side by side with a swipe control between them, the predicted damage layer overlaid on both, and the raw per-pixel predictions available as an optional toggle. Opacity, contrast, hue, and saturation sliders help make damage visible in imagery that was captured in poor light.

On the map, each building is shaded according to its damage fraction in five bands, cut at one fifth, two fifths, three fifths, and four fifths, running from white through pale peach and orange to red and dark red. This is a presentational scale only. It should not be read as a severity classification in the sense used by structural engineers, because the model was never taught degrees of damage. It only ever learned to separate damaged from intact, and the shading reflects how much of a roof it flagged rather than how badly the building was hurt.

The validation report. This compares the model's predictions against the analyst's own validation labels and reports overall accuracy, precision, recall, and F1 for each class, a macro-averaged F1, and a confusion matrix. Precision answers the question: of the buildings the model called damaged, what share really were? Recall answers the opposite: of the buildings that really were damaged, what share did the model find? These two errors have different humanitarian costs, and the report deliberately does not merge them into a single figure.

The assessment report. This summarises the whole layer: the total number of building outlines, how many were excluded as cloud-covered, how many of the remainder were predicted damaged and what percentage that represents, the accuracy metrics at the chosen threshold, a precision-and-recall curve across all thresholds, and the estimated total of damaged buildings with its 95 per cent confidence interval. Where no validation labels exist, the report still shows the prediction counts but withholds the accuracy metrics and the estimate rather than presenting an unvalidated number.

Downloads. Results can be exported as a GeoPackage, an open standard file that opens in common mapping software, along with the training artefacts and the building outlines used. This matters for accountability: a partner receiving a HASTE layer can inspect it independently rather than taking a summary figure on trust.

07

Data Sources

Imagery. HASTE holds no imagery of its own; the analyst supplies it. Only GeoTIFF files are accepted. The documentation records that in past activations imagery has come from the following providers.

  • Planet, a commercial operator of a large constellation of small satellites that publishes disaster imagery openly.
  • Maxar, now Vantor, a commercial provider of high-resolution imagery with its own open data programme for disasters.
  • The Airbus Foundation.
  • Sentinel-1 and Sentinel-2, the radar and optical satellites of the European Union's Copernicus programme, whose data is free to all.
  • Products from the Copernicus Emergency Management Service.
  • Aerial imagery from the United States National Oceanic and Atmospheric Administration.

The software itself has no live connection to any of these providers. Placeholder functions for fetching imagery directly from Maxar and Planet exist in the code but are empty. In practice the analyst supplies a link to a file, and for security reasons those links may point only at Azure Blob Storage or Amazon S3, the two hosting services on the platform's permitted list. Public disaster imagery from the major providers is generally published on one of those, which is why the restriction is workable.

HASTE does adjust for the provider in one respect. Different satellites record their colour bands in different orders, and some record bands the human eye cannot see, so the platform holds a lookup table of band orders for Planet Scope, Planet Skysat, Maxar, Sentinel-2, and one partner-specific format, and uses it to assemble a correct colour picture. Where the source is unknown, HASTE falls back to the labels embedded in the file, and failing that assumes the first three bands are red, green, and blue.

Building outlines. Overture Maps by default, described in section 04, and analyst-supplied outlines where better local data exists. HASTE reads the Overture data anonymously from a public store, taking the most recent release available and falling back to a fixed February 2026 release if it cannot determine one. Only polygons are kept.

What HASTE does not use. The platform draws on imagery and building outlines and nothing else. It does not read ground reports, weather data, sensor networks, social media, or population figures. Its picture of a disaster is strictly what a camera in orbit could see, which is a narrower thing than what happened.

Personal data. The platform is not designed to identify people, does not ingest or output personal data, and does not treat person-scale features in imagery as signal.

08

Evidence of Performance

Everything in this section is the developer's own account. The benchmark results, the deployment record, and the field precision and recall figures all originate from the Microsoft paper and repository. None of them has been independently reproduced or externally validated, here or elsewhere, so they should be read as what the team that built the platform reports about it rather than as third-party verification. That is not an accusation of overstatement. It is the provenance of the evidence, and it is the first thing a sceptical reader is entitled to know.

Validated accuracy. The research paper published alongside the platform, HASTE: A Platform for Rapid Post-Disaster Building Damage Assessment (arXiv:2607.11838), reports experiments on xBD, a public benchmark dataset of paired pre-event and post-event satellite imagery with expert damage annotations. The team collapsed the benchmark's minor, major, and destroyed categories into a single damaged category and measured how well each embedding method performed as the number of labels was varied.

The discrimination score in Table 1 is the area under the receiver operating characteristic curve, usually shortened to AUROC. In plain terms it is the probability that the model ranks a randomly chosen damaged building above a randomly chosen intact one. It runs from 0 to 1, a coin flip scores 0.5, and 1.0 would mean the model never gets a pair the wrong way round. The measure says nothing about where the threshold should sit: a model can rank buildings well and still mislabel a great many of them once a cutoff is applied.

ApproachLabels usedDiscrimination score
Strongest embedding1 per cent0.84
Strongest embedding10 per cent0.91
Fully supervised ResNet-50All labels0.88
Table 1. Reported performance on the xBD benchmark as the number of labels was varied.

The practical claim is that a handful of labels plus a good general-purpose image description gets close to a conventionally trained model. The headline figure is a modest deficit rather than a match: at one per cent of labels the score is 0.84 against the fully supervised 0.88, and it is at ten per cent, where it reaches 0.91, that the fast route actually overtakes the baseline. What one per cent buys is not equal accuracy but most of the accuracy, hours sooner and without a machine-learning engineer, which is a real trade and a different claim.

The table does not say which of the two embedding methods produced these scores, and this report cannot resolve it from the published material. Since the lightweight option is the default an analyst would run, and the heavier one is the more capable, that gap matters to anyone reading the figures as a prediction of what they will get.

Deployment record. The figures that follow are evidence of adoption rather than of correctness, and only the Rolling Fork assessment carries an accuracy measurement against field ground truth.

The same paper reports thirty-one field deployments since early 2023, including four cities assessed within three days of the February 2023 Türkiye earthquakes, a tornado assessment in Rolling Fork delivered in under two hours at 0.86 precision and 0.80 recall against field ground truth, and the August 2023 Maui wildfire, where imagery available at nine in the morning yielded an assessment by one in the afternoon identifying roughly 1,700 damaged buildings.

For Hurricane Melissa in Jamaica in late 2025, four areas covering about 2,300 square kilometres were assessed. In Black River, some 110,000 building outlines were examined, of which around 65,000 were obscured by cloud.

AreaRecallPrecisionEstimated damaged
Black River96 per cent82 per cent31,000 buildings
Montego Bay86 per cent71 per centNot reported
Table 2. Two of the areas assessed during the Hurricane Melissa response.

The variation between Black River and Montego Bay, on the same event with the same team days apart, is substantial, and the very large share of cloud-obscured buildings in Black River is a reminder that a headline damage estimate can rest on a minority of the buildings actually present.

09

Human Oversight and Governance

There is no autonomous mode. A person selects the imagery, provides every label the model learns from, reviews the predictions, validates a sample, and decides whether to distribute the result.

Outputs distributed publicly, including through the Humanitarian Data Exchange and partner mapping systems, carry notices warning against over-reliance and encouraging cross-validation against ground reports and other imagery. The notices frame the output as exploratory rather than definitive, and name HASTE, the imagery provider, and the building-outline dataset so that a downstream user can assess provenance for themselves.

The documentation states that where a widespread inaccuracy or pattern of misuse is identified, the laboratory may publish guidance recommending temporary suspension or restricted use of the workflow.

Labels are not reused. Each event's labels belong to that event and are not accumulated into a global training set, which follows directly from the train-per-event design and also limits the accumulation of one analyst's interpretive habits across many responses.

10

Limitations and Caveats

The repository documents its own weaknesses at length. The most consequential are these.

Every performance figure in this report is developer-supplied. This one is not in the repository's own list, and it is the limitation a sceptical reader should weigh first. As section 08 sets out, none of the reported results has been independently reproduced. The rest of this report reads the source code and reports what it finds, which is first-hand; the performance evidence does not have that standing, and the two should not be given the same weight.

The output depends heavily on who did the labelling. Two competent analysts working the same event from the same imagery can produce materially different results. The documentation gives a concrete example from the Hurricane Melissa response, where an initial set of 153 labels produced predictions that described buildings as around 20 per cent damaged when they were in fact totally destroyed; the error was caught by visual inspection and corrected by adding a further 107 labels and retraining.

Imagery quality governs everything. Cloud, haze, low light, and imagery captured at an angle rather than from directly overhead all degrade performance, sometimes severely. Planet imagery in particular sits below the roughly 30 centimetre resolution that building-detection models generally prefer, which is part of why the platform leans on external building outlines rather than trying to find buildings itself.

Building-outline coverage is uneven in a way that matters. OpenStreetMap and Microsoft Building Footprints are weakest in the Global South, in informal settlements, in conflict-affected areas, and where construction has been rapid and recent. These are precisely the populations most exposed to disaster. A building with no outline is invisible to HASTE regardless of how badly it was damaged, and the 50 square metre minimum area removes further small structures from the headline figure.

The model does not generalise, by design. A model fitted to one event is not expected to transfer to another, which also means HASTE cannot be operated as a standing monitoring system.

False positives and false negatives have distinct causes. False positives arise from shadows, ordinary construction and demolition unrelated to the disaster, atmospheric artefacts, and vegetation change. False negatives arise from subtle structural damage visible only from an angle, damage hidden by cloud or shadow, and damage at a scale finer than the imagery can resolve.

No contextual data is incorporated. There is no ground truth, no weather, no sensors, no population data. Flood extent is intersected with buildings but water depth is never estimated, so HASTE can say a building is in water and not how deep.

The confidence interval understates real uncertainty. It quantifies the error introduced by sampling and nothing else, as section 05 explains.

Some findings from reading the source code should be added to the project's own list, since they bear on how much weight an output can carry.

There is no genuinely held-out validation data in the trained-model route. The measure used to decide which version of the model to keep is computed on random cut-outs of the same imagery the model was trained on. It is therefore a measure of fit rather than of generalisation. The independent check in HASTE comes later, from the human validation sample, which is the number a reader should rely on.

The meaning of each class is fixed by its position in a list. Damaged Building is understood downstream as the third class and Cloud as the fourth. An analyst who reorders the classes when setting up a project will get damage figures computed from the wrong category, without any error being raised.

Several defaults disagree between the worked example, the web form, and the server. This is most visible for the number of training passes. Two analysts following the documentation by different routes may not be running the same configuration.

Predictions are matched to buildings by position in a file rather than by an identifier. This is fast, and it means that anything which reorders or filters the building list between steps would silently misattribute damage. The developers are evidently aware of the risk, since the code goes to some trouble to preserve row positions even for buildings it cannot assess.

The platform is not validated for production or autonomous use. It has not been designed, tested, or validated for production deployment or autonomous decision-making, and its outputs are not authoritative damage assessments. The documentation states plainly that users should not rely solely on HASTE outputs for decisions affecting safety, property, or human life, and should not use them to trigger public alerts or resource deployments without independent verification.

11

Practical Nature of the Platform

HASTE is a full web application rather than a single page or a script. It consists of a browser interface, a set of programming interfaces that the browser front end calls, background workers that handle long jobs such as preparing imagery and training models, a tile server that streams large satellite images into the map view, and a shared Python library holding the analysis logic.

It can be run in two ways. A complete local instance can be started on one machine using Docker, a tool that packages software with everything it needs to run, which requires no cloud account and is intended for evaluation. A production instance is deployed to Microsoft Azure with a single command, and the deploying organisation controls it entirely.

The local configuration is explicitly flagged as unsuitable for production, since it disables authentication and uses an in-memory storage emulator. A separate hardening checklist is provided for real deployments.

The repository shows active and careful security practice, including automated code scanning and secret scanning, and documented handling of known vulnerabilities in the underlying geospatial libraries where a patched version was not available. Sample data from Hurricane Melissa and the Lahaina wildfire is published so that the platform can be tried without sourcing imagery first.

12

Intended Audience and Use

HASTE is aimed at trained humanitarian and disaster-response practitioners working with post-event imagery, and at researchers studying rapid damage-assessment methods. The documentation names non-governmental organisations, United Nations agencies, and government users as the intended downstream consumers of its outputs.

Its stated purpose is to contribute information to preliminary damage assessment in the first hours and days, supplementing rather than replacing expert assessment, and to indicate where damage may be concentrated so that attention can be directed.

It is explicitly not intended for use as any of the following.

  • An authoritative damage register.
  • Ground truth for insurance, governmental, or public-reporting purposes.
  • The sole basis for search-and-rescue tasking or resource allocation.
  • An automated alerting system of any kind.

13

Application: the Mariupol Corridor Severity Model

This section describes work by the Ethical Tech CoLab rather than by the platform's developers, and it describes work in progress. Nothing below has been published as a result, and no HASTE-derived figure currently enters the model it discusses. It is included because the reasoning about when this platform is and is not the right instrument is more useful stated against a real case than in the abstract.

The gap in our own model. The CoLab's Mariupol Corridor Severity Model is a daily civilian-danger index for the 2022 siege, built from six components. The sixth is infrastructure damage, the cumulative share of the city's structures visibly damaged or destroyed, and it is drawn from satellite damage assessments published by UNOSAT, the United Nations Satellite Centre. UNOSAT publishes on particular dates rather than continuously. Across a siege of seventy-seven days the model has five anchor points: 0.02 on 5 March, 0.04 on 14 March, 0.14 on 26 March, 0.32 on 7 May, and 0.33 on 20 May. Every day in between is read off a straight line drawn between them.

The model is candid about what this costs it. In its own words, the word daily describes the resolution of the output and not the resolution of the evidence, and it cannot distinguish 18 March from 19 March on any evidence about what happened on those two days. The practical danger is specific: a stretch of the siege that UNOSAT did not image is rendered as a gently rising line, so an absence of published assessment can be mistaken for an absence of destruction.

Why HASTE is the plausible instrument. The constraint that limits UNOSAT is not imagery, it is the analytic labour of a formal before-and-after building-by-building assessment. HASTE is built for exactly that constraint. It needs only post-event imagery, it fits a disposable model to one event from a small number of hand-marked examples, and it returns a per-building damage fraction with a stated margin of error. In principle it could produce assessments for dates that lie between the UNOSAT anchors, replacing an interpolated line with measured points and letting the damage component move on the evidence rather than on the arithmetic.

What has actually been done. So far, scoping. The model still runs on the five UNOSAT anchors and the straight lines between them, and the per-building display in the tool remains illustrative until real geodata is supplied. We are working through whether HASTE-style assessment of additional 2022 imagery dates is feasible and, more importantly, whether the result would be honest enough to publish.

What would have to be true first. The obstacles are the ones this report has already described, met in a hard case. They are set out here rather than discovered later.

  • Archival imagery. HASTE assesses whatever imagery it is given, and it holds none. A retrospective study needs cloud-free 2022 coverage of Mariupol on the specific dates that would fill the gaps, which is a sourcing problem before it is a modelling one.
  • Building outlines. The platform can only attach a prediction to a building it has an outline for, and section 10 records that outline coverage is weakest in exactly the places under greatest pressure. Outlines derived before the siege will also disagree with imagery of a city whose ground and structures have moved.
  • The event type. HASTE is benchmarked on xBD, a dataset of disaster damage. Shelling and airstrike damage is not what the reported figures were measured on, and the platform's own design assumption is that a model fitted to one kind of event is not expected to transfer to another.
  • Human validation. Every HASTE result depends on an analyst marking training examples and judging an independent validation sample. For a retrospective conflict case that means someone competent to read 2022 imagery of this city, and their judgment, not the model's, sets the ceiling on what the number is worth.
  • Status of the output. HASTE describes its outputs as preliminary signals requiring expert validation rather than authoritative damage assessments. The Mariupol model already declares its damage component a lower bound. A HASTE-derived series would have to carry both caveats, not shed them by being newer.

The risk the pairing creates. A denser series looks better evidenced whether or not it is. Replacing five anchors with, say, twenty assessed dates would produce a damage curve that appears to respond to events day by day, and a reader would reasonably infer that the model now knows more about 18 March than it did. It would know more only if each new point were validated to the standard the five UNOSAT anchors carry. Both projects state their uncertainty explicitly and neither should be allowed to launder the other's: HASTE's confidence interval accounts for sampling error and nothing else, and the severity model's daily resolution is a property of its output rather than of its evidence. If HASTE assessments do enter the model, they will be labelled as such, dated, and reported alongside the UNOSAT figures rather than blended into them.

14

Conclusion

HASTE's most useful contribution is not a modelling advance but a reallocation of labour. It moves the machine-learning work out of the way so that the scarce resource, the judgment of someone who can look at an image and know what damage looks like in that country, is applied where it counts: to choosing the imagery, marking the examples, and checking the answer. The model is small, disposable, and fitted to one event, and the platform treats it as such.

The reporting design is equally deliberate. By requiring an independent human validation sample before it will publish an accuracy figure, by separating precision from recall rather than averaging them away, and by attaching a confidence interval to its headline number, the platform makes it harder to mistake a fast estimate for a survey.

The honest reading is that HASTE is fast, transparent about its assumptions, and constrained by things it does not control. Its ceiling is set by the resolution of the imagery available and the completeness of the building outlines beneath it, and both of those are weakest in the places where humanitarian need is greatest. That is not a flaw in the software so much as a statement about the wider data landscape, but it means the platform's value in any given response depends on conditions decided long before the disaster occurred. Read alongside its own documented limitations, it is a credible example of a machine-learning tool built to assist expert judgment rather than to displace it.

References

Sources

  1. 01Microsoft AI for Good Lab. HASTE: A Platform for Rapid Post-Disaster Building Damage Assessment. arXiv:2607.11838.
  2. 02Microsoft AI for Good Lab. HASTE, open-source platform released under the MIT Licence.
  3. 03Overture Maps Foundation. Buildings theme, combining OpenStreetMap with machine-derived building datasets from Microsoft and Google.
  4. 04OpenStreetMap contributors. OpenStreetMap building data.
  5. 05Microsoft. Global ML Building Footprints.
  6. 06xBD. Public benchmark dataset of paired pre-event and post-event satellite imagery with expert damage annotations.
  7. 07MOSAIKS. A generalizable and accessible approach to machine learning with global satellite imagery. Nature Communications, 2021.
  8. 08Meta AI. DINOv2, a family of vision models trained on 142 million unlabelled images, 2023.
  9. 09European Commission. Copernicus Emergency Management Service, Rapid Mapping.
  10. 10European Union. Copernicus Sentinel-1 and Sentinel-2 radar and optical satellite data.
  11. 11Planet. Open disaster imagery from a constellation of small satellites.
  12. 12Maxar, now Vantor. Open Data Program for disaster response imagery.
  13. 13United States National Oceanic and Atmospheric Administration. Emergency response aerial imagery.
  14. 14OCHA. Humanitarian Data Exchange, the channel through which HASTE outputs have been released openly.
  15. 15Ethical Tech CoLab. Mariupol Corridor Severity Model, a daily civilian-danger index for the 2022 siege, discussed in section 13.
  16. 16UNOSAT, United Nations Satellite Centre. Ukraine damage assessments, activation CE20220223UKR, the source of the five anchor points in section 13.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file +

Publications · Academic report

HASTE: High-speed Assessment and Satellite Tracking for Emergencies

Rapid Post-Disaster Building Damage Assessment

Software by the Microsoft AI for Good Lab · Report by Ethical Tech CoLab

NYU Center for Global Affairs

July 2026

HASTE was developed by the Microsoft AI for Good Lab and released as open-source software under the MIT Licence. The contributors recorded in the repository's commit history are Meygha Machado, Caleb Robinson, Joaquín Rivero, Cameron Birge, Marcelo Duarte, and Anthony Cintron Roman. The accompanying research paper additionally credits Anthony Ortiz, Simone Fobi Nsutezo, Kevin White, Inbal Becker-Reshef, and Juan M. Lavista Ferres. This plain-language report was prepared under the Ethical Tech CoLab at the NYU Center for Global Affairs as part of masters research (2026), on the fork of the project held at Ethical-Tech-CoLab/haste.

HASTE is an open-source platform that lets a trained analyst, working without code, fit a disposable machine-learning model to a single disaster and produce a building-by-building damage estimate within hours. This report reads the platform's own source code alongside its documentation to set out what every adjustable and hardcoded setting means, what the published performance figures do and do not establish, and where the answer it produces is constrained by things the software does not control.

Key figures

31

field deployments since early 2023, developer-reported (§08)

3

labelled buildings, across at least two categories, before the in-browser classifier will train

0.1

damage threshold, one tenth of a building's visible area, the most consequential single number in the platform (§05)

0.84

area under the ROC curve at one per cent of labels, against 0.88 fully supervised (§08)

00

Foreword

In the hours after an earthquake, a hurricane, or a wildfire, the single most useful thing a relief coordinator can hold is a map of which buildings are still standing. Where should search teams go first? Which neighbourhoods need shelter, and how many people are likely to need it? Which roads lead to places that no longer exist? Those questions have to be answered before anyone on the ground has been able to walk the affected area, and often before the affected area is safe to walk at all.

The traditional answers are slow. Teams fly helicopters over the damage, photograph it, and interpret the photographs by hand. Official mapping services task a satellite, wait for a cloud-free pass, and publish a product days later. Both are careful and both are valuable, but neither reliably delivers inside the first days, which is the window in which decisions about people, supplies, and attention are actually being made.

HASTE, which stands for High-speed Assessment and Satellite Tracking for Emergencies, is an open-source research platform built to shorten that gap. It lets a trained analyst who is not a machine-learning engineer take fresh satellite or aerial imagery of a disaster zone, mark a small number of examples by hand, and have a computer extend those examples across the whole affected area to produce a building-by-building damage estimate. This report explains, in non-technical language, what HASTE does, how it does it, what each of its settings means, and what it cannot be trusted to do.

01

Executive Summary

HASTE is a web-based platform that turns post-disaster satellite imagery into an estimate of which individual buildings have been damaged. It was developed by the Microsoft AI for Good Lab and released as open-source software under the MIT licence. The version reviewed here is the copy held in the Ethical Tech CoLab repository, which is a fork of the original project at microsoft/haste.

The platform's central design choice is to train a fresh model for each disaster rather than maintain one global model that tries to recognise damage everywhere. A hurricane in Jamaica and an earthquake in Türkiye leave visually different traces on visually different building stock, and a model tuned to one is not expected to work on the other. HASTE accepts that limitation deliberately in exchange for speed: a model fitted to one event, from one analyst's labels, can be ready in minutes.

A human being is required at every stage, and there is no automatic mode. The project documentation states repeatedly that outputs are preliminary signals requiring expert validation rather than authoritative damage assessments. Section 09 sets out where the human sits in the workflow.

HASTE offers two routes from imagery to an answer. The faster route, Rapid Building Assessment, computes a numerical fingerprint for every building in the area, asks the analyst to label a handful of them, and trains a very small classifier inside the web browser, which scores all the rest in seconds. The slower route, Damage Mapping, asks the analyst to draw damaged and undamaged areas by hand and trains a full image-segmentation model on a graphics processor, producing a continuous, pixel-level damage map.

Both routes end at the same place: a per-building damage figure, a set of accuracy measures computed against a human-labelled validation sample, and an estimate of the total number of damaged buildings with a stated margin of error.

The platform depends on outside data it does not produce. Imagery comes from commercial and public providers such as Planet, Maxar, Airbus, the European Union's Copernicus programme, and the United States National Oceanic and Atmospheric Administration. Building outlines come from the Overture Maps Foundation. Where those outlines are missing or wrong, HASTE has nothing to attach its predictions to.

According to the research paper published alongside the platform, HASTE has been used in thirty-one field deployments since early 2023, and its outputs have been released openly through the Humanitarian Data Exchange.

02

Background and Rationale

The problem. Rapid damage assessment is a timing problem before it is a technical one. Imagery of a disaster zone often becomes available within a day. An interpretation of that imagery detailed enough to direct a response team to a particular neighbourhood usually does not. The interval between the two is where humanitarian decisions are made with the least information and the greatest consequences.

The gap. Established products have real strengths and known constraints. Copernicus Emergency Management Service Rapid Mapping, the European Union's free on-demand crisis mapping service, delivers standardised map products within hours to days of an activation, but must be formally activated by an authorised user and covers only large-scale emergencies. Manual aerial surveys are accurate but limited in geographic reach and slow to process. Neither easily absorbs the specific situational context that a particular responding organisation cares about, such as one parish, one road corridor, or one category of structure.

The nearer comparator is the automated damage classification literature that grew out of the xView2 challenge, on whose xBD dataset HASTE is itself benchmarked. Against that work the claim is not that nothing existed. It is that existing machine learning approaches assume a globally pretrained model applied to a new disaster, and HASTE trades that for a disposable per event model an analyst fits by hand, in a browser, without writing code.

The response. The proposition HASTE was built around, carried over from earlier in-browser damage-assessment research at the same laboratory, is that human oversight should be structural rather than advisory: the operator is not reviewing a machine's conclusion after the fact, the operator is the source of everything the machine knows about this event.

The design also reflects a practical constraint on who does this work. The people who understand what damage looks like in a given country are rarely the people who can write machine-learning code. HASTE is presented as a no-code platform so that the person supplying the expert judgment and the person operating the model can be the same person.

03

Objectives

The platform is designed to do the following.

  • Produce a building-level damage estimate from post-disaster imagery quickly enough to be useful inside the first days of a response.
  • Allow a non-programmer to fit a model to a specific event using only a map interface, a mouse, and their own visual judgment.
  • Keep a human in control of every consequential step, including the decision to release a result at all.
  • Express uncertainty honestly, by measuring the model against a separate human-labelled sample and reporting a margin of error on the headline damage figure rather than a single confident number.
  • Produce outputs in standard geographic file formats so that they can be opened in the mapping software humanitarian organisations already use, and published openly where appropriate.
  • Remain deployable by others. The platform can be run on a laptop for evaluation or installed on an organisation's own cloud infrastructure, in which case that organisation, and not Microsoft, controls the imagery and the outputs.

04

How HASTE Works

The workflow has a shared beginning, then splits into two routes, then rejoins for validation and reporting.

The shared beginning. An analyst first creates a project, which is simply a container for one disaster event. It records a name, a description, the date of the event, and the affected countries. The date and the countries are stored for reference and do not affect any calculation.

Into that project the analyst adds an image layer, which is the imagery being assessed. It arrives as GeoTIFF, a standard image format that carries the geographic coordinates of every pixel alongside the picture itself. The analyst can upload several files covering the same area, and HASTE merges them into a single mosaic. Imagery from before the event is optional and is used for visual comparison, not for any calculation.

HASTE then obtains the outlines of every building in the area covered by the imagery. By default it downloads them automatically from Overture Maps, an open dataset maintained by the Overture Maps Foundation under the Linux Foundation, which combines OpenStreetMap with machine-derived building datasets from Microsoft and Google and, on the foundation's own account, covers roughly 2.3 billion buildings worldwide. An analyst who has better local data can upload their own outlines instead, as a GeoPackage file of up to 500 megabytes; HASTE converts it to the standard global coordinate system and trims it to the imagery area. When the layer is created the analyst chooses a workflow, Building or Standard, which determines which of the two routes is available.

Route A, Rapid Building Assessment. This route is available when building outlines exist and produces an answer in minutes with no separate training job. First, HASTE computes an embedding for every building. An embedding is a compact list of numbers that describes what the imagery around that building looks like: its texture, its colour, its edges. Two buildings that look alike receive similar lists of numbers. Nothing in the embedding knows anything about damage; it is simply a numerical description of appearance.

The analyst then opens a map, clicks buildings, and assigns each one to one of three categories: Intact, Damaged, or Cloudy, the last meaning that cloud or shadow makes the building impossible to judge. A left click labels a building, a right click removes the label, and holding the control key while dragging labels many at once.

Once at least three buildings have been labelled across at least two categories, a very small statistical model called a logistic regression trains automatically in the browser and immediately predicts a category for every other building in view. The analyst can flip between what they labelled and what the model predicted, see where it is wrong, label a few more examples, and watch the prediction improve. This loop takes seconds, which is what makes the route fast. When satisfied, the analyst runs a final pass that scores every building in the layer and saves the result as a geographic data file.

Route B, Damage Mapping with a trained model. This route does not require building outlines and produces a continuous damage picture across every pixel of the imagery rather than a verdict per building. The analyst draws polygons, rectangles, or circles over a small part of the image and assigns each shape a class. The default classes are Background, Building, and Damaged Building; additional classes of No Damage and Flood Extent are offered for particular event types. The project documentation advises drawing at least five to ten shapes per class, considers seventy to one hundred a good training set, and states that more than roughly one hundred and fifty is unnecessary.

Those shapes train an image-segmentation model, which is a model that assigns a class to every individual pixel rather than to a whole picture. The architecture used is a U-Net with a ResNeXt-50 encoder, a standard and well-understood design in satellite image analysis, started from weights pre-trained on ordinary photographs and then fitted to the analyst's labels. Training runs on a graphics processor, either locally in a container or on cloud computing capacity. The trained model is then run across the entire image layer, producing a per-pixel damage prediction that can be viewed alongside the imagery and downloaded.

Turning pixels into buildings. A per-pixel damage map is not directly useful to a responder, who needs to know about buildings. HASTE therefore overlays the building outlines onto the pixel predictions and, for each building, counts what proportion of the classified pixels inside that outline were called damaged. That proportion, between zero and one, is the building's damage fraction. It is the number that drives everything downstream. The same step also records what proportion of each building was obscured by cloud.

Validation. Both routes end with the same discipline. The analyst opens a validation tool that presents a random sample of roughly two hundred building outlines over the post-event imagery and asks them to judge each one independently as Damaged, Not Damaged, or Unknown. These human judgments are treated as the ground truth against which the model's predictions are scored. Buildings marked Unknown are excluded from the scoring entirely rather than being guessed at.

05

The Variables, Explained Simply

Every number an analyst can adjust, and every number fixed inside the code, is described below in ordinary terms: what it represents, why it is set where it is, and what it changes about the answer.

The label classes. These are the categories an analyst draws with. Each is stored internally as a class value, and later steps identify a class by that position rather than by its name.

ClassStored valueWhat it marks
Background1Ground and anything that is not a building.
Building2An intact structure.
Damaged Building3A structure showing damage. This is the class the damage figures are computed from.
Cloud4Areas where cloud or shadow makes the imagery impossible to read.
The four positional label classes and the values they are stored as.

Two of these choices carry more weight than they appear to. The reason Background exists as a class in its own right is that the model is only taught by the pixels the analyst actually marked, so if a damaged roof is labelled without the ground around it, the model is never told what the ground is, and can produce a blurry, spreading prediction without ever being penalised for it. Labelling a building together with its surroundings is what forces the model to learn a boundary. And cloudy buildings are excluded from the damage statistics rather than counted as intact, which would otherwise bias the result downward in exactly the wet, storm-affected conditions where damage assessment matters most.

No Damage and Flood Extent. Two additional classes offered for earthquake, fire, and flood event types respectively. Flood Extent lets the analyst mark standing water, which HASTE can intersect with building outlines to say which buildings are in water.

Normalisation means and standard deviations. Image pixels are stored as whole numbers, but models learn more stably from values between 0 and 1, so each channel is rescaled by subtracting a mean and dividing by a standard deviation. In the worked example shipped with the platform these are set to 0 and 255, which is a plain rescaling of ordinary 8-bit imagery. When an image layer is prepared automatically, HASTE instead sets every mean to zero and sets each channel's divisor to the brightness value at that channel's 98th percentile, so that the brightest two per cent of pixels are treated as glare and clipped rather than being allowed to compress everything else into a narrow band. This matters because imagery from different sensors arrives on different numeric scales, and a wrong divisor makes the picture unrecognisable to the model.

Ground resolution. HASTE never checks, records, or standardises how many metres of ground each pixel covers. It works on whatever grid the supplied imagery has. Several settings are expressed in pixels rather than metres as a result, and the physical area covered by a training tile therefore varies with the imagery source.

Learning rate, default 0.0001. How large a correction the model makes each time it discovers it was wrong. Too large and it lurches past the right answer; too small and it never gets there within the time available. The default is a conservative value appropriate to fine-tuning a model that already carries useful pre-trained weights.

Batch size, default 32 for training. How many image tiles the model examines before making one correction. Larger batches give a steadier, less noisy signal about which direction to move in, but require proportionally more memory on the graphics processor.

Maximum epochs. One epoch is one complete pass through the training data. A small number is intended here, for a clear reason: the model is being fitted to a few dozen hand-drawn shapes from a single event, and training it longer would mainly teach it to memorise those particular shapes rather than the general appearance of damage. The repository is inconsistent about the actual figure. The worked example specifies 10, the form in the web interface offers 3, and the server falls back to 1 when nothing is supplied, while the training script separately imposes a hardcoded minimum of 10. An analyst who selects 3 in the interface is therefore asking for a maximum below the enforced minimum, which anyone reproducing a result should know.

Training chip size, 256 pixels. The model does not see the whole scene at once. It is shown small square cut-outs 256 pixels on a side, drawn at random from the labelled area, and the training run feeds it 1,024 batches of them per epoch. The model's whole view of the disaster is assembled from these fragments.

Label buffer, nominally 3 metres. Applied to the Building class using the Background class. When a person traces a building, the traced line is never exactly on the wall. This setting draws a narrow band around each building polygon and treats it as Background, which stops the imprecision of human tracing from teaching the model that the pavement is part of the structure. The setting is written in metres but is applied by counting pixels, so it means three metres only when a pixel happens to represent one metre of ground. The script's own documentation notes this.

Patch size, default 2048, and padding, default 64. Used when running the finished model across the full image. Satellite scenes are far too large to process in one piece, so they are cut into square tiles. Predictions are least reliable at the very edge of a tile, where the model can see no context, so each tile is processed with 64 pixels of overlap that are then discarded. This is what stops a visible grid pattern from appearing across the finished damage map.

Initial weights, default none. Meaning the model starts from weights learned on ordinary photographs. An analyst can instead start from a previously trained HASTE model held in the model catalogue, which is offered only where the event type and imagery source match, since a model fitted to wildfire imagery is not a sensible starting point for flood imagery.

The cloud penalty. An optional setting, switched off by default, changes how the model is corrected in areas the analyst marked as Cloud. Rather than being told what class those pixels really are, which nobody knows, the model is penalised whenever it predicts damage beneath a cloud. In plain terms it is taught not to claim to see through weather.

Embedding backbone. The method used to convert the appearance of a building into numbers. The lightweight default is MOSAIKS, an approach published in Nature Communications in 2021 that passes the image through a large set of randomly generated filters rather than learned ones. The counter-intuitive finding of that work is that random filters, applied at sufficient scale, describe an image well enough for a simple linear model to make accurate predictions from them, at a tiny fraction of the computational cost. The alternative offered is DINOv2, a family of vision models released by Meta AI in 2023 that learned general visual features from 142 million unlabelled images, available in small and base sizes.

Output dimensions, default 1024. Applicable to MOSAIKS only. How many numbers describe each building. More numbers capture more nuance and cost more time and memory. When DINOv2 is used this setting is ignored, because the size of the description is fixed by the model itself: 384 numbers for the small version, 768 for the base version, 1024 for the large one.

Kernel size, default 7. The width in pixels of each random filter. Small filters respond to fine texture such as roof material and rubble; larger ones respond to coarser structure. Seven pixels is a middle setting suited to detecting the change in surface texture that collapse produces.

Resize factor, default 4. How much the small image crop around each building is enlarged before being described. In moderate-resolution satellite imagery a single house may span only a handful of pixels, too few for the filters to find anything. Enlarging the crop fourfold gives them something to work with. This is a real trade-off and not a free gain: enlarging an image does not create detail that the sensor never recorded, it only makes the recorded detail large enough to be measured. Choosing DINOv2 forces this setting back to 1, because that model brings its own fixed way of dividing an image into pieces.

Context padding, 8 pixels. Each building crop reaches a little beyond the traced outline, so that the description includes some of the ground immediately around the structure. Debris and scorch marks sit there rather than on the roof.

What is actually described. The crop is divided into a grid of small tiles, each tile is described separately, and only those tiles falling inside the building's outline are averaged together to produce the building's final list of numbers. Buildings that fall outside the imagery keep their place in the file with an empty entry rather than being dropped, because everything downstream matches buildings to predictions by position in the file rather than by name. That design decision is efficient and fragile in equal measure.

The crop cap, 192 source pixels. A limit that stops a single very large building, such as a warehouse or a stadium, from generating an enormous crop and exhausting the available memory. Oversized footprints are cropped from the centre instead.

The in-browser classifier. The model that turns a handful of labelled buildings into predictions for all of them is a logistic regression, one of the oldest and simplest classifiers in statistics. It fits a straight-line boundary through the space of building descriptions and reports, for each building, how far onto the damaged side of that boundary it falls. Its learning rate is 0.1 and it takes 500 steps, values fixed in the code rather than exposed to the analyst, chosen to complete in well under a second so that the label-and-see loop stays interactive.

Regularisation strength, 0.01. A deliberate penalty that discourages the model from placing heavy reliance on any single one of the thousand-odd numbers describing each building. Without it, a model trained on three examples would seize on whatever coincidental feature happened to separate those three, and would fail on the fourth. This single small number is the main defence against overfitting in a workflow explicitly built around very few labels.

Holdout fraction, 0.2 in practice. One fifth of the analyst's labels are held back from training and used only to score the model, which is what produces the live precision, recall, and F1 figures shown for the Damaged class in the side panel. Because those numbers come from labels the model was never shown, they are an honest indication of quality rather than a report of how well the model memorised its own training data. Two weaknesses are worth naming. The sample is tiny: with twenty labels the holdout is four buildings, and four buildings cannot tell you much. And the split is drawn afresh every time a label is added, so the displayed figures jump around from click to click, as the code comments acknowledge.

Measurement rings, 0, 10, and 20 metres. Not to be confused with the label buffer above. HASTE calculates each building's damage fraction three times: once inside the traced outline, once inside a ring extending ten metres beyond it, and once at twenty metres. There are two reasons. Building outlines and satellite imagery are frequently misaligned by several metres, particularly in dense cities and where the ground itself has moved, so the strict outline may sit over the neighbouring plot. And some kinds of damage, notably debris fields and burn scars, appear around a structure rather than on it. The wider measurements are recorded so that this can be examined rather than assumed.

Damage fraction. For a given buffer, the number of pixels classified as Damaged Building divided by the number of classified pixels of any kind inside that shape. Unclassified pixels, recorded as zero, are excluded from both the top and the bottom of that division, so a building half outside the imagery is judged on the half that was actually seen. The result is capped at 1.

Damage threshold, default 0.1. The proportion of damaged pixels above which a building is declared damaged. This is the most consequential single number in the entire platform, and it is set low on purpose. A tenth of a building's visible area showing damage indicators is treated as enough, reflecting a judgment that in the first days of a response, missing a damaged building is a worse error than flagging one that turns out to be intact. It also compensates for partial visibility: a structure photographed from overhead shows mainly its roof, and serious structural damage may present as only a modest patch of altered texture there. Because the threshold is a setting rather than a fixed rule, and because the platform also publishes a full precision-and-recall curve across all possible thresholds, an analyst who disagrees with this trade-off can see exactly what a different choice would cost.

A wrinkle a careful reader should know about. The exported data file also carries a simple yes-or-no damaged column, and that column is set whenever a building contains even one damaged pixel, with no threshold at all. The validation report reads that column; the assessment report applies the tenth-part threshold. The same model and the same human labels will therefore yield slightly different accuracy figures in the two reports. Nothing in the documentation flags this, and it is the kind of discrepancy that could confuse a reader comparing the two.

Minimum footprint area, default 50 square metres. Buildings smaller than this are excluded from the population used for the final headline estimate. The stated reasoning is that these are structures a human reviewer could not realistically have judged in the validation step, so including them in an extrapolation built from human judgments would be unsound. The consequence, which the platform does not hide, is that small informal structures are absent from the headline count, and those structures are common in precisely the settlements most exposed to disaster.

Cloud exclusion. Any building with a cloud fraction above zero is excluded from the damage count entirely. This is a strict rule, not a proportional one; a building need only be slightly obscured to be set aside.

The final estimate. HASTE does not publish a raw count of the buildings its model called damaged. It uses the human-validated sample to estimate the true damage rate and scales that rate up to the full population of buildings. The arithmetic is a standard finite-population survey estimate. The sample proportion, written as p-hat, is the number of buildings a human confirmed as damaged divided by the number of buildings the human judged either way, with Unknown buildings excluded from both figures. The population, written as N, is the count of building outlines larger than the minimum area, and the estimated number of damaged buildings is simply N multiplied by p-hat.

The sampling fraction, written as f. The size of the validated sample divided by the population. It enters the calculation through what statisticians call a finite-population correction, the term (1 minus f) in the variance. Its effect is intuitive: if you have checked half of all the buildings by hand, your uncertainty about the other half should be smaller than if you had checked one per cent of them. Standard survey formulas assume an infinite population and would overstate the uncertainty here.

The critical value z, fixed at 1.959963984540054. This is the constant that turns a standard error into a 95 per cent confidence interval, and it is written out as a literal number in the code specifically so that the platform does not have to load a heavy statistical library to obtain it. The reported interval runs from N times (p-hat minus z times the standard error) to N times (p-hat plus z times the standard error).

What the interval means in plain terms. If the same sampling exercise were repeated many times, an interval calculated this way would contain the true number of damaged buildings in about nineteen cases out of twenty. It accounts for the fact that a sample was taken. It does not account for a mislabelled sample, misaligned building outlines, or a model that is wrong in a consistent direction, and it should not be read as expressing total uncertainty about the figure.

06

Reading the Results

The visualiser. The pre-event and post-event imagery are shown side by side with a swipe control between them, the predicted damage layer overlaid on both, and the raw per-pixel predictions available as an optional toggle. Opacity, contrast, hue, and saturation sliders help make damage visible in imagery that was captured in poor light.

On the map, each building is shaded according to its damage fraction in five bands, cut at one fifth, two fifths, three fifths, and four fifths, running from white through pale peach and orange to red and dark red. This is a presentational scale only. It should not be read as a severity classification in the sense used by structural engineers, because the model was never taught degrees of damage. It only ever learned to separate damaged from intact, and the shading reflects how much of a roof it flagged rather than how badly the building was hurt.

The validation report. This compares the model's predictions against the analyst's own validation labels and reports overall accuracy, precision, recall, and F1 for each class, a macro-averaged F1, and a confusion matrix. Precision answers the question: of the buildings the model called damaged, what share really were? Recall answers the opposite: of the buildings that really were damaged, what share did the model find? These two errors have different humanitarian costs, and the report deliberately does not merge them into a single figure.

The assessment report. This summarises the whole layer: the total number of building outlines, how many were excluded as cloud-covered, how many of the remainder were predicted damaged and what percentage that represents, the accuracy metrics at the chosen threshold, a precision-and-recall curve across all thresholds, and the estimated total of damaged buildings with its 95 per cent confidence interval. Where no validation labels exist, the report still shows the prediction counts but withholds the accuracy metrics and the estimate rather than presenting an unvalidated number.

Downloads. Results can be exported as a GeoPackage, an open standard file that opens in common mapping software, along with the training artefacts and the building outlines used. This matters for accountability: a partner receiving a HASTE layer can inspect it independently rather than taking a summary figure on trust.

07

Data Sources

Imagery. HASTE holds no imagery of its own; the analyst supplies it. Only GeoTIFF files are accepted. The documentation records that in past activations imagery has come from the following providers.

  • Planet, a commercial operator of a large constellation of small satellites that publishes disaster imagery openly.
  • Maxar, now Vantor, a commercial provider of high-resolution imagery with its own open data programme for disasters.
  • The Airbus Foundation.
  • Sentinel-1 and Sentinel-2, the radar and optical satellites of the European Union's Copernicus programme, whose data is free to all.
  • Products from the Copernicus Emergency Management Service.
  • Aerial imagery from the United States National Oceanic and Atmospheric Administration.

The software itself has no live connection to any of these providers. Placeholder functions for fetching imagery directly from Maxar and Planet exist in the code but are empty. In practice the analyst supplies a link to a file, and for security reasons those links may point only at Azure Blob Storage or Amazon S3, the two hosting services on the platform's permitted list. Public disaster imagery from the major providers is generally published on one of those, which is why the restriction is workable.

HASTE does adjust for the provider in one respect. Different satellites record their colour bands in different orders, and some record bands the human eye cannot see, so the platform holds a lookup table of band orders for Planet Scope, Planet Skysat, Maxar, Sentinel-2, and one partner-specific format, and uses it to assemble a correct colour picture. Where the source is unknown, HASTE falls back to the labels embedded in the file, and failing that assumes the first three bands are red, green, and blue.

Building outlines. Overture Maps by default, described in section 04, and analyst-supplied outlines where better local data exists. HASTE reads the Overture data anonymously from a public store, taking the most recent release available and falling back to a fixed February 2026 release if it cannot determine one. Only polygons are kept.

What HASTE does not use. The platform draws on imagery and building outlines and nothing else. It does not read ground reports, weather data, sensor networks, social media, or population figures. Its picture of a disaster is strictly what a camera in orbit could see, which is a narrower thing than what happened.

Personal data. The platform is not designed to identify people, does not ingest or output personal data, and does not treat person-scale features in imagery as signal.

08

Evidence of Performance

Everything in this section is the developer's own account. The benchmark results, the deployment record, and the field precision and recall figures all originate from the Microsoft paper and repository. None of them has been independently reproduced or externally validated, here or elsewhere, so they should be read as what the team that built the platform reports about it rather than as third-party verification. That is not an accusation of overstatement. It is the provenance of the evidence, and it is the first thing a sceptical reader is entitled to know.

Validated accuracy. The research paper published alongside the platform, HASTE: A Platform for Rapid Post-Disaster Building Damage Assessment (arXiv:2607.11838), reports experiments on xBD, a public benchmark dataset of paired pre-event and post-event satellite imagery with expert damage annotations. The team collapsed the benchmark's minor, major, and destroyed categories into a single damaged category and measured how well each embedding method performed as the number of labels was varied.

The discrimination score in Table 1 is the area under the receiver operating characteristic curve, usually shortened to AUROC. In plain terms it is the probability that the model ranks a randomly chosen damaged building above a randomly chosen intact one. It runs from 0 to 1, a coin flip scores 0.5, and 1.0 would mean the model never gets a pair the wrong way round. The measure says nothing about where the threshold should sit: a model can rank buildings well and still mislabel a great many of them once a cutoff is applied.

ApproachLabels usedDiscrimination score
Strongest embedding1 per cent0.84
Strongest embedding10 per cent0.91
Fully supervised ResNet-50All labels0.88
Table 1. Reported performance on the xBD benchmark as the number of labels was varied.

The practical claim is that a handful of labels plus a good general-purpose image description gets close to a conventionally trained model. The headline figure is a modest deficit rather than a match: at one per cent of labels the score is 0.84 against the fully supervised 0.88, and it is at ten per cent, where it reaches 0.91, that the fast route actually overtakes the baseline. What one per cent buys is not equal accuracy but most of the accuracy, hours sooner and without a machine-learning engineer, which is a real trade and a different claim.

The table does not say which of the two embedding methods produced these scores, and this report cannot resolve it from the published material. Since the lightweight option is the default an analyst would run, and the heavier one is the more capable, that gap matters to anyone reading the figures as a prediction of what they will get.

Deployment record. The figures that follow are evidence of adoption rather than of correctness, and only the Rolling Fork assessment carries an accuracy measurement against field ground truth.

The same paper reports thirty-one field deployments since early 2023, including four cities assessed within three days of the February 2023 Türkiye earthquakes, a tornado assessment in Rolling Fork delivered in under two hours at 0.86 precision and 0.80 recall against field ground truth, and the August 2023 Maui wildfire, where imagery available at nine in the morning yielded an assessment by one in the afternoon identifying roughly 1,700 damaged buildings.

For Hurricane Melissa in Jamaica in late 2025, four areas covering about 2,300 square kilometres were assessed. In Black River, some 110,000 building outlines were examined, of which around 65,000 were obscured by cloud.

AreaRecallPrecisionEstimated damaged
Black River96 per cent82 per cent31,000 buildings
Montego Bay86 per cent71 per centNot reported
Table 2. Two of the areas assessed during the Hurricane Melissa response.

The variation between Black River and Montego Bay, on the same event with the same team days apart, is substantial, and the very large share of cloud-obscured buildings in Black River is a reminder that a headline damage estimate can rest on a minority of the buildings actually present.

09

Human Oversight and Governance

There is no autonomous mode. A person selects the imagery, provides every label the model learns from, reviews the predictions, validates a sample, and decides whether to distribute the result.

Outputs distributed publicly, including through the Humanitarian Data Exchange and partner mapping systems, carry notices warning against over-reliance and encouraging cross-validation against ground reports and other imagery. The notices frame the output as exploratory rather than definitive, and name HASTE, the imagery provider, and the building-outline dataset so that a downstream user can assess provenance for themselves.

The documentation states that where a widespread inaccuracy or pattern of misuse is identified, the laboratory may publish guidance recommending temporary suspension or restricted use of the workflow.

Labels are not reused. Each event's labels belong to that event and are not accumulated into a global training set, which follows directly from the train-per-event design and also limits the accumulation of one analyst's interpretive habits across many responses.

10

Limitations and Caveats

The repository documents its own weaknesses at length. The most consequential are these.

Every performance figure in this report is developer-supplied. This one is not in the repository's own list, and it is the limitation a sceptical reader should weigh first. As section 08 sets out, none of the reported results has been independently reproduced. The rest of this report reads the source code and reports what it finds, which is first-hand; the performance evidence does not have that standing, and the two should not be given the same weight.

The output depends heavily on who did the labelling. Two competent analysts working the same event from the same imagery can produce materially different results. The documentation gives a concrete example from the Hurricane Melissa response, where an initial set of 153 labels produced predictions that described buildings as around 20 per cent damaged when they were in fact totally destroyed; the error was caught by visual inspection and corrected by adding a further 107 labels and retraining.

Imagery quality governs everything. Cloud, haze, low light, and imagery captured at an angle rather than from directly overhead all degrade performance, sometimes severely. Planet imagery in particular sits below the roughly 30 centimetre resolution that building-detection models generally prefer, which is part of why the platform leans on external building outlines rather than trying to find buildings itself.

Building-outline coverage is uneven in a way that matters. OpenStreetMap and Microsoft Building Footprints are weakest in the Global South, in informal settlements, in conflict-affected areas, and where construction has been rapid and recent. These are precisely the populations most exposed to disaster. A building with no outline is invisible to HASTE regardless of how badly it was damaged, and the 50 square metre minimum area removes further small structures from the headline figure.

The model does not generalise, by design. A model fitted to one event is not expected to transfer to another, which also means HASTE cannot be operated as a standing monitoring system.

False positives and false negatives have distinct causes. False positives arise from shadows, ordinary construction and demolition unrelated to the disaster, atmospheric artefacts, and vegetation change. False negatives arise from subtle structural damage visible only from an angle, damage hidden by cloud or shadow, and damage at a scale finer than the imagery can resolve.

No contextual data is incorporated. There is no ground truth, no weather, no sensors, no population data. Flood extent is intersected with buildings but water depth is never estimated, so HASTE can say a building is in water and not how deep.

The confidence interval understates real uncertainty. It quantifies the error introduced by sampling and nothing else, as section 05 explains.

Some findings from reading the source code should be added to the project's own list, since they bear on how much weight an output can carry.

There is no genuinely held-out validation data in the trained-model route. The measure used to decide which version of the model to keep is computed on random cut-outs of the same imagery the model was trained on. It is therefore a measure of fit rather than of generalisation. The independent check in HASTE comes later, from the human validation sample, which is the number a reader should rely on.

The meaning of each class is fixed by its position in a list. Damaged Building is understood downstream as the third class and Cloud as the fourth. An analyst who reorders the classes when setting up a project will get damage figures computed from the wrong category, without any error being raised.

Several defaults disagree between the worked example, the web form, and the server. This is most visible for the number of training passes. Two analysts following the documentation by different routes may not be running the same configuration.

Predictions are matched to buildings by position in a file rather than by an identifier. This is fast, and it means that anything which reorders or filters the building list between steps would silently misattribute damage. The developers are evidently aware of the risk, since the code goes to some trouble to preserve row positions even for buildings it cannot assess.

The platform is not validated for production or autonomous use. It has not been designed, tested, or validated for production deployment or autonomous decision-making, and its outputs are not authoritative damage assessments. The documentation states plainly that users should not rely solely on HASTE outputs for decisions affecting safety, property, or human life, and should not use them to trigger public alerts or resource deployments without independent verification.

11

Practical Nature of the Platform

HASTE is a full web application rather than a single page or a script. It consists of a browser interface, a set of programming interfaces that the browser front end calls, background workers that handle long jobs such as preparing imagery and training models, a tile server that streams large satellite images into the map view, and a shared Python library holding the analysis logic.

It can be run in two ways. A complete local instance can be started on one machine using Docker, a tool that packages software with everything it needs to run, which requires no cloud account and is intended for evaluation. A production instance is deployed to Microsoft Azure with a single command, and the deploying organisation controls it entirely.

The local configuration is explicitly flagged as unsuitable for production, since it disables authentication and uses an in-memory storage emulator. A separate hardening checklist is provided for real deployments.

The repository shows active and careful security practice, including automated code scanning and secret scanning, and documented handling of known vulnerabilities in the underlying geospatial libraries where a patched version was not available. Sample data from Hurricane Melissa and the Lahaina wildfire is published so that the platform can be tried without sourcing imagery first.

12

Intended Audience and Use

HASTE is aimed at trained humanitarian and disaster-response practitioners working with post-event imagery, and at researchers studying rapid damage-assessment methods. The documentation names non-governmental organisations, United Nations agencies, and government users as the intended downstream consumers of its outputs.

Its stated purpose is to contribute information to preliminary damage assessment in the first hours and days, supplementing rather than replacing expert assessment, and to indicate where damage may be concentrated so that attention can be directed.

It is explicitly not intended for use as any of the following.

  • An authoritative damage register.
  • Ground truth for insurance, governmental, or public-reporting purposes.
  • The sole basis for search-and-rescue tasking or resource allocation.
  • An automated alerting system of any kind.

13

Application: the Mariupol Corridor Severity Model

This section describes work by the Ethical Tech CoLab rather than by the platform's developers, and it describes work in progress. Nothing below has been published as a result, and no HASTE-derived figure currently enters the model it discusses. It is included because the reasoning about when this platform is and is not the right instrument is more useful stated against a real case than in the abstract.

The gap in our own model. The CoLab's Mariupol Corridor Severity Model is a daily civilian-danger index for the 2022 siege, built from six components. The sixth is infrastructure damage, the cumulative share of the city's structures visibly damaged or destroyed, and it is drawn from satellite damage assessments published by UNOSAT, the United Nations Satellite Centre. UNOSAT publishes on particular dates rather than continuously. Across a siege of seventy-seven days the model has five anchor points: 0.02 on 5 March, 0.04 on 14 March, 0.14 on 26 March, 0.32 on 7 May, and 0.33 on 20 May. Every day in between is read off a straight line drawn between them.

The model is candid about what this costs it. In its own words, the word daily describes the resolution of the output and not the resolution of the evidence, and it cannot distinguish 18 March from 19 March on any evidence about what happened on those two days. The practical danger is specific: a stretch of the siege that UNOSAT did not image is rendered as a gently rising line, so an absence of published assessment can be mistaken for an absence of destruction.

Why HASTE is the plausible instrument. The constraint that limits UNOSAT is not imagery, it is the analytic labour of a formal before-and-after building-by-building assessment. HASTE is built for exactly that constraint. It needs only post-event imagery, it fits a disposable model to one event from a small number of hand-marked examples, and it returns a per-building damage fraction with a stated margin of error. In principle it could produce assessments for dates that lie between the UNOSAT anchors, replacing an interpolated line with measured points and letting the damage component move on the evidence rather than on the arithmetic.

What has actually been done. So far, scoping. The model still runs on the five UNOSAT anchors and the straight lines between them, and the per-building display in the tool remains illustrative until real geodata is supplied. We are working through whether HASTE-style assessment of additional 2022 imagery dates is feasible and, more importantly, whether the result would be honest enough to publish.

What would have to be true first. The obstacles are the ones this report has already described, met in a hard case. They are set out here rather than discovered later.

  • Archival imagery. HASTE assesses whatever imagery it is given, and it holds none. A retrospective study needs cloud-free 2022 coverage of Mariupol on the specific dates that would fill the gaps, which is a sourcing problem before it is a modelling one.
  • Building outlines. The platform can only attach a prediction to a building it has an outline for, and section 10 records that outline coverage is weakest in exactly the places under greatest pressure. Outlines derived before the siege will also disagree with imagery of a city whose ground and structures have moved.
  • The event type. HASTE is benchmarked on xBD, a dataset of disaster damage. Shelling and airstrike damage is not what the reported figures were measured on, and the platform's own design assumption is that a model fitted to one kind of event is not expected to transfer to another.
  • Human validation. Every HASTE result depends on an analyst marking training examples and judging an independent validation sample. For a retrospective conflict case that means someone competent to read 2022 imagery of this city, and their judgment, not the model's, sets the ceiling on what the number is worth.
  • Status of the output. HASTE describes its outputs as preliminary signals requiring expert validation rather than authoritative damage assessments. The Mariupol model already declares its damage component a lower bound. A HASTE-derived series would have to carry both caveats, not shed them by being newer.

The risk the pairing creates. A denser series looks better evidenced whether or not it is. Replacing five anchors with, say, twenty assessed dates would produce a damage curve that appears to respond to events day by day, and a reader would reasonably infer that the model now knows more about 18 March than it did. It would know more only if each new point were validated to the standard the five UNOSAT anchors carry. Both projects state their uncertainty explicitly and neither should be allowed to launder the other's: HASTE's confidence interval accounts for sampling error and nothing else, and the severity model's daily resolution is a property of its output rather than of its evidence. If HASTE assessments do enter the model, they will be labelled as such, dated, and reported alongside the UNOSAT figures rather than blended into them.

14

Conclusion

HASTE's most useful contribution is not a modelling advance but a reallocation of labour. It moves the machine-learning work out of the way so that the scarce resource, the judgment of someone who can look at an image and know what damage looks like in that country, is applied where it counts: to choosing the imagery, marking the examples, and checking the answer. The model is small, disposable, and fitted to one event, and the platform treats it as such.

The reporting design is equally deliberate. By requiring an independent human validation sample before it will publish an accuracy figure, by separating precision from recall rather than averaging them away, and by attaching a confidence interval to its headline number, the platform makes it harder to mistake a fast estimate for a survey.

The honest reading is that HASTE is fast, transparent about its assumptions, and constrained by things it does not control. Its ceiling is set by the resolution of the imagery available and the completeness of the building outlines beneath it, and both of those are weakest in the places where humanitarian need is greatest. That is not a flaw in the software so much as a statement about the wider data landscape, but it means the platform's value in any given response depends on conditions decided long before the disaster occurred. Read alongside its own documented limitations, it is a credible example of a machine-learning tool built to assist expert judgment rather than to displace it.

References

Sources

  1. 01Microsoft AI for Good Lab. HASTE: A Platform for Rapid Post-Disaster Building Damage Assessment. arXiv:2607.11838.
  2. 02Microsoft AI for Good Lab. HASTE, open-source platform released under the MIT Licence.
  3. 03Overture Maps Foundation. Buildings theme, combining OpenStreetMap with machine-derived building datasets from Microsoft and Google.
  4. 04OpenStreetMap contributors. OpenStreetMap building data.
  5. 05Microsoft. Global ML Building Footprints.
  6. 06xBD. Public benchmark dataset of paired pre-event and post-event satellite imagery with expert damage annotations.
  7. 07MOSAIKS. A generalizable and accessible approach to machine learning with global satellite imagery. Nature Communications, 2021.
  8. 08Meta AI. DINOv2, a family of vision models trained on 142 million unlabelled images, 2023.
  9. 09European Commission. Copernicus Emergency Management Service, Rapid Mapping.
  10. 10European Union. Copernicus Sentinel-1 and Sentinel-2 radar and optical satellite data.
  11. 11Planet. Open disaster imagery from a constellation of small satellites.
  12. 12Maxar, now Vantor. Open Data Program for disaster response imagery.
  13. 13United States National Oceanic and Atmospheric Administration. Emergency response aerial imagery.
  14. 14OCHA. Humanitarian Data Exchange, the channel through which HASTE outputs have been released openly.
  15. 15Ethical Tech CoLab. Mariupol Corridor Severity Model, a daily civilian-danger index for the 2022 siege, discussed in section 13.
  16. 16UNOSAT, United Nations Satellite Centre. Ukraine damage assessments, activation CE20220223UKR, the source of the five anchor points in section 13.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file diff --git a/static-site/print/haste/index.txt b/static-site/print/haste/index.txt index 9bc15be3c..8b63e58a9 100644 --- a/static-site/print/haste/index.txt +++ b/static-site/print/haste/index.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -12:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","print","haste",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","haste","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -14:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -16:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","print","haste",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","haste","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:"$Sreact.suspense" -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1c:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -29,7 +29,7 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 19:[] 10:"$W19" 11:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1e:Ta14, @page { size: 210mm 297mm; margin: 20mm 18mm; } @@ -185,6 +185,6 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 65:["$","div","6",{"children":["$undefined",["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"Archival imagery. HASTE assesses whatever imagery it is given, and it holds none. A retrospective study needs cloud-free 2022 coverage of Mariupol on the specific dates that would fill the gaps, which is a sourcing problem before it is a modelling one."}],["$","li","1",{"className":"pl-1","children":"Building outlines. The platform can only attach a prediction to a building it has an outline for, and section 10 records that outline coverage is weakest in exactly the places under greatest pressure. Outlines derived before the siege will also disagree with imagery of a city whose ground and structures have moved."}],["$","li","2",{"className":"pl-1","children":"The event type. HASTE is benchmarked on xBD, a dataset of disaster damage. Shelling and airstrike damage is not what the reported figures were measured on, and the platform's own design assumption is that a model fitted to one kind of event is not expected to transfer to another."}],["$","li","3",{"className":"pl-1","children":"Human validation. Every HASTE result depends on an analyst marking training examples and judging an independent validation sample. For a retrospective conflict case that means someone competent to read 2022 imagery of this city, and their judgment, not the model's, sets the ceiling on what the number is worth."}],["$","li","4",{"className":"pl-1","children":"Status of the output. HASTE describes its outputs as preliminary signals requiring expert validation rather than authoritative damage assessments. The Mariupol model already declares its damage component a lower bound. A HASTE-derived series would have to carry both caveats, not shed them by being newer."}]]}]]}] 66:["$","p","7",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"The risk the pairing creates."}]," ","A denser series looks better evidenced whether or not it is. Replacing five anchors with, say, twenty assessed dates would produce a damage curve that appears to respond to events day by day, and a reader would reasonably infer that the model now knows more about 18 March than it did. It would know more only if each new point were validated to the standard the five UNOSAT anchors carry. Both projects state their uncertainty explicitly and neither should be allowed to launder the other's: HASTE's confidence interval accounts for sampling error and nothing else, and the severity model's daily resolution is a property of its output rather than of its evidence. If HASTE assessments do enter the model, they will be labelled as such, dated, and reported alongside the UNOSAT figures rather than blended into them."]}] 1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -67:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +67:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 18:null 1d:[["$","title","0",{"children":"HASTE: High-speed Assessment and Satellite Tracking for Emergencies (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L67","6",{}]] diff --git a/static-site/print/mariupol-severity-model/__next._full.txt b/static-site/print/mariupol-severity-model/__next._full.txt index 4dba8baf1..38bfabe8f 100644 --- a/static-site/print/mariupol-severity-model/__next._full.txt +++ b/static-site/print/mariupol-severity-model/__next._full.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -12:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","print","mariupol-severity-model",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","mariupol-severity-model","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -14:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -16:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","print","mariupol-severity-model",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","mariupol-severity-model","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:"$Sreact.suspense" -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1c:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -29,7 +29,7 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 19:[] 10:"$W19" 11:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1e:Ta14, @page { size: 210mm 297mm; margin: 20mm 18mm; } @@ -174,6 +174,6 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 5a:["$","li","20",{"children":[["$","span",null,{"className":"print-ref-num","children":"21"}],["$","span",null,{"children":"World Health Organization. WHO Housing and Health Guidelines, 2018, healthy indoor minimum of 18 degrees Celsius."}]]}] 5b:["$","li","21",{"children":[["$","span",null,{"className":"print-ref-num","children":"22"}],["$","span",null,{"children":"World Bank. Ukraine disability prevalence as recorded in Ministry of Social Policy pension-system statistics."}]]}] 1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -5c:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +5c:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 18:null 1d:[["$","title","0",{"children":"The Mariupol Corridor Severity Model (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L5c","6",{}]] diff --git a/static-site/print/mariupol-severity-model/__next._head.txt b/static-site/print/mariupol-severity-model/__next._head.txt index 69aa33e84..be3220f67 100644 --- a/static-site/print/mariupol-severity-model/__next._head.txt +++ b/static-site/print/mariupol-severity-model/__next._head.txt @@ -1,6 +1,6 @@ 1:"$Sreact.fragment" -2:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -3:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +2:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +3:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 4:"$Sreact.suspense" -5:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +5:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 0:{"rsc":["$","$1","h",{"children":[null,["$","$L2",null,{"children":[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]]}],["$","div",null,{"hidden":true,"children":["$","$L3",null,{"children":["$","$4",null,{"name":"Next.Metadata","children":[["$","title","0",{"children":"The Mariupol Corridor Severity Model (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L5","6",{}]]}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/print/mariupol-severity-model/__next._index.txt b/static-site/print/mariupol-severity-model/__next._index.txt index f98a4fd90..18b9d82b6 100644 --- a/static-site/print/mariupol-severity-model/__next._index.txt +++ b/static-site/print/mariupol-severity-model/__next._index.txt @@ -1,13 +1,13 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -d:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +d:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} 8:["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$Ld",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$Ld",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$Ld",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$Ld",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] diff --git a/static-site/print/mariupol-severity-model/__next._tree.txt b/static-site/print/mariupol-severity-model/__next._tree.txt index d649f1c53..c279a8857 100644 --- a/static-site/print/mariupol-severity-model/__next._tree.txt +++ b/static-site/print/mariupol-severity-model/__next._tree.txt @@ -1,7 +1,7 @@ -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] 0:{"tree":{"name":"","param":null,"prefetchHints":16,"slots":{"children":{"name":"print","param":null,"prefetchHints":0,"slots":{"children":{"name":"slug","param":{"type":"d","key":"mariupol-severity-model","siblings":null},"prefetchHints":0,"slots":{"children":{"name":"__PAGE__","param":null,"prefetchHints":0,"slots":null}}}}}}},"staleTime":300,"buildId":"static-snapshot"} diff --git a/static-site/print/mariupol-severity-model/__next.print.txt b/static-site/print/mariupol-severity-model/__next.print.txt index 460afe242..fa617e997 100644 --- a/static-site/print/mariupol-severity-model/__next.print.txt +++ b/static-site/print/mariupol-severity-model/__next.print.txt @@ -1,4 +1,4 @@ 1:"$Sreact.fragment" -2:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -3:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +2:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +3:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] 0:{"rsc":["$","$1","c",{"children":[null,[["$","style",null,{"children":"\n /* The site is dark-first, and a dark color-scheme makes Chrome paint\n the printed sheet's margin area dark too. Paper is light. */\n html { color-scheme: light !important; }\n html, body { background: #ffffff !important; }\n body::after { content: none !important; }\n header, footer, .site-bg, .aura { display: none !important; }\n "}],["$","$L2",null,{"parallelRouterKey":"children","template":["$","$L3",null,{}]}]]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/print/mariupol-severity-model/index.html b/static-site/print/mariupol-severity-model/index.html index 009c4248c..410b9cf71 100644 --- a/static-site/print/mariupol-severity-model/index.html +++ b/static-site/print/mariupol-severity-model/index.html @@ -1,4 +1,4 @@ -The Mariupol Corridor Severity Model (print edition) · NYU Ethical Tech CoLab

Publications · Academic report

The Mariupol Corridor Severity Model

A Daily Measure of Civilian Danger During the Siege of Mariupol, March to May 2022

Ethical Tech CoLab

NYU Center for Global Affairs

July 2026

Christine, GitHub user ChristineLumen. Prepared as masters research at the NYU Center for Global Affairs, under the Ethical Tech CoLab. Repository version 0.9.

A siege does not kill people at a steady rate, and it does not kill them only by shelling. Some days the shells fall. Other days the guns are quiet, and what is killing people is that there has been no heating for three weeks, no running water, no medicine, and no convoy has come through. From a distance, the quiet days can look like improvement. This model produces one number for every day of the siege of Mariupol, built so that a lull in shelling cannot by itself lower the assessment.

Key figures

77

days scored, one severity number for every day from 5 March to 20 May 2022

54

of those days on which the deprivation clock was the dominant driver, against none for the three violence components

0.742

the composite score any single saturated component forces, above the 0.70 threshold for the Critical phase

7.5

weeks between 9 March, when the model first says evacuation is warranted, and the negotiated mechanism arriving on 30 April

00

Foreword

A siege does not kill people at a steady rate, and it does not kill them only by shelling. Some days the shells fall and the count of the dead rises. Other days the guns are comparatively quiet, and what is killing people is that there has been no heating for three weeks, no running water, no medicine, and no convoy has come through. From the outside, and especially from the outside at a distance, the quiet days can look like improvement.

Between March and May of 2022 the city of Mariupol was encircled. Corridors for civilians to leave were announced, and then did not hold. Weeks passed. A negotiated evacuation involving the United Nations and the International Committee of the Red Cross eventually took people out of the Azovstal steelworks at the very end of April. The question this repository asks is a narrow and uncomfortable one: on which day did the facts on the ground first become bad enough that the legal obligations owed to those civilians were plainly engaged, and how long afterwards did anything actually arrive?

This report explains, in non-technical language, what the model measures, how it turns observations into a single daily number, why each of its ingredients was chosen, and what it cannot do. It is written for readers who work in policy, humanitarian operations, or law, and who have never written a line of computer code.

01

Executive Summary

This repository contains a research prototype: an experimental piece of software that calculates, for every single day of the siege of Mariupol from 5 March to 20 May 2022, a number between 0 and 1 describing how dangerous it was for civilians to remain in the city. The repository calls this number the composite severity, written as S.

The severity number is built from six ingredients, which the repository calls components. Three of them describe violence: how many attacks happened each day, how close the fighting was to where people lived, and how much of that violence was aimed at civilians or broke promises of safe passage. Three of them describe the slower forms of harm that a siege produces: the cold in a city whose heating has been destroyed, the accumulating time under encirclement with no aid arriving, and the progressive destruction of the buildings people shelter in.

The six ingredients are deliberately not averaged in the ordinary way. The model uses a method of combination that pulls the answer close to whichever ingredient is worst. The reasoning is stated plainly in the repository's methodology: a population that is safe from shelling but freezing without water is not in medium danger. A plain average would have made the lull in shelling after mid-March look like recovery. It was not.

The resulting daily number is sorted into one of five phases, from Minimal to Critical, each carrying an indicated posture ranging from routine monitoring to immediate protective action. The presentation deliberately follows the five-category convention used by the INFORM Severity Index and by the Integrated Food Security Phase Classification, so that humanitarian readers recognise the shape of the output.

The severity number is then multiplied by a vulnerability weight of 1.114, derived from the share of the pre-siege population who were children, elderly, or living with a disability, to produce what the repository calls a priority index.

The repository's central claim is a claim about timing. It argues that publicly available information was sufficient to establish, well before the end of March, that the factual conditions triggering obligations under International Humanitarian Law were met, and that the negotiated evacuation mechanism did not arrive until 30 April. The repository states its conclusion in one sentence: the binding constraint was not information but consent.

The tool is a single web page that runs in an ordinary browser. Its author describes it as retrospective, not validated against how people actually behaved, and explicitly not operational guidance. It is the first of three intended axes of a fuller decision framework; the other two, feasibility and destination viability, are not built.

This report finds the model's reasoning clear and unusually well documented, and its candour about its own weaknesses genuine. It also identifies several points where a reviewer should verify a citation or a data figure before the work is published or relied upon.

02

Background and Rationale

The problem. During a siege, the people who might intervene are almost always working from partial information, and the information that is available degrades exactly when conditions are worst. Reporters leave or are killed. Communications fail. The number of recorded incidents falls, not because less is happening but because fewer people are left to record it. A decision-maker watching a feed of reported attacks can therefore watch a city get quieter on the page while it gets deadlier on the ground.

The gap. Existing humanitarian severity measures are built to rank crises against each other, usually monthly and usually at the level of a whole country or a whole emergency. They are not built to answer the question a siege poses, which is a question about a single place on a single day: is it, today, sufficiently dangerous here that the obligations owed to these civilians are plainly engaged? Nor do the standard measures put any weight on whether promises of safe passage were kept, which in a siege is often the decisive variable.

The response. This repository proposes a daily, place-specific severity signal assembled entirely from information that was publicly available at the time: recorded conflict events, temperature, satellite damage assessments, population estimates, and the documented record of what corridors were announced and what became of them. It then attaches each of the six ingredients to the legal provision it bears on, so that a rise in the number can be read as a rise in the strength of a specific legal claim rather than as a general worsening.

The research question the repository sets itself is stated at the top of its own interface: how can and should artificial intelligence tools be used to enhance the protection and evacuation of civilians during armed conflict, and in what ways do they support or challenge the implementation of existing obligations under International Humanitarian Law? The model is offered as one empirical answer, worked through a single case.

03

Objectives

The model is designed to:

  • Produce a transparent daily measure of how dangerous it was for civilians to remain in Mariupol on any given date between 5 March and 20 May 2022.
  • Give equal analytical standing to the slow harms of a siege, cold, deprivation, and the destruction of shelter, alongside the fast harms of shelling, so that a lull in attacks cannot by itself lower the assessment.
  • Turn the question of consent into a measurable quantity, by scoring how many announced corridors were honoured and on whose terms passage was permitted.
  • Anchor each component to the International Humanitarian Law provision it bears on, so that the model measures the factual predicate for an obligation rather than pronouncing on the obligation itself.
  • Generate, for any selected date, a written situation assessment naming the corridor regime in force, the dominant driver of severity, and the legal provisions engaged.
  • Make every step of the calculation inspectable, including the arithmetic, so that a reader who disagrees with a threshold can see exactly what changing it would do.

04

How the Model Works

The model belongs to a family of tools called composite indicators: measures that take several unlike quantities and squash them into a single score. The standard reference manual for this kind of work, the Handbook on Constructing Composite Indicators published jointly by the OECD and the European Commission's Joint Research Centre in 2008, is cited by the repository and shapes its approach.

The first problem such a tool must solve is that its ingredients are measured in different units. Attacks per day, degrees Celsius, and percentage of buildings destroyed cannot be added together as they stand. The solution used here is called min-max normalisation: for each ingredient the author fixes a floor and a ceiling, and then rescales the raw observation so that the floor becomes 0 and the ceiling becomes 1. A value of 0 means no concern on that measure. A value of 1 means as bad as the model's scale allows, and the interface marks such a value as saturated.

The ceilings are therefore not neutral. They are judgements about what counts as the worst case, and where they are set determines how much any given ingredient can contribute. The repository is explicit that its ceilings are documented conventions rather than statistically derived values, and that they should be re-estimated as percentiles of an observed distribution before publication.

Component one, hostility intensity. The number of violent events recorded inside the city on an average day: battles, shelling, air and drone strikes, and explosions or remote violence. The raw number comes from the Armed Conflict Location and Event Data Project, known as ACLED, an independent non-profit organisation that reads news reports, non-governmental organisation reports, and local sources and converts them into a structured record of individual violent incidents. The daily event count is divided by a ceiling of 10 events per day. It does very little in the Mariupol case: the highest daily rate in the author's extraction is 4.8 events per day, during 10 to 15 March, which yields a score of 0.48. It is never the dominant driver of severity on any day of the siege window, and the repository treats that as a finding.

Component two, kinetic proximity. How close the armed activity was to where civilians lived. The number of armed actions per day recorded within 5 kilometres of the city, plus one half of the number recorded within 5 to 15 kilometres, all divided by a ceiling of 6. The halving is what the methodology calls distance-decay weighting: fighting fifteen kilometres away is a smaller threat to a civilian in the city than fighting five kilometres away, but it is not no threat at all, because it shapes where the front is going and whether routes out remain usable. Two cities can record the same number of attacks per day while one has them at its edges and the other in its centre, so the repository keeps closeness visible in its own right. It peaks at 0.556 during the 10 to 15 March window and never becomes dominant.

Component three, protection risk. The model's most distinctive component, and the one that carries most of its legal argument. It is the plain average of four sub-indicators, each scored from 0 to 1 and each given equal weight. The first two describe how the violence was conducted: the civilian-targeting share, the proportion of all recorded violent events whose target was civilian, which reaches 0.483 in the 10 to 15 March window; and lethality, deaths per recorded event divided by a ceiling of 15, which reaches 12.9 deaths per event in the same window for a score of 0.86. The methodology identifies that window as the retaliation phase and treats the combination of high civilian targeting and high lethality as its empirical signature.

Corridor violation history. The third sub-indicator, and the first of the model's two original contributions. It is the proportion of previously announced corridors on the relevant route that were shelled or blocked. During the first regime this sits at 1.00, because both prior attempts had collapsed. It falls to 0.67 during the self-evacuation period, to 0.60 during the organised evacuation, and rises again to 0.70 at the end.

Consent and filtration exposure. The fourth sub-indicator is an ordered ladder describing the legal and political basis on which people were moving, running as follows.

  • 0.25, negotiated with a neutral third-party escort.
  • 0.40, negotiated bilaterally between the parties.
  • 0.50, announced unilaterally by one party.
  • 0.75, no ceasefire in effect at all.
  • 0.90, movement subject to screening by the adversary, the process reported as filtration.

Choosing a destination controlled by the adversary raises the effective value to 0.90 or above.

Two of the four sub-indicators measure what was done to civilians; two measure what was promised to them and whether it held. The repository states that the last two operationalise the consent conditions attached to evacuation and to the passage of relief under the Fourth Geneva Convention and Additional Protocol I. The practical effect is that a corridor which exists on paper but not in fact raises the severity score rather than lowering it. That is the correct direction of travel, and it is not how a simple access indicator would behave. The four sub-indicators are averaged without weighting, which is an explicit choice rather than an oversight, but it is a choice: the proportion of attacks aimed at civilians and the legal basis of a corridor are treated as equally informative about protection risk.

Component four, cold burden. How far the temperature fell below a habitable minimum. It is the shortfall of the mean daily temperature below a threshold of 18 degrees Celsius, divided by 28, with a floor at minus 10 degrees. At 18 degrees or above the component scores 0; at minus 10 degrees or below it scores 1. Outdoor temperature is used as a proxy for indoor temperature, which is the component's crucial and defensible assumption: district heating, electricity and water supply in Mariupol were destroyed from around 2 March, and in a city where the heating has stopped, the inside of a building converges on the temperature outside it.

Cold matters a great deal here, because the siege ran through the end of a Ukrainian winter. On the model's own temperature series the coldest point is around minus 6 degrees in the second week of March. Cold is the dominant component on 23 of the 77 days in the window, more than any of the three violence components achieve on any day. A purely kinetic assessment would have registered none of this.

One note on the 18 degree figure. The repository attributes this threshold to the Sphere Handbook. A full-text check of the fourth edition, published in 2018, does not locate an 18 degree indoor or habitability figure anywhere in it. The comparable published figure in the sector is UNHCR's Emergency Handbook, which gives a comfortable shelter interior as 15 to 19 degrees Celsius, and the World Health Organization's housing guidance, which uses 18 degrees as a healthy indoor minimum. The threshold itself is therefore well within the range the sector uses; the attribution appears to need correction.

Component five, the deprivation clock. Time. Specifically, the number of days the city has been encircled, counted from 2 March, divided by a ceiling of 60 days. The definition subtracts three days from the count for each aid convoy that reached the city, on the reasoning that a delivery of supplies buys back some of the accumulated deficit. In the Mariupol case the number of convoys that reached the city proper is recorded as zero, so this relief credit never operates. The repository notes that the ICRC convoy was repeatedly blocked at Berdiansk.

The reasoning for a clock is that hunger, thirst, untreated illness, and the exhaustion of medicines are cumulative. They do not require a shot to be fired and they do not improve on their own. A model that only counted incidents would treat a quiet fiftieth day of encirclement as safer than a loud fifth day. The deprivation clock is the model's way of insisting that it is not. It does more work than any other component: it is the dominant driver on 54 of the 77 days, it reaches its ceiling of 1.0 on 1 May, and it remains saturated for the rest of the window. Because of how the six components are combined, this single saturated component is enough on its own to hold the whole model in its top severity phase. This is both the model's sharpest argument and its most significant vulnerability.

Component six, infrastructure damage. The cumulative share of the city's structures visibly damaged or destroyed, taken from satellite damage assessments published by UNOSAT, the United Nations Satellite Centre, which operates within the United Nations Institute for Training and Research. UNOSAT analysts compare before-and-after satellite images and classify individual buildings as destroyed, severely damaged, moderately damaged or possibly damaged. The Ukraine work sits under a single activation code, CE20220223UKR. UNOSAT publishes on particular dates, not every day, so the model fixes five anchor points and draws a straight line between them: 0.02 on 5 March, 0.04 on 14 March, 0.14 on 26 March, 0.32 on 7 May, and 0.33 on 20 May.

The repository says plainly that this is a lower bound. Satellites see roofs. Damage to the inside of a building whose roof is intact is invisible to this method, and UNOSAT itself labels its assessments preliminary and not validated in the field. The true share of unusable buildings on any given day was therefore higher than the number the model uses. The component rises steadily but never becomes dominant, because its ceiling of 1.0 would require essentially the whole city to be visibly destroyed, and the assessed figure reaches only about a third.

Combining the six. Having produced six numbers between 0 and 1, the model combines them with what mathematicians call a generalised power mean, using an exponent of 6. In plain terms the recipe is: raise each of the six scores to the sixth power, take the ordinary average of those six powered numbers, then take the sixth root of the result. The effect of raising numbers to a high power before averaging is that large values dominate small ones very heavily. The answer always lands somewhere between the plain average of the six and the single worst of the six, and with an exponent of 6 it lands close to the worst. The repository calls this the humanitarian weakest-link principle.

The technical name for what is at stake here is compensability, meaning the extent to which a good score on one measure is allowed to cancel out a terrible score on another. A plain average is fully compensatory: three comfortable measures will drag a catastrophic fourth up toward the middle. The OECD and Joint Research Centre handbook treats this as an explicit decision the designer of any index must make and defend, and the INFORM Severity Index makes a similar choice for similar reasons, using geometric rather than arithmetic averaging to limit compensability. This model goes considerably further in the same direction than INFORM does, and the repository is right to flag its exponent as an adjustable convention rather than a discovered fact.

One consequence deserves to be stated in plain terms, because it is not obvious from the formula and it matters for how the results should be read. If any single one of the six components reaches its ceiling of 1.0, the combined score cannot fall below approximately 0.742 no matter what the other five are doing. Since the threshold for the model's top severity phase is 0.70, one saturated component is by itself sufficient to place the city in the highest category. From 1 May onward the deprivation clock is saturated, and the model's output is effectively pinned by that fact alone.

The repository states, correctly and importantly, that the sixth powers are an internal weighting device and are not shares of causation. The number 0.33 for infrastructure damage, raised to the sixth power, becomes 0.0013 and effectively disappears from the sum. That does not mean the destruction of a third of the city mattered little. It means this particular arithmetic is designed to report the worst thing happening, and everything that is not the worst thing recedes.

The five severity phases. The resulting score is placed in one of five bands, each carrying a stated posture.

PhaseScore rangeLabelIndicated posture
1below 0.20Minimalroutine monitoring
20.20 to 0.40Concerncontingency planning; register vulnerable groups
30.40 to 0.55Seriousevacuation planning; negotiate access
40.55 to 0.70Highevacuation warranted
50.70 and aboveCriticalimmediate protective action

The five-band shape is borrowed deliberately. The INFORM Severity Index, run by ACAPS with the European Commission's Joint Research Centre as scientific lead, sorts every humanitarian crisis in the world into five severity categories. The Integrated Food Security Phase Classification, the partnership of United Nations agencies and non-governmental organisations that classifies food crises, uses a five-phase ladder that humanitarian readers know well. Adopting the same shape makes the output legible to people who already work with those systems. The thresholds themselves, however, are the author's own conventions. They were not derived from data and they have not been validated by expert elicitation. The repository says so.

Vulnerability weighting. The final step adjusts for who was in the city. The weight is 1 plus 0.3 times the share of the population who were children or elderly, plus 0.3 times the share living with a disability, with a pre-siege population of 343,598, children numbering 32,926, elderly 73,723, and people with disabilities approximately 24,288. These figures yield a vulnerability weight of 1.114, and the priority index is simply the severity score multiplied by that weight.

Vw = 1 + 0.3 x (children + elderly) / N + 0.3 x (disabled / N)

N is the pre-siege population of 343,598. The figures above give Vw of 1.114.

The population figures come from WorldPop, a research programme based at the University of Southampton that estimates how many people live in each small square of ground by taking official census totals for large areas and distributing them using satellite evidence about where buildings are. The boundaries used are from GADM, a free global database of administrative area outlines. The disability share of 7.07 per cent is taken from Ministry of Social Policy pension-system statistics as reported by the World Bank.

The repository is candid that the 0.3 increments are placeholders pending calibration. They are chosen to mirror the categories of person given special protection under the Fourth Geneva Convention and under Article 11 of the Convention on the Rights of Persons with Disabilities, which requires States to take all necessary measures to ensure the protection and safety of persons with disabilities in situations of armed conflict. The weight is applied as a multiplier on exposure rather than added in as a seventh component, mirroring the INFORM Severity separation between how bad the threat is and how many vulnerable people it falls on. The repository calls 1.114 conservative, because the World Bank source itself notes that true disability prevalence in Ukraine is likely closer to the World Health Organization's international rate of around 16 per cent than to the 7.07 per cent recorded in the pension system.

Two honest observations about this step. The repository notes that elderly and disabled populations overlap and that the same people are therefore counted twice; it flags this as needing correction where joint distributions exist. And, more consequentially for reading the results, the vulnerability weight is a single constant applied to all 77 days. Within this one case it never changes the ordering of anything. Its purpose is comparison between cases in the fuller decision framework the repository envisages, which has not been built.

05

A Worked Example: 14 March 2022

The repository works through a single day in full, and it is worth following because it demonstrates the model's central argument better than any description. 14 March 2022 was the day the first private cars got out of Mariupol and reached Zaporizhzhia. The six components on that day were:

ComponentCalculationScore
Hostility intensity4.8 events per day divided by 100.480
Kinetic proximity(2.17 plus half of 2.33) divided by 60.556
Protection riskaverage of 0.483, 0.860, 0.670, 0.7500.691
Cold burden(18 minus minus 5.3) divided by 280.832
Deprivation clock12 days divided by 600.200
Infrastructure damageUNOSAT anchor for 14 March0.040

The plain average of these six is 0.47. The worst of the six is 0.83. The weakest-link combination gives 0.66, which places the day in Phase 4 of 5, High, with the indicated posture that evacuation is warranted. Multiplied by the vulnerability weight, the priority index is 0.73.

The reading the repository draws from this is the important part. On the day the first cars escaped, the gravest single threat to the people still inside was not shelling. It was cold: minus 5 degrees, no heating, day twelve of encirclement. A kinetic assessment, watching only the tempo of attacks, would have missed it entirely.

06

Reading the Results

The tool is a single web page. The user selects a date, by dragging along a timeline, stepping day by day, or jumping to one of the four corridor regimes or to a documented event. The page then shows, for that date: which corridor regime was in force and where its route ran on a schematic map of the front line; a satellite view; the observed values of every variable that feeds the model; the six component scores with a short sentence explaining each one; the composite severity with its phase and posture; and a generated written assessment naming the legal provisions engaged.

Three presentational choices deserve mention. The interface prints a single driver sentence identifying which component is currently worst, so that the reader is told not only how bad it is but why. It draws a small number line showing the plain average, the model's rating, and the worst component side by side, which makes the effect of the weakest-link method visible rather than hidden. And the full arithmetic, including every sixth power and the running sum, is available behind a toggle, so a sceptical reader can check the sum by hand.

Recomputing the model exactly as published produces the following trajectory across the 77 days. The series opens on 5 March at 0.51, Phase 3, Serious. It rises through the first week and enters Phase 4, evacuation warranted, on 9 March. It reaches a local peak of 0.675 on 13 March, driven by the combination of the coldest part of the window and the retaliation phase in the protection-risk component. It then falls away through the second half of March, reaching a trough of about 0.38 on 26 March, which is Phase 2, Concern. It falls because the weather improves and because the recorded tempo of attacks drops sharply after 15 March.

From late March it climbs again, steadily and without interruption, as the deprivation clock accumulates. It re-enters Phase 4 in mid-April, crosses 0.70 into Phase 5, Critical, on 28 April, and reaches 0.743 on 1 May, where it stays, flat, until the window closes on 20 May.

Two things follow from this shape, and both are worth stating carefully. The first supports the repository's argument. The threshold at which the model says evacuation is warranted was first crossed on 9 March. The organised evacuation involving the United Nations and the International Committee of the Red Cross began on 30 April. That is a gap of roughly seven and a half weeks between the modelled signal and the mechanism, and the substance of the repository's central finding survives the recomputation.

The second qualifies it. The repository's README describes severity as crossing critical thresholds in mid-March, roughly eight weeks before the mechanism arrived. On the model's own five-phase scale, what happens in mid-March is entry into Phase 4, High, not Phase 5, Critical. The model's Critical phase is not reached until 28 April, and the global maximum of the whole series is at its very end, not in March. The mid-March feature is a local peak, and it is a real and meaningful one, but the summary wording somewhat overstates what the model outputs.

It is also worth noting that across all 77 days the dominant component is either cold burden, on 23 days, or the deprivation clock, on 54 days. On no day of the siege is the worst-scoring component one of the three that measure violence. That is a striking result and it is the model's substantive point made arithmetically. It is also, in part, an artefact of where the ceilings were set: intensity was divided by 10 when the observed maximum was 4.8, while the deprivation clock was divided by 60 days in a siege that ran for more than 60 days. Both readings are true at once, and an honest use of this model requires holding both.

07

Where the Numbers Come From

This section matters more than its length suggests, because the phrase daily model can be read in two quite different ways, and only one of them is accurate here.

The repository does not contain a dataset. Its data folder and its imagery folder hold instructions for downloading optional extra layers, not the layers themselves. Every number the model actually computes with is written directly into the single web page as a fixed constant.

The violence figures come from the author's own extraction from ACLED for Mariupol raion, but they are not daily figures. They are summarised into four windows, and held constant inside each window. Phase I covers 24 February to 9 March, Phase II covers 10 to 15 March, Phase III covers 16 March to 5 May, and Phase IV, after 5 May, is labelled extrapolated in the repository's own tables. Every day inside a window carries the same event rate, the same civilian targeting share, the same lethality, and the same proximity figures.

The temperature series is not an ERA5 extraction. ERA5 is the European Centre for Medium-Range Weather Forecasts reanalysis, a reconstruction of past weather that fills every gap in space and time by feeding decades of real observations through a modern forecasting model, and it would supply an hourly record for Mariupol. The repository names ERA5 as its intended source and states plainly that the current series is an interpolated climatology, a smooth curve drawn through ten anchor temperatures, pending the full extraction. The infrastructure damage series is likewise interpolated between five UNOSAT anchor points.

The consequence, stated plainly. The model produces a value for every day, but its inputs move on far fewer than 77 occasions. Day-to-day variation within a phase comes only from the temperature curve, the deprivation clock ticking forward by one day, and the damage line advancing by a fraction. The word daily describes the resolution of the output, not the resolution of the evidence. A policy reader should hold this firmly in mind: the shape of the curve is real and is defensible, but the model cannot distinguish 18 March from 19 March on any evidence about what happened on those two days.

There are optional layers. If a user downloads the UNOSAT building-damage geodata from the Humanitarian Data Exchange, the humanitarian data-sharing platform run by the United Nations Office for the Coordination of Humanitarian Affairs, the tool will display real per-building assessments and filter them by date. Until that file is supplied, the per-building colouring is generated from a district-level chronology and is labelled illustrative throughout the interface. The interface also offers building footprints from OpenStreetMap, the volunteer map of the world, retrieved through its Overpass query service, and dated satellite crops the user can export from the Copernicus Browser, the European Space Agency's free viewer for Sentinel-2 imagery. The repository warns explicitly that the undated base satellite mosaic must not be presented as the state of the city on any particular date, which is a scrupulous piece of labelling.

08

The Corridor Record

The model divides the siege window into four regimes, each with its own scores for corridor violation history and consent exposure. These are the two variables that carry the model's legal argument, so it is worth setting the regimes out as the repository defines them.

Announced corridors that failed, 5 to 13 March. Violation history 1.00, consent 0.50. The repository records that corridors announced for 5 to 7 March collapsed under renewed shelling, with each party attributing responsibility to the other, and that a proposal on 7 March routing corridors toward Russia and Belarus was rejected by Kyiv. The consent value of 0.50 reflects unilateral announcement rather than negotiated agreement.

Self-evacuation, 14 March to 29 April. Violation history 0.67, consent 0.75. The repository describes private vehicles departing toward Zaporizhzhia without any ceasefire in effect, crossing Russian-held territory and multiple checkpoints between Manhush and Vasylivka, with passage intermittently permitted and convoys turned back on some days. The consent value of 0.75 is the model's score for movement occurring with no agreement at all. Eastbound movement into Russian-controlled territory via Bezimenne is recorded as continuing in parallel throughout.

Organised humanitarian evacuation, 30 April to 7 May. Violation history 0.60, consent 0.25. The repository records that following the United Nations Secretary-General's meetings in Moscow on 26 April, an arrangement involving the United Nations and the International Committee of the Red Cross took civilians out of the Azovstal plant, with evacuees exiting through a Russian-controlled screening point at Bezimenne before onward transfer to Zaporizhzhia, and some proceeding to Russian-controlled territory rather than to Zaporizhzhia. The consent value of 0.25 is the model's lowest, reserved for passage negotiated with a neutral third-party escort.

No organised corridor, 8 to 20 May. Violation history 0.70, consent 0.90. The repository records that no agreed civilian passage was in effect and that remaining movement was subject to filtration screening, citing reporting by the Office of the United Nations High Commissioner for Human Rights and its Human Rights Monitoring Mission in Ukraine, including reported onward transfers to the Russian Federation.

The repository adds one observation that deserves emphasis, because it is the reason the whole exercise has a point. The macro front line in the Zaporizhzhia sector was static throughout the siege window. The routes out did not become unsafe because the battlefield shifted underneath them. They were unsafe, or usable, according to whether passage was permitted. In the repository's phrase, corridor insecurity was a consent problem across stable territory, not a shifting-battlefield problem.

All conduct summaries above are the repository's, attributed by it to reporting by the Office for the Coordination of Humanitarian Affairs, the International Committee of the Red Cross, and the Office of the High Commissioner for Human Rights. They are descriptive, and the repository is careful to state that they are not adjudicative.

09

Grounding in International Humanitarian Law

The model does not decide legal questions. What it claims to do is measure the factual predicate on which certain obligations turn: whether the danger was imminent, and whether objects indispensable to survival were being denied. That is a modest and appropriate framing, and it is the right one.

Fourth Geneva Convention, Article 17. The parties to a conflict shall endeavour to conclude local agreements for the removal from besieged or encircled areas of the wounded, sick, infirm and aged, children and maternity cases, and for the passage into such areas of medical personnel and equipment and ministers of religion. Two points of precision matter, and the repository would be stronger for stating them. The obligation is to endeavour, that is, to make a good-faith effort to reach an agreement; it is an obligation of conduct, not of result. And the categories named are narrower than civilians in general. Article 17 is the provision the repository points to as the mechanism that arrived late, and that identification is apt, but Article 17 standing alone is a comparatively weak legal hook.

Fourth Geneva Convention, Article 23. Each party shall allow the free passage of medical and hospital stores intended for civilians, and of essential foodstuffs, clothing and tonics intended for children under fifteen, expectant mothers and maternity cases. Passage may be made conditional where there are serious reasons to fear diversion, ineffective monitoring, or definite military advantage to the adversary. The narrowness of the food limb, which does not cover the general adult civilian population, is a known gap in the 1949 text.

Additional Protocol I, Article 70. Relief actions that are humanitarian and impartial shall be undertaken for civilian populations not adequately provided for, subject to the agreement of the parties concerned, and parties shall allow and facilitate the rapid and unimpeded passage of relief consignments even where destined for the adverse party's civilians. The dominant reading of the consent requirement is that it may not be withheld arbitrarily. Article 70 substantially closes the gap left by Article 23.

Additional Protocol I, Articles 51, 57 and 58. Article 51 prohibits attacks on civilians and prohibits indiscriminate attacks, including those expected to cause civilian harm excessive in relation to the concrete and direct military advantage anticipated. Article 57 requires constant care to spare civilians, verification of targets, choice of means that minimise civilian harm, cancellation of attacks that turn out to be unlawful, and effective advance warning where circumstances permit. Article 58 places duties on the defending party to move its own civilians away from military objectives so far as feasible, expressly without prejudice to the prohibition on forcible transfers.

Fourth Geneva Convention, Article 49. Forcible transfers and deportations of protected persons out of occupied territory are prohibited regardless of motive. Evacuation is permitted only where the security of the population or imperative military reasons demand it, must stay within the occupied territory unless materially impossible, and must be reversed as soon as hostilities in the area cease. This is the provision the repository invokes in connection with documented filtration transfers, and it is the correct one.

Customary International Humanitarian Law, Rules 15, 24, 53, 54 and 55. As codified in the study published by the International Committee of the Red Cross: constant care and feasible precautions in military operations; removal of civilians from the vicinity of military objectives so far as feasible; the prohibition on starvation of the civilian population as a method of warfare; the prohibition on attacking, destroying, removing or rendering useless objects indispensable to civilian survival; and the duty to allow and facilitate rapid and unimpeded passage of impartial humanitarian relief, subject to a right of control.

The legal background against which all of this sits, and which the repository assumes without stating, is that a siege is not unlawful merely because it is a siege. The International Committee of the Red Cross puts it directly: sieges are not prohibited as such under International Humanitarian Law. What makes a siege unlawful is how it is conducted, and specifically whether starvation is used as a method of warfare, whether objects indispensable to survival are attacked or rendered useless, whether relief is arbitrarily refused, whether attacks within the encirclement are indiscriminate or disproportionate, and whether the required precautions are taken. The model's six components map onto precisely that list, which is the strongest argument in its favour as a design.

The distinctive legal move the model makes is to score consent. Under Article 17 the parties must endeavour to agree; under Article 70 and Rule 55 relief must be allowed and facilitated, subject to a genuine but limited right of control. These are obligations about willingness, and willingness has historically been assessed in prose rather than in numbers. The corridor violation and consent sub-indicators are an attempt to measure the distance between what was announced and what was honoured. Whether the specific values on the ladder are right is arguable, and the repository says they are conventions. That the gap is worth measuring at all seems to this reviewer correct.

10

Methodological Choices

Why the worst condition dominates. The choice of exponent 6 is the single most consequential decision in the model, and the repository defends it on substantive rather than technical grounds: in a siege, the things that kill people do not average out. The defence is sound. The magnitude of the exponent is a separate question from its direction, and the repository treats it as adjustable.

Why vulnerability multiplies rather than adds. Being a child in a freezing city does not add a further hazard; it changes how the existing hazards land. The model therefore applies vulnerability as a multiplier on exposure rather than averaging it in as a seventh component, which mirrors the structure of the INFORM Severity Index.

Why the ordinal consent ladder rather than a binary. A corridor is not simply open or closed. It can be announced by one side and not the other, negotiated between the parties, escorted by a neutral third party, or nominally available but subject to screening by an adversary at the far end. Each of those states carries a different level of protection, and the five-step ladder is a reasonable, if rough, way of ordering them.

Why the model was built retrospectively on a closed case. Using a siege whose outcome is known allows every input to be checked against published sources and every claim to be traced. It also, as the repository acknowledges, means the design has been shaped by knowledge that a real-time user would not have.

Where the thresholds come from. From the author's judgement, informed by published conventions. Not from data, and not yet from expert consensus. The repository states this without hedging, which is to its credit.

11

Limitations and Caveats

The repository sets out seven limitations of its own. They are real limitations, not token ones, and the summary below preserves them while adding four observations that emerged from reviewing the code and recomputing the series.

The evidence degrades exactly when it is needed most. Event reporting from inside a besieged city collapses at the point of greatest severity, because the people who would report have left, died, or lost communications. Every count of attacks in this model is a count of reported attacks. ACLED itself advises that fatality figures are the most biased and least accurate component of conflict reporting and recommends using measures other than fatalities to assess intensity, which is a direct caution about the lethality sub-indicator.

The model is retrospective. It was built knowing how the siege ended. A real-time deployment would face uncertainty this prototype does not.

It has never been validated against behaviour. Modelled severity has not been compared to observed departure flows, or to any other record of what people actually did. There is no test in this repository of whether a high score corresponds to anything.

Judgement is embedded in the numbers. The normalisation ceilings, the exponent of 6, the values on the consent ladder, the 0.3 vulnerability increments, and the five phase thresholds are all conventions chosen by the author. Each is defensible and each is adjustable, and the repository says they should be re-estimated as distribution percentiles before publication.

It is one axis of three. Severity measures how dangerous it was to stay. It says nothing about whether an evacuation was operationally possible, or whether the place people would be moved to was safe. Those two axes, feasibility and destination viability, are described in the repository but not implemented. A reader must not treat a high severity score as a recommendation to move people.

The tool is dual use. The repository raises this itself, and it is the most serious ethical point in the document. The same fusion of event data, population data and route information that helps prioritise an evacuation could support targeting or screening. Movement data about civilians under siege is protection-sensitive by nature.

Infrastructure damage is a lower bound. And the per-building display is illustrative until the real UNOSAT geodata is supplied.

To these the following are added from this review.

The daily resolution is nominal. The violence inputs change on three occasions across 77 days. This is the caveat most likely to be missed by a non-technical reader looking at a smooth daily curve.

A single saturated component pins the result. Because any one component at 1.0 forces the composite to approximately 0.742, and the Critical threshold is 0.70, the deprivation clock alone holds the model in its top phase for the last three weeks of the window regardless of anything else. The model is making a defensible substantive claim there, that fifty-odd days of encirclement without relief is by itself a critical condition. But the reader should understand that after 1 May the number is reporting the passage of time and nothing more, and that the choice of a 60-day ceiling is what makes this happen when it happens.

The ceilings determine which component wins. Intensity is divided by 10 against an observed maximum of 4.8, so it can never exceed 0.48. The deprivation clock is divided by 60 in a siege lasting longer than 60 days, so it necessarily saturates. Under a weakest-link rule, the component with the most generous ceiling relative to its observed range will dominate almost by construction. That the violence components never lead is therefore partly a substantive finding and partly a consequence of the scaling, and the two should not be conflated.

The vulnerability weight carries no information within this case. As a single constant applied to all 77 days, it rescales the output without changing any comparison. It would begin to do work only in the cross-case framework the repository has not yet built.

The relief credit is never exercised. Because no convoy reached the city, the three-day credit in the deprivation clock never operates. It is therefore an untested part of the design.

Above all, the repository is explicit and repeated on the central point: this is a research prototype, its outputs are indicative, and it is not operational guidance. Nothing in this report should be read as suggesting otherwise.

12

Points a Reviewer Should Verify

This section exists because the repository is a working paper at version 0.9 and would benefit from a short list of things to check before it goes further. None of these undermines the model's argument; all are the kind of thing that a reviewer or an examiner will find.

The temperature threshold citation. The 18 degree threshold is attributed to the Sphere Handbook. A full-text check of the 2018 fourth edition does not find an 18 degree indoor or habitability figure. The nearest published sector figures are UNHCR's Emergency Handbook, which gives 15 to 19 degrees Celsius for a comfortable shelter interior, and the World Health Organization's housing guidance, which uses 18 degrees. The threshold looks well chosen; the citation appears to need correcting.

The March UNOSAT anchor. The anchor for 14 March is given as 773 of 17,594 structures, about 4 per cent, for a combined Livoberezhnyi and Zhovtnevyi area of interest. The published UNOSAT assessment using imagery of 14 March 2022 reports 433 of 9,279 structures damaged in Livoberezhnyi District, about 5 per cent. The May figure in the repository, 5,647 structures and about 32 per cent, matches the published updated assessment exactly. The March figure should be traced back to its specific product.

The README wording. The README summarises severity as crossing critical thresholds in mid-March. On the model's own five-phase scale it crosses into Phase 4, High, on 9 March and does not reach Phase 5, Critical, until 28 April. The claim about the timing gap survives either way, but the wording and the chart should agree.

The citation file. It still contains placeholder fields for the author's surname and for the repository address.

The date range. The repository's own summary uses an en dash to write the date range as March to May. Where the work is prepared for publication in a United Nations or similar house style, ranges are normally written out.

13

Practical Nature and Intended Audience

The tool is one web page. It can be published as a static website or served from a folder on a laptop. It requires an internet connection for the base map layers but no installation, no account, and no server. That design choice makes it easy to circulate and easy to archive, which for a piece of research about a contested event is a real advantage: what a reader sees is exactly what the author wrote.

The audience is a mixed one, which the design reflects. The written assessment and the driver sentence are aimed at humanitarian coordinators and political decision-makers. The legal anchors are aimed at legal advisers and accountability practitioners. The visible arithmetic is aimed at reviewers who want to disagree with a threshold and see immediately what changes.

Its most likely honest use today is as a teaching and argument instrument. It is a good way to show a room of people why a lull in shelling is not improvement, why the destination of a corridor is a protection variable rather than a logistical detail, and why the question of when an obligation crystallised can be asked with dates attached.

14

Conclusion

The contribution this repository makes is not a number. It is an argument carried out in arithmetic: that the harms of a siege are not additive, that the worst condition should govern, and that whether a promise of safe passage was kept is a measurable fact about protection rather than background colour. Each of those claims is contestable. Each is stated here in a form clear enough to be contested, which is more than most such tools allow.

The model's own numbers say that the conditions the law responds to were plainly present in Mariupol from the second week of March 2022, and that the mechanism the law envisages arrived at the end of April. The repository draws from this the conclusion that the constraint was consent rather than information. That conclusion does not follow from the arithmetic alone, and it should not be presented as though it does. But the arithmetic makes it a question that can be asked precisely, with a date attached, and against evidence that was available at the time to anyone who cared to assemble it.

The prototype is modest about itself in all the right places. Its weights are conventions, its data are coarser than its daily output suggests, it has never been validated against behaviour, and it measures only one of the three things a real evacuation decision requires. Its author says all of this without being asked. The proper next steps are the ones the repository already names: re-estimate the bounds against observed distributions, complete the ERA5 and UNOSAT extractions, submit the thresholds to expert elicitation, and build the two missing axes. Until then it should be read as what it says it is, which is a careful and honest piece of work in progress.

References

Sources

  1. 01ACLED. Armed Conflict Location and Event Data Project, conflict events and fatalities for Mariupol raion.
  2. 02OECD and the Joint Research Centre. (2008). Handbook on Constructing Composite Indicators. OECD Publishing.
  3. 03ACAPS and the Joint Research Centre of the European Commission. INFORM Severity Index.
  4. 04Integrated Food Security Phase Classification. IPC five-phase severity classification.
  5. 05UNITAR and UNOSAT. Satellite-derived damage assessments for Ukraine, activation CE20220223UKR.
  6. 06United Nations Office for the Coordination of Humanitarian Affairs. Humanitarian Data Exchange, UNOSAT building damage geodata for Ukraine.
  7. 07WorldPop, University of Southampton. Gridded population estimates.
  8. 08GADM. Database of Global Administrative Areas.
  9. 09ECMWF and the Copernicus Climate Change Service. ERA5 reanalysis, named by the repository as its intended temperature source.
  10. 10European Space Agency. Copernicus Browser, Sentinel-2 imagery exports.
  11. 11OpenStreetMap contributors. Building footprints retrieved through the Overpass query service.
  12. 12International Committee of the Red Cross. Article 17, Geneva Convention (IV) relative to the Protection of Civilian Persons in Time of War, 1949.
  13. 13International Committee of the Red Cross. Article 23, Geneva Convention (IV), 1949.
  14. 14International Committee of the Red Cross. Article 49, Geneva Convention (IV), 1949.
  15. 15International Committee of the Red Cross. Articles 51, 57, 58 and 70, Protocol Additional to the Geneva Conventions of 12 August 1949, and relating to the Protection of Victims of International Armed Conflicts (Protocol I), 1977.
  16. 16Henckaerts, J.M. and Doswald-Beck, L. Customary International Humanitarian Law. International Committee of the Red Cross, Rules 15, 24, 53, 54 and 55.
  17. 17United Nations. Article 11, Convention on the Rights of Persons with Disabilities, situations of risk and humanitarian emergencies.
  18. 18Office of the United Nations High Commissioner for Human Rights and the Human Rights Monitoring Mission in Ukraine. Reporting on filtration screening and onward transfers.
  19. 19Sphere Association. The Sphere Handbook, fourth edition, 2018. Cited by the repository for the 18 degree threshold; see Section 12.
  20. 20UNHCR. Emergency Handbook, comfortable shelter interior of 15 to 19 degrees Celsius.
  21. 21World Health Organization. WHO Housing and Health Guidelines, 2018, healthy indoor minimum of 18 degrees Celsius.
  22. 22World Bank. Ukraine disability prevalence as recorded in Ministry of Social Policy pension-system statistics.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file +

Publications · Academic report

The Mariupol Corridor Severity Model

A Daily Measure of Civilian Danger During the Siege of Mariupol, March to May 2022

Ethical Tech CoLab

NYU Center for Global Affairs

July 2026

Christine, GitHub user ChristineLumen. Prepared as masters research at the NYU Center for Global Affairs, under the Ethical Tech CoLab. Repository version 0.9.

A siege does not kill people at a steady rate, and it does not kill them only by shelling. Some days the shells fall. Other days the guns are quiet, and what is killing people is that there has been no heating for three weeks, no running water, no medicine, and no convoy has come through. From a distance, the quiet days can look like improvement. This model produces one number for every day of the siege of Mariupol, built so that a lull in shelling cannot by itself lower the assessment.

Key figures

77

days scored, one severity number for every day from 5 March to 20 May 2022

54

of those days on which the deprivation clock was the dominant driver, against none for the three violence components

0.742

the composite score any single saturated component forces, above the 0.70 threshold for the Critical phase

7.5

weeks between 9 March, when the model first says evacuation is warranted, and the negotiated mechanism arriving on 30 April

00

Foreword

A siege does not kill people at a steady rate, and it does not kill them only by shelling. Some days the shells fall and the count of the dead rises. Other days the guns are comparatively quiet, and what is killing people is that there has been no heating for three weeks, no running water, no medicine, and no convoy has come through. From the outside, and especially from the outside at a distance, the quiet days can look like improvement.

Between March and May of 2022 the city of Mariupol was encircled. Corridors for civilians to leave were announced, and then did not hold. Weeks passed. A negotiated evacuation involving the United Nations and the International Committee of the Red Cross eventually took people out of the Azovstal steelworks at the very end of April. The question this repository asks is a narrow and uncomfortable one: on which day did the facts on the ground first become bad enough that the legal obligations owed to those civilians were plainly engaged, and how long afterwards did anything actually arrive?

This report explains, in non-technical language, what the model measures, how it turns observations into a single daily number, why each of its ingredients was chosen, and what it cannot do. It is written for readers who work in policy, humanitarian operations, or law, and who have never written a line of computer code.

01

Executive Summary

This repository contains a research prototype: an experimental piece of software that calculates, for every single day of the siege of Mariupol from 5 March to 20 May 2022, a number between 0 and 1 describing how dangerous it was for civilians to remain in the city. The repository calls this number the composite severity, written as S.

The severity number is built from six ingredients, which the repository calls components. Three of them describe violence: how many attacks happened each day, how close the fighting was to where people lived, and how much of that violence was aimed at civilians or broke promises of safe passage. Three of them describe the slower forms of harm that a siege produces: the cold in a city whose heating has been destroyed, the accumulating time under encirclement with no aid arriving, and the progressive destruction of the buildings people shelter in.

The six ingredients are deliberately not averaged in the ordinary way. The model uses a method of combination that pulls the answer close to whichever ingredient is worst. The reasoning is stated plainly in the repository's methodology: a population that is safe from shelling but freezing without water is not in medium danger. A plain average would have made the lull in shelling after mid-March look like recovery. It was not.

The resulting daily number is sorted into one of five phases, from Minimal to Critical, each carrying an indicated posture ranging from routine monitoring to immediate protective action. The presentation deliberately follows the five-category convention used by the INFORM Severity Index and by the Integrated Food Security Phase Classification, so that humanitarian readers recognise the shape of the output.

The severity number is then multiplied by a vulnerability weight of 1.114, derived from the share of the pre-siege population who were children, elderly, or living with a disability, to produce what the repository calls a priority index.

The repository's central claim is a claim about timing. It argues that publicly available information was sufficient to establish, well before the end of March, that the factual conditions triggering obligations under International Humanitarian Law were met, and that the negotiated evacuation mechanism did not arrive until 30 April. The repository states its conclusion in one sentence: the binding constraint was not information but consent.

The tool is a single web page that runs in an ordinary browser. Its author describes it as retrospective, not validated against how people actually behaved, and explicitly not operational guidance. It is the first of three intended axes of a fuller decision framework; the other two, feasibility and destination viability, are not built.

This report finds the model's reasoning clear and unusually well documented, and its candour about its own weaknesses genuine. It also identifies several points where a reviewer should verify a citation or a data figure before the work is published or relied upon.

02

Background and Rationale

The problem. During a siege, the people who might intervene are almost always working from partial information, and the information that is available degrades exactly when conditions are worst. Reporters leave or are killed. Communications fail. The number of recorded incidents falls, not because less is happening but because fewer people are left to record it. A decision-maker watching a feed of reported attacks can therefore watch a city get quieter on the page while it gets deadlier on the ground.

The gap. Existing humanitarian severity measures are built to rank crises against each other, usually monthly and usually at the level of a whole country or a whole emergency. They are not built to answer the question a siege poses, which is a question about a single place on a single day: is it, today, sufficiently dangerous here that the obligations owed to these civilians are plainly engaged? Nor do the standard measures put any weight on whether promises of safe passage were kept, which in a siege is often the decisive variable.

The response. This repository proposes a daily, place-specific severity signal assembled entirely from information that was publicly available at the time: recorded conflict events, temperature, satellite damage assessments, population estimates, and the documented record of what corridors were announced and what became of them. It then attaches each of the six ingredients to the legal provision it bears on, so that a rise in the number can be read as a rise in the strength of a specific legal claim rather than as a general worsening.

The research question the repository sets itself is stated at the top of its own interface: how can and should artificial intelligence tools be used to enhance the protection and evacuation of civilians during armed conflict, and in what ways do they support or challenge the implementation of existing obligations under International Humanitarian Law? The model is offered as one empirical answer, worked through a single case.

03

Objectives

The model is designed to:

  • Produce a transparent daily measure of how dangerous it was for civilians to remain in Mariupol on any given date between 5 March and 20 May 2022.
  • Give equal analytical standing to the slow harms of a siege, cold, deprivation, and the destruction of shelter, alongside the fast harms of shelling, so that a lull in attacks cannot by itself lower the assessment.
  • Turn the question of consent into a measurable quantity, by scoring how many announced corridors were honoured and on whose terms passage was permitted.
  • Anchor each component to the International Humanitarian Law provision it bears on, so that the model measures the factual predicate for an obligation rather than pronouncing on the obligation itself.
  • Generate, for any selected date, a written situation assessment naming the corridor regime in force, the dominant driver of severity, and the legal provisions engaged.
  • Make every step of the calculation inspectable, including the arithmetic, so that a reader who disagrees with a threshold can see exactly what changing it would do.

04

How the Model Works

The model belongs to a family of tools called composite indicators: measures that take several unlike quantities and squash them into a single score. The standard reference manual for this kind of work, the Handbook on Constructing Composite Indicators published jointly by the OECD and the European Commission's Joint Research Centre in 2008, is cited by the repository and shapes its approach.

The first problem such a tool must solve is that its ingredients are measured in different units. Attacks per day, degrees Celsius, and percentage of buildings destroyed cannot be added together as they stand. The solution used here is called min-max normalisation: for each ingredient the author fixes a floor and a ceiling, and then rescales the raw observation so that the floor becomes 0 and the ceiling becomes 1. A value of 0 means no concern on that measure. A value of 1 means as bad as the model's scale allows, and the interface marks such a value as saturated.

The ceilings are therefore not neutral. They are judgements about what counts as the worst case, and where they are set determines how much any given ingredient can contribute. The repository is explicit that its ceilings are documented conventions rather than statistically derived values, and that they should be re-estimated as percentiles of an observed distribution before publication.

Component one, hostility intensity. The number of violent events recorded inside the city on an average day: battles, shelling, air and drone strikes, and explosions or remote violence. The raw number comes from the Armed Conflict Location and Event Data Project, known as ACLED, an independent non-profit organisation that reads news reports, non-governmental organisation reports, and local sources and converts them into a structured record of individual violent incidents. The daily event count is divided by a ceiling of 10 events per day. It does very little in the Mariupol case: the highest daily rate in the author's extraction is 4.8 events per day, during 10 to 15 March, which yields a score of 0.48. It is never the dominant driver of severity on any day of the siege window, and the repository treats that as a finding.

Component two, kinetic proximity. How close the armed activity was to where civilians lived. The number of armed actions per day recorded within 5 kilometres of the city, plus one half of the number recorded within 5 to 15 kilometres, all divided by a ceiling of 6. The halving is what the methodology calls distance-decay weighting: fighting fifteen kilometres away is a smaller threat to a civilian in the city than fighting five kilometres away, but it is not no threat at all, because it shapes where the front is going and whether routes out remain usable. Two cities can record the same number of attacks per day while one has them at its edges and the other in its centre, so the repository keeps closeness visible in its own right. It peaks at 0.556 during the 10 to 15 March window and never becomes dominant.

Component three, protection risk. The model's most distinctive component, and the one that carries most of its legal argument. It is the plain average of four sub-indicators, each scored from 0 to 1 and each given equal weight. The first two describe how the violence was conducted: the civilian-targeting share, the proportion of all recorded violent events whose target was civilian, which reaches 0.483 in the 10 to 15 March window; and lethality, deaths per recorded event divided by a ceiling of 15, which reaches 12.9 deaths per event in the same window for a score of 0.86. The methodology identifies that window as the retaliation phase and treats the combination of high civilian targeting and high lethality as its empirical signature.

Corridor violation history. The third sub-indicator, and the first of the model's two original contributions. It is the proportion of previously announced corridors on the relevant route that were shelled or blocked. During the first regime this sits at 1.00, because both prior attempts had collapsed. It falls to 0.67 during the self-evacuation period, to 0.60 during the organised evacuation, and rises again to 0.70 at the end.

Consent and filtration exposure. The fourth sub-indicator is an ordered ladder describing the legal and political basis on which people were moving, running as follows.

  • 0.25, negotiated with a neutral third-party escort.
  • 0.40, negotiated bilaterally between the parties.
  • 0.50, announced unilaterally by one party.
  • 0.75, no ceasefire in effect at all.
  • 0.90, movement subject to screening by the adversary, the process reported as filtration.

Choosing a destination controlled by the adversary raises the effective value to 0.90 or above.

Two of the four sub-indicators measure what was done to civilians; two measure what was promised to them and whether it held. The repository states that the last two operationalise the consent conditions attached to evacuation and to the passage of relief under the Fourth Geneva Convention and Additional Protocol I. The practical effect is that a corridor which exists on paper but not in fact raises the severity score rather than lowering it. That is the correct direction of travel, and it is not how a simple access indicator would behave. The four sub-indicators are averaged without weighting, which is an explicit choice rather than an oversight, but it is a choice: the proportion of attacks aimed at civilians and the legal basis of a corridor are treated as equally informative about protection risk.

Component four, cold burden. How far the temperature fell below a habitable minimum. It is the shortfall of the mean daily temperature below a threshold of 18 degrees Celsius, divided by 28, with a floor at minus 10 degrees. At 18 degrees or above the component scores 0; at minus 10 degrees or below it scores 1. Outdoor temperature is used as a proxy for indoor temperature, which is the component's crucial and defensible assumption: district heating, electricity and water supply in Mariupol were destroyed from around 2 March, and in a city where the heating has stopped, the inside of a building converges on the temperature outside it.

Cold matters a great deal here, because the siege ran through the end of a Ukrainian winter. On the model's own temperature series the coldest point is around minus 6 degrees in the second week of March. Cold is the dominant component on 23 of the 77 days in the window, more than any of the three violence components achieve on any day. A purely kinetic assessment would have registered none of this.

One note on the 18 degree figure. The repository attributes this threshold to the Sphere Handbook. A full-text check of the fourth edition, published in 2018, does not locate an 18 degree indoor or habitability figure anywhere in it. The comparable published figure in the sector is UNHCR's Emergency Handbook, which gives a comfortable shelter interior as 15 to 19 degrees Celsius, and the World Health Organization's housing guidance, which uses 18 degrees as a healthy indoor minimum. The threshold itself is therefore well within the range the sector uses; the attribution appears to need correction.

Component five, the deprivation clock. Time. Specifically, the number of days the city has been encircled, counted from 2 March, divided by a ceiling of 60 days. The definition subtracts three days from the count for each aid convoy that reached the city, on the reasoning that a delivery of supplies buys back some of the accumulated deficit. In the Mariupol case the number of convoys that reached the city proper is recorded as zero, so this relief credit never operates. The repository notes that the ICRC convoy was repeatedly blocked at Berdiansk.

The reasoning for a clock is that hunger, thirst, untreated illness, and the exhaustion of medicines are cumulative. They do not require a shot to be fired and they do not improve on their own. A model that only counted incidents would treat a quiet fiftieth day of encirclement as safer than a loud fifth day. The deprivation clock is the model's way of insisting that it is not. It does more work than any other component: it is the dominant driver on 54 of the 77 days, it reaches its ceiling of 1.0 on 1 May, and it remains saturated for the rest of the window. Because of how the six components are combined, this single saturated component is enough on its own to hold the whole model in its top severity phase. This is both the model's sharpest argument and its most significant vulnerability.

Component six, infrastructure damage. The cumulative share of the city's structures visibly damaged or destroyed, taken from satellite damage assessments published by UNOSAT, the United Nations Satellite Centre, which operates within the United Nations Institute for Training and Research. UNOSAT analysts compare before-and-after satellite images and classify individual buildings as destroyed, severely damaged, moderately damaged or possibly damaged. The Ukraine work sits under a single activation code, CE20220223UKR. UNOSAT publishes on particular dates, not every day, so the model fixes five anchor points and draws a straight line between them: 0.02 on 5 March, 0.04 on 14 March, 0.14 on 26 March, 0.32 on 7 May, and 0.33 on 20 May.

The repository says plainly that this is a lower bound. Satellites see roofs. Damage to the inside of a building whose roof is intact is invisible to this method, and UNOSAT itself labels its assessments preliminary and not validated in the field. The true share of unusable buildings on any given day was therefore higher than the number the model uses. The component rises steadily but never becomes dominant, because its ceiling of 1.0 would require essentially the whole city to be visibly destroyed, and the assessed figure reaches only about a third.

Combining the six. Having produced six numbers between 0 and 1, the model combines them with what mathematicians call a generalised power mean, using an exponent of 6. In plain terms the recipe is: raise each of the six scores to the sixth power, take the ordinary average of those six powered numbers, then take the sixth root of the result. The effect of raising numbers to a high power before averaging is that large values dominate small ones very heavily. The answer always lands somewhere between the plain average of the six and the single worst of the six, and with an exponent of 6 it lands close to the worst. The repository calls this the humanitarian weakest-link principle.

The technical name for what is at stake here is compensability, meaning the extent to which a good score on one measure is allowed to cancel out a terrible score on another. A plain average is fully compensatory: three comfortable measures will drag a catastrophic fourth up toward the middle. The OECD and Joint Research Centre handbook treats this as an explicit decision the designer of any index must make and defend, and the INFORM Severity Index makes a similar choice for similar reasons, using geometric rather than arithmetic averaging to limit compensability. This model goes considerably further in the same direction than INFORM does, and the repository is right to flag its exponent as an adjustable convention rather than a discovered fact.

One consequence deserves to be stated in plain terms, because it is not obvious from the formula and it matters for how the results should be read. If any single one of the six components reaches its ceiling of 1.0, the combined score cannot fall below approximately 0.742 no matter what the other five are doing. Since the threshold for the model's top severity phase is 0.70, one saturated component is by itself sufficient to place the city in the highest category. From 1 May onward the deprivation clock is saturated, and the model's output is effectively pinned by that fact alone.

The repository states, correctly and importantly, that the sixth powers are an internal weighting device and are not shares of causation. The number 0.33 for infrastructure damage, raised to the sixth power, becomes 0.0013 and effectively disappears from the sum. That does not mean the destruction of a third of the city mattered little. It means this particular arithmetic is designed to report the worst thing happening, and everything that is not the worst thing recedes.

The five severity phases. The resulting score is placed in one of five bands, each carrying a stated posture.

PhaseScore rangeLabelIndicated posture
1below 0.20Minimalroutine monitoring
20.20 to 0.40Concerncontingency planning; register vulnerable groups
30.40 to 0.55Seriousevacuation planning; negotiate access
40.55 to 0.70Highevacuation warranted
50.70 and aboveCriticalimmediate protective action

The five-band shape is borrowed deliberately. The INFORM Severity Index, run by ACAPS with the European Commission's Joint Research Centre as scientific lead, sorts every humanitarian crisis in the world into five severity categories. The Integrated Food Security Phase Classification, the partnership of United Nations agencies and non-governmental organisations that classifies food crises, uses a five-phase ladder that humanitarian readers know well. Adopting the same shape makes the output legible to people who already work with those systems. The thresholds themselves, however, are the author's own conventions. They were not derived from data and they have not been validated by expert elicitation. The repository says so.

Vulnerability weighting. The final step adjusts for who was in the city. The weight is 1 plus 0.3 times the share of the population who were children or elderly, plus 0.3 times the share living with a disability, with a pre-siege population of 343,598, children numbering 32,926, elderly 73,723, and people with disabilities approximately 24,288. These figures yield a vulnerability weight of 1.114, and the priority index is simply the severity score multiplied by that weight.

Vw = 1 + 0.3 x (children + elderly) / N + 0.3 x (disabled / N)

N is the pre-siege population of 343,598. The figures above give Vw of 1.114.

The population figures come from WorldPop, a research programme based at the University of Southampton that estimates how many people live in each small square of ground by taking official census totals for large areas and distributing them using satellite evidence about where buildings are. The boundaries used are from GADM, a free global database of administrative area outlines. The disability share of 7.07 per cent is taken from Ministry of Social Policy pension-system statistics as reported by the World Bank.

The repository is candid that the 0.3 increments are placeholders pending calibration. They are chosen to mirror the categories of person given special protection under the Fourth Geneva Convention and under Article 11 of the Convention on the Rights of Persons with Disabilities, which requires States to take all necessary measures to ensure the protection and safety of persons with disabilities in situations of armed conflict. The weight is applied as a multiplier on exposure rather than added in as a seventh component, mirroring the INFORM Severity separation between how bad the threat is and how many vulnerable people it falls on. The repository calls 1.114 conservative, because the World Bank source itself notes that true disability prevalence in Ukraine is likely closer to the World Health Organization's international rate of around 16 per cent than to the 7.07 per cent recorded in the pension system.

Two honest observations about this step. The repository notes that elderly and disabled populations overlap and that the same people are therefore counted twice; it flags this as needing correction where joint distributions exist. And, more consequentially for reading the results, the vulnerability weight is a single constant applied to all 77 days. Within this one case it never changes the ordering of anything. Its purpose is comparison between cases in the fuller decision framework the repository envisages, which has not been built.

05

A Worked Example: 14 March 2022

The repository works through a single day in full, and it is worth following because it demonstrates the model's central argument better than any description. 14 March 2022 was the day the first private cars got out of Mariupol and reached Zaporizhzhia. The six components on that day were:

ComponentCalculationScore
Hostility intensity4.8 events per day divided by 100.480
Kinetic proximity(2.17 plus half of 2.33) divided by 60.556
Protection riskaverage of 0.483, 0.860, 0.670, 0.7500.691
Cold burden(18 minus minus 5.3) divided by 280.832
Deprivation clock12 days divided by 600.200
Infrastructure damageUNOSAT anchor for 14 March0.040

The plain average of these six is 0.47. The worst of the six is 0.83. The weakest-link combination gives 0.66, which places the day in Phase 4 of 5, High, with the indicated posture that evacuation is warranted. Multiplied by the vulnerability weight, the priority index is 0.73.

The reading the repository draws from this is the important part. On the day the first cars escaped, the gravest single threat to the people still inside was not shelling. It was cold: minus 5 degrees, no heating, day twelve of encirclement. A kinetic assessment, watching only the tempo of attacks, would have missed it entirely.

06

Reading the Results

The tool is a single web page. The user selects a date, by dragging along a timeline, stepping day by day, or jumping to one of the four corridor regimes or to a documented event. The page then shows, for that date: which corridor regime was in force and where its route ran on a schematic map of the front line; a satellite view; the observed values of every variable that feeds the model; the six component scores with a short sentence explaining each one; the composite severity with its phase and posture; and a generated written assessment naming the legal provisions engaged.

Three presentational choices deserve mention. The interface prints a single driver sentence identifying which component is currently worst, so that the reader is told not only how bad it is but why. It draws a small number line showing the plain average, the model's rating, and the worst component side by side, which makes the effect of the weakest-link method visible rather than hidden. And the full arithmetic, including every sixth power and the running sum, is available behind a toggle, so a sceptical reader can check the sum by hand.

Recomputing the model exactly as published produces the following trajectory across the 77 days. The series opens on 5 March at 0.51, Phase 3, Serious. It rises through the first week and enters Phase 4, evacuation warranted, on 9 March. It reaches a local peak of 0.675 on 13 March, driven by the combination of the coldest part of the window and the retaliation phase in the protection-risk component. It then falls away through the second half of March, reaching a trough of about 0.38 on 26 March, which is Phase 2, Concern. It falls because the weather improves and because the recorded tempo of attacks drops sharply after 15 March.

From late March it climbs again, steadily and without interruption, as the deprivation clock accumulates. It re-enters Phase 4 in mid-April, crosses 0.70 into Phase 5, Critical, on 28 April, and reaches 0.743 on 1 May, where it stays, flat, until the window closes on 20 May.

Two things follow from this shape, and both are worth stating carefully. The first supports the repository's argument. The threshold at which the model says evacuation is warranted was first crossed on 9 March. The organised evacuation involving the United Nations and the International Committee of the Red Cross began on 30 April. That is a gap of roughly seven and a half weeks between the modelled signal and the mechanism, and the substance of the repository's central finding survives the recomputation.

The second qualifies it. The repository's README describes severity as crossing critical thresholds in mid-March, roughly eight weeks before the mechanism arrived. On the model's own five-phase scale, what happens in mid-March is entry into Phase 4, High, not Phase 5, Critical. The model's Critical phase is not reached until 28 April, and the global maximum of the whole series is at its very end, not in March. The mid-March feature is a local peak, and it is a real and meaningful one, but the summary wording somewhat overstates what the model outputs.

It is also worth noting that across all 77 days the dominant component is either cold burden, on 23 days, or the deprivation clock, on 54 days. On no day of the siege is the worst-scoring component one of the three that measure violence. That is a striking result and it is the model's substantive point made arithmetically. It is also, in part, an artefact of where the ceilings were set: intensity was divided by 10 when the observed maximum was 4.8, while the deprivation clock was divided by 60 days in a siege that ran for more than 60 days. Both readings are true at once, and an honest use of this model requires holding both.

07

Where the Numbers Come From

This section matters more than its length suggests, because the phrase daily model can be read in two quite different ways, and only one of them is accurate here.

The repository does not contain a dataset. Its data folder and its imagery folder hold instructions for downloading optional extra layers, not the layers themselves. Every number the model actually computes with is written directly into the single web page as a fixed constant.

The violence figures come from the author's own extraction from ACLED for Mariupol raion, but they are not daily figures. They are summarised into four windows, and held constant inside each window. Phase I covers 24 February to 9 March, Phase II covers 10 to 15 March, Phase III covers 16 March to 5 May, and Phase IV, after 5 May, is labelled extrapolated in the repository's own tables. Every day inside a window carries the same event rate, the same civilian targeting share, the same lethality, and the same proximity figures.

The temperature series is not an ERA5 extraction. ERA5 is the European Centre for Medium-Range Weather Forecasts reanalysis, a reconstruction of past weather that fills every gap in space and time by feeding decades of real observations through a modern forecasting model, and it would supply an hourly record for Mariupol. The repository names ERA5 as its intended source and states plainly that the current series is an interpolated climatology, a smooth curve drawn through ten anchor temperatures, pending the full extraction. The infrastructure damage series is likewise interpolated between five UNOSAT anchor points.

The consequence, stated plainly. The model produces a value for every day, but its inputs move on far fewer than 77 occasions. Day-to-day variation within a phase comes only from the temperature curve, the deprivation clock ticking forward by one day, and the damage line advancing by a fraction. The word daily describes the resolution of the output, not the resolution of the evidence. A policy reader should hold this firmly in mind: the shape of the curve is real and is defensible, but the model cannot distinguish 18 March from 19 March on any evidence about what happened on those two days.

There are optional layers. If a user downloads the UNOSAT building-damage geodata from the Humanitarian Data Exchange, the humanitarian data-sharing platform run by the United Nations Office for the Coordination of Humanitarian Affairs, the tool will display real per-building assessments and filter them by date. Until that file is supplied, the per-building colouring is generated from a district-level chronology and is labelled illustrative throughout the interface. The interface also offers building footprints from OpenStreetMap, the volunteer map of the world, retrieved through its Overpass query service, and dated satellite crops the user can export from the Copernicus Browser, the European Space Agency's free viewer for Sentinel-2 imagery. The repository warns explicitly that the undated base satellite mosaic must not be presented as the state of the city on any particular date, which is a scrupulous piece of labelling.

08

The Corridor Record

The model divides the siege window into four regimes, each with its own scores for corridor violation history and consent exposure. These are the two variables that carry the model's legal argument, so it is worth setting the regimes out as the repository defines them.

Announced corridors that failed, 5 to 13 March. Violation history 1.00, consent 0.50. The repository records that corridors announced for 5 to 7 March collapsed under renewed shelling, with each party attributing responsibility to the other, and that a proposal on 7 March routing corridors toward Russia and Belarus was rejected by Kyiv. The consent value of 0.50 reflects unilateral announcement rather than negotiated agreement.

Self-evacuation, 14 March to 29 April. Violation history 0.67, consent 0.75. The repository describes private vehicles departing toward Zaporizhzhia without any ceasefire in effect, crossing Russian-held territory and multiple checkpoints between Manhush and Vasylivka, with passage intermittently permitted and convoys turned back on some days. The consent value of 0.75 is the model's score for movement occurring with no agreement at all. Eastbound movement into Russian-controlled territory via Bezimenne is recorded as continuing in parallel throughout.

Organised humanitarian evacuation, 30 April to 7 May. Violation history 0.60, consent 0.25. The repository records that following the United Nations Secretary-General's meetings in Moscow on 26 April, an arrangement involving the United Nations and the International Committee of the Red Cross took civilians out of the Azovstal plant, with evacuees exiting through a Russian-controlled screening point at Bezimenne before onward transfer to Zaporizhzhia, and some proceeding to Russian-controlled territory rather than to Zaporizhzhia. The consent value of 0.25 is the model's lowest, reserved for passage negotiated with a neutral third-party escort.

No organised corridor, 8 to 20 May. Violation history 0.70, consent 0.90. The repository records that no agreed civilian passage was in effect and that remaining movement was subject to filtration screening, citing reporting by the Office of the United Nations High Commissioner for Human Rights and its Human Rights Monitoring Mission in Ukraine, including reported onward transfers to the Russian Federation.

The repository adds one observation that deserves emphasis, because it is the reason the whole exercise has a point. The macro front line in the Zaporizhzhia sector was static throughout the siege window. The routes out did not become unsafe because the battlefield shifted underneath them. They were unsafe, or usable, according to whether passage was permitted. In the repository's phrase, corridor insecurity was a consent problem across stable territory, not a shifting-battlefield problem.

All conduct summaries above are the repository's, attributed by it to reporting by the Office for the Coordination of Humanitarian Affairs, the International Committee of the Red Cross, and the Office of the High Commissioner for Human Rights. They are descriptive, and the repository is careful to state that they are not adjudicative.

09

Grounding in International Humanitarian Law

The model does not decide legal questions. What it claims to do is measure the factual predicate on which certain obligations turn: whether the danger was imminent, and whether objects indispensable to survival were being denied. That is a modest and appropriate framing, and it is the right one.

Fourth Geneva Convention, Article 17. The parties to a conflict shall endeavour to conclude local agreements for the removal from besieged or encircled areas of the wounded, sick, infirm and aged, children and maternity cases, and for the passage into such areas of medical personnel and equipment and ministers of religion. Two points of precision matter, and the repository would be stronger for stating them. The obligation is to endeavour, that is, to make a good-faith effort to reach an agreement; it is an obligation of conduct, not of result. And the categories named are narrower than civilians in general. Article 17 is the provision the repository points to as the mechanism that arrived late, and that identification is apt, but Article 17 standing alone is a comparatively weak legal hook.

Fourth Geneva Convention, Article 23. Each party shall allow the free passage of medical and hospital stores intended for civilians, and of essential foodstuffs, clothing and tonics intended for children under fifteen, expectant mothers and maternity cases. Passage may be made conditional where there are serious reasons to fear diversion, ineffective monitoring, or definite military advantage to the adversary. The narrowness of the food limb, which does not cover the general adult civilian population, is a known gap in the 1949 text.

Additional Protocol I, Article 70. Relief actions that are humanitarian and impartial shall be undertaken for civilian populations not adequately provided for, subject to the agreement of the parties concerned, and parties shall allow and facilitate the rapid and unimpeded passage of relief consignments even where destined for the adverse party's civilians. The dominant reading of the consent requirement is that it may not be withheld arbitrarily. Article 70 substantially closes the gap left by Article 23.

Additional Protocol I, Articles 51, 57 and 58. Article 51 prohibits attacks on civilians and prohibits indiscriminate attacks, including those expected to cause civilian harm excessive in relation to the concrete and direct military advantage anticipated. Article 57 requires constant care to spare civilians, verification of targets, choice of means that minimise civilian harm, cancellation of attacks that turn out to be unlawful, and effective advance warning where circumstances permit. Article 58 places duties on the defending party to move its own civilians away from military objectives so far as feasible, expressly without prejudice to the prohibition on forcible transfers.

Fourth Geneva Convention, Article 49. Forcible transfers and deportations of protected persons out of occupied territory are prohibited regardless of motive. Evacuation is permitted only where the security of the population or imperative military reasons demand it, must stay within the occupied territory unless materially impossible, and must be reversed as soon as hostilities in the area cease. This is the provision the repository invokes in connection with documented filtration transfers, and it is the correct one.

Customary International Humanitarian Law, Rules 15, 24, 53, 54 and 55. As codified in the study published by the International Committee of the Red Cross: constant care and feasible precautions in military operations; removal of civilians from the vicinity of military objectives so far as feasible; the prohibition on starvation of the civilian population as a method of warfare; the prohibition on attacking, destroying, removing or rendering useless objects indispensable to civilian survival; and the duty to allow and facilitate rapid and unimpeded passage of impartial humanitarian relief, subject to a right of control.

The legal background against which all of this sits, and which the repository assumes without stating, is that a siege is not unlawful merely because it is a siege. The International Committee of the Red Cross puts it directly: sieges are not prohibited as such under International Humanitarian Law. What makes a siege unlawful is how it is conducted, and specifically whether starvation is used as a method of warfare, whether objects indispensable to survival are attacked or rendered useless, whether relief is arbitrarily refused, whether attacks within the encirclement are indiscriminate or disproportionate, and whether the required precautions are taken. The model's six components map onto precisely that list, which is the strongest argument in its favour as a design.

The distinctive legal move the model makes is to score consent. Under Article 17 the parties must endeavour to agree; under Article 70 and Rule 55 relief must be allowed and facilitated, subject to a genuine but limited right of control. These are obligations about willingness, and willingness has historically been assessed in prose rather than in numbers. The corridor violation and consent sub-indicators are an attempt to measure the distance between what was announced and what was honoured. Whether the specific values on the ladder are right is arguable, and the repository says they are conventions. That the gap is worth measuring at all seems to this reviewer correct.

10

Methodological Choices

Why the worst condition dominates. The choice of exponent 6 is the single most consequential decision in the model, and the repository defends it on substantive rather than technical grounds: in a siege, the things that kill people do not average out. The defence is sound. The magnitude of the exponent is a separate question from its direction, and the repository treats it as adjustable.

Why vulnerability multiplies rather than adds. Being a child in a freezing city does not add a further hazard; it changes how the existing hazards land. The model therefore applies vulnerability as a multiplier on exposure rather than averaging it in as a seventh component, which mirrors the structure of the INFORM Severity Index.

Why the ordinal consent ladder rather than a binary. A corridor is not simply open or closed. It can be announced by one side and not the other, negotiated between the parties, escorted by a neutral third party, or nominally available but subject to screening by an adversary at the far end. Each of those states carries a different level of protection, and the five-step ladder is a reasonable, if rough, way of ordering them.

Why the model was built retrospectively on a closed case. Using a siege whose outcome is known allows every input to be checked against published sources and every claim to be traced. It also, as the repository acknowledges, means the design has been shaped by knowledge that a real-time user would not have.

Where the thresholds come from. From the author's judgement, informed by published conventions. Not from data, and not yet from expert consensus. The repository states this without hedging, which is to its credit.

11

Limitations and Caveats

The repository sets out seven limitations of its own. They are real limitations, not token ones, and the summary below preserves them while adding four observations that emerged from reviewing the code and recomputing the series.

The evidence degrades exactly when it is needed most. Event reporting from inside a besieged city collapses at the point of greatest severity, because the people who would report have left, died, or lost communications. Every count of attacks in this model is a count of reported attacks. ACLED itself advises that fatality figures are the most biased and least accurate component of conflict reporting and recommends using measures other than fatalities to assess intensity, which is a direct caution about the lethality sub-indicator.

The model is retrospective. It was built knowing how the siege ended. A real-time deployment would face uncertainty this prototype does not.

It has never been validated against behaviour. Modelled severity has not been compared to observed departure flows, or to any other record of what people actually did. There is no test in this repository of whether a high score corresponds to anything.

Judgement is embedded in the numbers. The normalisation ceilings, the exponent of 6, the values on the consent ladder, the 0.3 vulnerability increments, and the five phase thresholds are all conventions chosen by the author. Each is defensible and each is adjustable, and the repository says they should be re-estimated as distribution percentiles before publication.

It is one axis of three. Severity measures how dangerous it was to stay. It says nothing about whether an evacuation was operationally possible, or whether the place people would be moved to was safe. Those two axes, feasibility and destination viability, are described in the repository but not implemented. A reader must not treat a high severity score as a recommendation to move people.

The tool is dual use. The repository raises this itself, and it is the most serious ethical point in the document. The same fusion of event data, population data and route information that helps prioritise an evacuation could support targeting or screening. Movement data about civilians under siege is protection-sensitive by nature.

Infrastructure damage is a lower bound. And the per-building display is illustrative until the real UNOSAT geodata is supplied.

To these the following are added from this review.

The daily resolution is nominal. The violence inputs change on three occasions across 77 days. This is the caveat most likely to be missed by a non-technical reader looking at a smooth daily curve.

A single saturated component pins the result. Because any one component at 1.0 forces the composite to approximately 0.742, and the Critical threshold is 0.70, the deprivation clock alone holds the model in its top phase for the last three weeks of the window regardless of anything else. The model is making a defensible substantive claim there, that fifty-odd days of encirclement without relief is by itself a critical condition. But the reader should understand that after 1 May the number is reporting the passage of time and nothing more, and that the choice of a 60-day ceiling is what makes this happen when it happens.

The ceilings determine which component wins. Intensity is divided by 10 against an observed maximum of 4.8, so it can never exceed 0.48. The deprivation clock is divided by 60 in a siege lasting longer than 60 days, so it necessarily saturates. Under a weakest-link rule, the component with the most generous ceiling relative to its observed range will dominate almost by construction. That the violence components never lead is therefore partly a substantive finding and partly a consequence of the scaling, and the two should not be conflated.

The vulnerability weight carries no information within this case. As a single constant applied to all 77 days, it rescales the output without changing any comparison. It would begin to do work only in the cross-case framework the repository has not yet built.

The relief credit is never exercised. Because no convoy reached the city, the three-day credit in the deprivation clock never operates. It is therefore an untested part of the design.

Above all, the repository is explicit and repeated on the central point: this is a research prototype, its outputs are indicative, and it is not operational guidance. Nothing in this report should be read as suggesting otherwise.

12

Points a Reviewer Should Verify

This section exists because the repository is a working paper at version 0.9 and would benefit from a short list of things to check before it goes further. None of these undermines the model's argument; all are the kind of thing that a reviewer or an examiner will find.

The temperature threshold citation. The 18 degree threshold is attributed to the Sphere Handbook. A full-text check of the 2018 fourth edition does not find an 18 degree indoor or habitability figure. The nearest published sector figures are UNHCR's Emergency Handbook, which gives 15 to 19 degrees Celsius for a comfortable shelter interior, and the World Health Organization's housing guidance, which uses 18 degrees. The threshold looks well chosen; the citation appears to need correcting.

The March UNOSAT anchor. The anchor for 14 March is given as 773 of 17,594 structures, about 4 per cent, for a combined Livoberezhnyi and Zhovtnevyi area of interest. The published UNOSAT assessment using imagery of 14 March 2022 reports 433 of 9,279 structures damaged in Livoberezhnyi District, about 5 per cent. The May figure in the repository, 5,647 structures and about 32 per cent, matches the published updated assessment exactly. The March figure should be traced back to its specific product.

The README wording. The README summarises severity as crossing critical thresholds in mid-March. On the model's own five-phase scale it crosses into Phase 4, High, on 9 March and does not reach Phase 5, Critical, until 28 April. The claim about the timing gap survives either way, but the wording and the chart should agree.

The citation file. It still contains placeholder fields for the author's surname and for the repository address.

The date range. The repository's own summary uses an en dash to write the date range as March to May. Where the work is prepared for publication in a United Nations or similar house style, ranges are normally written out.

13

Practical Nature and Intended Audience

The tool is one web page. It can be published as a static website or served from a folder on a laptop. It requires an internet connection for the base map layers but no installation, no account, and no server. That design choice makes it easy to circulate and easy to archive, which for a piece of research about a contested event is a real advantage: what a reader sees is exactly what the author wrote.

The audience is a mixed one, which the design reflects. The written assessment and the driver sentence are aimed at humanitarian coordinators and political decision-makers. The legal anchors are aimed at legal advisers and accountability practitioners. The visible arithmetic is aimed at reviewers who want to disagree with a threshold and see immediately what changes.

Its most likely honest use today is as a teaching and argument instrument. It is a good way to show a room of people why a lull in shelling is not improvement, why the destination of a corridor is a protection variable rather than a logistical detail, and why the question of when an obligation crystallised can be asked with dates attached.

14

Conclusion

The contribution this repository makes is not a number. It is an argument carried out in arithmetic: that the harms of a siege are not additive, that the worst condition should govern, and that whether a promise of safe passage was kept is a measurable fact about protection rather than background colour. Each of those claims is contestable. Each is stated here in a form clear enough to be contested, which is more than most such tools allow.

The model's own numbers say that the conditions the law responds to were plainly present in Mariupol from the second week of March 2022, and that the mechanism the law envisages arrived at the end of April. The repository draws from this the conclusion that the constraint was consent rather than information. That conclusion does not follow from the arithmetic alone, and it should not be presented as though it does. But the arithmetic makes it a question that can be asked precisely, with a date attached, and against evidence that was available at the time to anyone who cared to assemble it.

The prototype is modest about itself in all the right places. Its weights are conventions, its data are coarser than its daily output suggests, it has never been validated against behaviour, and it measures only one of the three things a real evacuation decision requires. Its author says all of this without being asked. The proper next steps are the ones the repository already names: re-estimate the bounds against observed distributions, complete the ERA5 and UNOSAT extractions, submit the thresholds to expert elicitation, and build the two missing axes. Until then it should be read as what it says it is, which is a careful and honest piece of work in progress.

References

Sources

  1. 01ACLED. Armed Conflict Location and Event Data Project, conflict events and fatalities for Mariupol raion.
  2. 02OECD and the Joint Research Centre. (2008). Handbook on Constructing Composite Indicators. OECD Publishing.
  3. 03ACAPS and the Joint Research Centre of the European Commission. INFORM Severity Index.
  4. 04Integrated Food Security Phase Classification. IPC five-phase severity classification.
  5. 05UNITAR and UNOSAT. Satellite-derived damage assessments for Ukraine, activation CE20220223UKR.
  6. 06United Nations Office for the Coordination of Humanitarian Affairs. Humanitarian Data Exchange, UNOSAT building damage geodata for Ukraine.
  7. 07WorldPop, University of Southampton. Gridded population estimates.
  8. 08GADM. Database of Global Administrative Areas.
  9. 09ECMWF and the Copernicus Climate Change Service. ERA5 reanalysis, named by the repository as its intended temperature source.
  10. 10European Space Agency. Copernicus Browser, Sentinel-2 imagery exports.
  11. 11OpenStreetMap contributors. Building footprints retrieved through the Overpass query service.
  12. 12International Committee of the Red Cross. Article 17, Geneva Convention (IV) relative to the Protection of Civilian Persons in Time of War, 1949.
  13. 13International Committee of the Red Cross. Article 23, Geneva Convention (IV), 1949.
  14. 14International Committee of the Red Cross. Article 49, Geneva Convention (IV), 1949.
  15. 15International Committee of the Red Cross. Articles 51, 57, 58 and 70, Protocol Additional to the Geneva Conventions of 12 August 1949, and relating to the Protection of Victims of International Armed Conflicts (Protocol I), 1977.
  16. 16Henckaerts, J.M. and Doswald-Beck, L. Customary International Humanitarian Law. International Committee of the Red Cross, Rules 15, 24, 53, 54 and 55.
  17. 17United Nations. Article 11, Convention on the Rights of Persons with Disabilities, situations of risk and humanitarian emergencies.
  18. 18Office of the United Nations High Commissioner for Human Rights and the Human Rights Monitoring Mission in Ukraine. Reporting on filtration screening and onward transfers.
  19. 19Sphere Association. The Sphere Handbook, fourth edition, 2018. Cited by the repository for the 18 degree threshold; see Section 12.
  20. 20UNHCR. Emergency Handbook, comfortable shelter interior of 15 to 19 degrees Celsius.
  21. 21World Health Organization. WHO Housing and Health Guidelines, 2018, healthy indoor minimum of 18 degrees Celsius.
  22. 22World Bank. Ukraine disability prevalence as recorded in Ministry of Social Policy pension-system statistics.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file diff --git a/static-site/print/mariupol-severity-model/index.txt b/static-site/print/mariupol-severity-model/index.txt index 4dba8baf1..38bfabe8f 100644 --- a/static-site/print/mariupol-severity-model/index.txt +++ b/static-site/print/mariupol-severity-model/index.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -12:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","print","mariupol-severity-model",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","mariupol-severity-model","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -14:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -16:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","print","mariupol-severity-model",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","mariupol-severity-model","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:"$Sreact.suspense" -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1c:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -29,7 +29,7 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 19:[] 10:"$W19" 11:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1e:Ta14, @page { size: 210mm 297mm; margin: 20mm 18mm; } @@ -174,6 +174,6 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 5a:["$","li","20",{"children":[["$","span",null,{"className":"print-ref-num","children":"21"}],["$","span",null,{"children":"World Health Organization. WHO Housing and Health Guidelines, 2018, healthy indoor minimum of 18 degrees Celsius."}]]}] 5b:["$","li","21",{"children":[["$","span",null,{"className":"print-ref-num","children":"22"}],["$","span",null,{"children":"World Bank. Ukraine disability prevalence as recorded in Ministry of Social Policy pension-system statistics."}]]}] 1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -5c:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +5c:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 18:null 1d:[["$","title","0",{"children":"The Mariupol Corridor Severity Model (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L5c","6",{}]] diff --git a/static-site/print/provenance-search/__next._full.txt b/static-site/print/provenance-search/__next._full.txt index dc64c06bf..182a8476c 100644 --- a/static-site/print/provenance-search/__next._full.txt +++ b/static-site/print/provenance-search/__next._full.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -12:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","print","provenance-search",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","provenance-search","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -14:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -16:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","print","provenance-search",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","provenance-search","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:"$Sreact.suspense" -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1c:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -29,7 +29,7 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 19:[] 10:"$W19" 11:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1e:Ta14, @page { size: 210mm 297mm; margin: 20mm 18mm; } @@ -153,6 +153,6 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 45:["$","p","11",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"The written rationale is not the arithmetic."}]," ","The plain-language explanation shown beside the score is composed by the language model and describes the substance of the case. It does not narrate the calculation, and a reader should not assume the two are saying the same thing."]}] 46:["$","p","12",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Status."}]," ","This is a research prototype by a single developer, deployed as a public demonstration. It is not an accredited due-diligence service, it carries no professional indemnity, and its passport is not a certificate. Its own attestation says so."]}] 1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -47:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +47:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 18:null 1d:[["$","title","0",{"children":"Provenance Search (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L47","6",{}]] diff --git a/static-site/print/provenance-search/__next._head.txt b/static-site/print/provenance-search/__next._head.txt index 837bdfab3..d1b4b8058 100644 --- a/static-site/print/provenance-search/__next._head.txt +++ b/static-site/print/provenance-search/__next._head.txt @@ -1,6 +1,6 @@ 1:"$Sreact.fragment" -2:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -3:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +2:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +3:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 4:"$Sreact.suspense" -5:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +5:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 0:{"rsc":["$","$1","h",{"children":[null,["$","$L2",null,{"children":[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]]}],["$","div",null,{"hidden":true,"children":["$","$L3",null,{"children":["$","$4",null,{"name":"Next.Metadata","children":[["$","title","0",{"children":"Provenance Search (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L5","6",{}]]}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/print/provenance-search/__next._index.txt b/static-site/print/provenance-search/__next._index.txt index f98a4fd90..18b9d82b6 100644 --- a/static-site/print/provenance-search/__next._index.txt +++ b/static-site/print/provenance-search/__next._index.txt @@ -1,13 +1,13 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -d:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +d:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} 8:["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$Ld",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$Ld",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$Ld",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$Ld",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] diff --git a/static-site/print/provenance-search/__next._tree.txt b/static-site/print/provenance-search/__next._tree.txt index 62ef1d203..932f7c8d3 100644 --- a/static-site/print/provenance-search/__next._tree.txt +++ b/static-site/print/provenance-search/__next._tree.txt @@ -1,7 +1,7 @@ -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] 0:{"tree":{"name":"","param":null,"prefetchHints":16,"slots":{"children":{"name":"print","param":null,"prefetchHints":0,"slots":{"children":{"name":"slug","param":{"type":"d","key":"provenance-search","siblings":null},"prefetchHints":0,"slots":{"children":{"name":"__PAGE__","param":null,"prefetchHints":0,"slots":null}}}}}}},"staleTime":300,"buildId":"static-snapshot"} diff --git a/static-site/print/provenance-search/__next.print.txt b/static-site/print/provenance-search/__next.print.txt index 460afe242..fa617e997 100644 --- a/static-site/print/provenance-search/__next.print.txt +++ b/static-site/print/provenance-search/__next.print.txt @@ -1,4 +1,4 @@ 1:"$Sreact.fragment" -2:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -3:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +2:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +3:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] 0:{"rsc":["$","$1","c",{"children":[null,[["$","style",null,{"children":"\n /* The site is dark-first, and a dark color-scheme makes Chrome paint\n the printed sheet's margin area dark too. Paper is light. */\n html { color-scheme: light !important; }\n html, body { background: #ffffff !important; }\n body::after { content: none !important; }\n header, footer, .site-bg, .aura { display: none !important; }\n "}],["$","$L2",null,{"parallelRouterKey":"children","template":["$","$L3",null,{}]}]]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/print/provenance-search/index.html b/static-site/print/provenance-search/index.html index dcfb4dd0a..28cb4d1fa 100644 --- a/static-site/print/provenance-search/index.html +++ b/static-site/print/provenance-search/index.html @@ -1,4 +1,4 @@ -Provenance Search (print edition) · NYU Ethical Tech CoLab

Publications · Academic report

Provenance Search

An Automated Ownership-History Check for Artworks and Cultural Objects

Ethical Tech CoLab

NYU Center for Global Affairs

July 2026

Developed under the Ethical Tech CoLab at the NYU Center for Global Affairs as part of masters research, 2026. The repository's commit history records a single contributor, working under the identifier yagorocha-web.

Every object in a museum, a saleroom, or a private collection carries an invisible second history alongside its artistic one: the record of who has owned it, when, and how it passed from hand to hand. The problem is not that this information does not exist. It is that it is scattered. Provenance Search assembles what the free public sources say into a single ownership timeline, marks the places where the timeline has holes, and puts a transparent number on how much of the picture is actually supported by retrieved evidence.

Key figures

7

free public sources queried at once, one primary web search and six corroborating references

5

watchlist domains that trigger a high-severity flag by fixed rule, never by model judgment

30

points deducted from the confidence score for each custody gap, the heaviest penalty in the calculation

12

limitations the report states about itself, beginning with the restricted databases it does not search

01

Executive Summary

Provenance Search is a small web application that takes a description of an artwork and returns a structured summary of that artwork's documented ownership history, together with a numerical confidence score and a list of risk flags. It is published as a live public demonstration and is intended for academic and demonstration purposes.

The tool queries seven free public information sources. One of them, a commercial web-search service called Tavily, is treated as the primary research engine and is deliberately restricted to a fixed list of authoritative websites, including INTERPOL, UNESCO, the Getty, the German Lost Art Database, the Central Registry of Information on Looted Cultural Property, the United States Federal Bureau of Investigation, and major auction houses. The remaining six sources, drawn from museum collections and encyclopedic and structured reference data, are used to corroborate and to supply exact dates.

A general-purpose artificial-intelligence model, Google's Gemini, performs two jobs: it can identify an artwork from a photograph, and it arranges the retrieved facts into a chronological ownership timeline. It is instructed to use only facts present in the retrieved material, and to mark any period of ownership that the sources do not account for as an explicit gap.

The most important design decision in the project is that the confidence score is not produced by the artificial-intelligence model. It is calculated by a short, fixed, published arithmetic rule written into the server software. The same set of findings will always produce the same score, and any reader can check the arithmetic.

The tool is candid in its own output about the limits of what it has done. The digital signature attached to each passport states that the record attests to process, not to underlying truth. That framing is accurate and is the right way for a policy reader to understand the entire system.

Provenance Search does not query the restricted law-enforcement and commercial databases that professional due diligence relies on. It reads the public web pages of some of the organisations that maintain those databases. The distinction is central to interpreting its output.

02

Background and Rationale

The problem. The trade in art and cultural objects is among the largest asset markets with comparatively little mandatory disclosure. Ownership history is held in records that were built for different purposes by institutions with no obligation to share, and there is no single register that a buyer, journalist, heir, or customs officer can consult to see the whole chain.

Why gaps matter. Provenance research does not usually produce a clean finding of theft. It produces an absence. An object whose recorded ownership jumps from 1932 to 1948 with nothing in between is not thereby proven to have been looted, but the years between 1933 and 1945 are precisely the window in which Jewish and other persecuted owners across continental Europe were dispossessed by confiscation, forced sale, and sale under duress. An unexplained gap in that period is what a suppressed transfer looks like from the outside. The professional convention, reflected in museum guidance in both the United States and Europe, is to treat such a gap as a trigger for further research rather than as a verdict.

The gap in tooling. The public resources that do exist are good but fragmented, and most require the researcher to know which one to consult and what to type into it. Assembling a first-pass picture across all of them is slow, repetitive clerical work. It is exactly the kind of work that software can usefully do, provided the software is honest about the difference between finding a record and establishing a fact.

The response. Provenance Search automates the first pass. It runs the same query against several sources at once, gathers what comes back, arranges it in date order, names the holes, and puts a transparent number on how much of the picture is actually supported by retrieved evidence. It is presented by its own interface as a tool that flags what cannot be verified rather than one that confirms what can.

Relationship to related work. The repository's package metadata, its earliest commit, and the identifier embedded in every passport signature all carry the name arts and artifacts, which is also the name of a sibling repository in the same organisation. Provenance Search is best understood as the deployed, publicly hosted web version of that line of work. This report describes only what is present in this repository.

03

Objectives

The tool is designed to:

  1. Assemble, from free and public sources only, whatever documented ownership history exists for a named artwork, without requiring the user to hold a subscription to any commercial database.
  2. Present that history as a dated chronological timeline in which every entry carries the source it came from, so that a reader can follow any claim back to its origin.
  3. Report the absence of information as a finding in its own right, rather than presenting an incomplete chain as though it were complete.
  4. Raise an explicit, high-severity alert when the search turns up material on the public sites of the recognised stolen-art and looted-art registries.
  5. Express overall reliability as a single number produced by a fixed published rule, so that the score cannot drift with the mood of a language model and can be audited by anyone.
  6. Work in the setting where the question is most often asked, including on a mobile telephone in a gallery, by allowing the object to be photographed rather than described.

04

How the Tool Works

The system has four stages. A user's request passes through all of them in a single operation that takes a few seconds.

Stage one, describing the object. The user supplies what they know through one of three routes.

  • A text form with fields for title, artist, period or date, medium, and an optional last known sale price.
  • An uploaded photograph of the object.
  • A photograph taken there and then with the device camera, which is the mode intended for use in a museum or a saleroom.

If a photograph is supplied, it is sent to Gemini's image-reading capability, which returns its best guess at title, artist, period, and medium, together with its own self-reported certainty about that identification and a short note. Those values are written into the form, and the search then runs automatically. Only the title and the artist are strictly required to proceed. Large photographs are reduced in size before they are sent, so that a high-resolution telephone image does not exceed the limits of the free service.

Stage two, querying the sources. The title and artist are combined into a search phrase and sent to seven sources at the same time. Each source returns one of three verdicts about the object: a hit was found and it raises no alarm; a hit was found on a registry of lost or stolen property; or nothing matching was found.

Stage three, assembling the timeline. Everything the seven sources returned is gathered into a single block of text and passed to Gemini with a set of written instructions. The model is told that the restricted web search is the primary basis for the timeline and that the museum and reference sources are supplementary corroboration. It is told to use only facts present in the material supplied. It is told that where the sources leave a period of ownership unaccounted for, it must insert an entry marked as a gap with a note explaining what is missing, on the stated principle that a gap is itself a fact worth reporting.

The model is also given one narrow permission to go beyond the retrieved material. For works it recognises as very well documented, where the live sources returned little or nothing, it may fill in widely known ownership history from its own training. Every such entry must be labelled as general knowledge, must be marked unverified, must carry no source link, and may never contradict what a live source actually said. Whenever this permission is used, the software itself adds a medium-severity flag to the result, so the reader sees that part of the timeline rests on the model's memory rather than on a citation.

Stage four, scoring and signing. The assembled timeline and flag list are then handed back to the server's own code, which does three things without any further involvement from the artificial-intelligence model. It adds a high-severity flag for every hit on a registry domain, independently of whether the model noticed it. It computes the confidence score by fixed arithmetic. And it attaches a signature block recording the time of the check and a digital fingerprint, a short string of characters derived mathematically from the title, the artist, and the timestamp, which allows a later reader to detect whether those details have been altered.

05

The Variables Explained

This section is the heart of the report. It sets out every input the tool takes, every rule it applies, and how each one reaches the final result.

Title and artist. These are the only required fields. They are combined into the phrase sent to every source, so they determine everything that follows. A misspelled artist name or a title in the wrong language will quietly produce a thin result rather than an error, which is a real practical caution for users.

Period or date, and medium. These are optional and are not used to filter the searches. They are passed to the model as descriptive context, helping it distinguish between different works that share a title and helping it judge whether a returned record is really the same object.

Last known sale price. This optional figure exists to support a single check. If the user supplies a price, the model is asked whether that price is clearly out of line with a comparable figure actually present in the retrieved sources. If, and only if, such a comparable exists and the supplied price is inconsistent with it, the valuation is marked anomalous. The instruction is deliberately conservative: with no price supplied, or no comparable found, the answer is always no. A price far above or below the plausible market level is a recognised signal in art-market due diligence, since valuation is one of the few numbers that has to be stated openly and is therefore one of the few that can be checked against the record.

Each of the seven sources returns one of three verdicts.

  • Clear means the source found at least one matching record and nothing alarming.
  • Flagged means the restricted web search returned a result hosted on one of the loss and stolen-property registries.
  • Not found means the source returned nothing, or was skipped because no access key was configured, or failed.

Only the primary web search can return the flagged verdict. The museum and reference sources can only ever say clear or not found, because they hold collection catalogues rather than loss reports.

DomainInstitution behind it
interpol.intINTERPOL's stolen works of art database
artloss.comThe Art Loss Register
lostart.deThe German Lost Art Foundation's database
lootedart.comThe Central Registry of Information on Looted Cultural Property
fbi.govThe FBI's stolen art file
The watchlist. If any result returned by the primary web search is hosted on one of these five domains, the software adds a risk flag of type watchlist match at high severity, records which domain it came from, and links to the page.

This rule is deterministic. It runs in the server's own code after the model has finished, it compares the address of each returned page against the list of five, and it does not ask the model's opinion. That is a deliberate safeguard: the single most consequential signal the tool can produce is the one signal that a language model is not permitted to suppress or to invent. It is also the one signal that disappears entirely if the web-search service is not configured, a dependency the repository documents plainly.

PenaltyDeductionWhen it applies
Custody gap30 points eachAny timeline entry the model marked as a period of unaccounted-for ownership
Thin corroboration25 pointsFewer than three of the seven sources returned anything at all
High-severity risk flag10 points eachAutomatic watchlist matches, and any high-severity flag the model raised from the retrieved material, such as a documented forced transfer or an unresolved legal claim
Anomalous valuation10 pointsThe supplied sale price was marked anomalous against a comparable figure found in the sources
The confidence score begins at 100 per cent, is reduced by these four penalties, and is then held within the range of 0 to 100. The result is divided by 100 and reported as a proportion, which the interface displays as a percentage.

The custody-gap penalty is the heaviest in the calculation, and the weighting reflects the professional convention that in provenance work an unexplained break in the chain is the primary warning sign, not a minor blemish. Its severity also means the score falls very fast. Two gaps alone remove 60 points. Three take the score to zero on their own.

The corroboration term measures corroboration rather than content. A finding supported by one source is a lead; a finding that four independent sources recognise is an established record. The threshold of three out of seven is a judgment call by the developer rather than a derived figure. Note also that the test counts any verdict other than not found, so a flagged result counts towards corroboration in the same way a clear result does, on the reasoning that a registry hit still demonstrates that the object is known to the record.

Two properties of this calculation deserve to be stated clearly, because they shape how the number should be interpreted. First, the score measures how well documented the ownership history is, not how likely it is that the object is legitimate. A famous work with a complete and well-known history that includes a documented wartime confiscation will score very low, because that confiscation registers as a break in title and attracts high-severity flags. An obscure object about which almost nothing is known may also score low, because too few sources returned anything. The two cases are very different in substance and can look similar in the number. The written rationale that accompanies the score is intended to distinguish them, and a reader should always read it.

Second, because the penalties are subtractive and large, the score reaches zero easily and then stops. Once it is at zero, further findings do not change it. Zero therefore means at least this bad rather than a measured floor. The repository's own worked example, a demonstration record for Egon Schiele's Portrait of Wally, scores zero for exactly this reason, with several custody gaps and several high-severity flags in combination.

ScoreColour
Below 40 per centRed
40 to 70 per centAmber
70 per cent or aboveGreen
The display bands. These are presentational thresholds only. Nothing in the software behaves differently according to which band a score falls into, and the bands carry no legal or institutional meaning.

06

Reading the Results

The passport. The output is a single structured record. It contains the artwork's details as searched, the confidence score, a short written rationale in plain language explaining what is and is not verified, the ownership timeline, the risk flags, the valuation assessment, the list of sources consulted with each one's verdict, and the signature block.

The timeline. Each entry gives a period, an owner, and where available a note, a source link, and the name of the source authority. Entries that represent gaps are shown with the owner field marked and tagged as a custody gap in the interface. Entries drawn from the model's own knowledge rather than a retrieved source are tagged as general knowledge. A reader can therefore see at a glance which parts of the chain are cited and which are not.

The risk flags. Each flag has a type, a severity of high, medium, or low, a plain-language detail sentence, and where applicable a link. Severity governs the colour of the flag in the interface and, for high-severity flags only, feeds the score.

The sources consulted panel. This lists all seven sources with their verdicts, so that a reader can see not only what was found but where nothing was found. This matters more than it might appear. A not-found verdict from a museum collection means only that the museum does not hold the object. It is not evidence of anything about the object's history.

The signature. Each passport records the identifier of the software version that produced it, the exact time, a digital fingerprint, and an attestation sentence. The attestation states that the passport records the results of automated queries to free public sources and attests to process, not to underlying truth. This is the single most important sentence in the output and should be read as governing everything above it.

07

The Data Sources in Plain Terms

Tavily. A commercial web-search service designed to be used by software rather than by a person browsing. In this project it is the primary research engine, and it is restricted to a fixed list of thirteen websites so that it cannot return results from the open web. The query sent is the title and artist followed by the words provenance, ownership, history, looting, theft, and restitution.

The thirteen permitted domains are:

  • metmuseum.org
  • getty.edu
  • interpol.int
  • unesco.org
  • artloss.com
  • lostart.de
  • lootedart.com
  • christies.com
  • sothebys.com
  • artnet.com
  • fbi.gov
  • ifar.org
  • wikipedia.org

For a reader not familiar with the field: INTERPOL maintains the only global database of police-certified records of stolen cultural objects, publicly searchable since 2021 through its free ID-Art application. The FBI's National Stolen Art File, established in 1997, is a publicly searchable United States register of stolen art and cultural property, populated only by law-enforcement agencies. The German Lost Art Foundation's Lost Art Database records cultural assets seized between 1933 and 1945 as a result of persecution, and objects whose history cannot exclude such a seizure; it is free and public. lootedart.com is the Central Registry of Information on Looted Cultural Property 1933 to 1945, established in 2001 by the Commission for Looted Art in Europe, and holds both a documentary database and an object database.

The Art Loss Register is a private London-based commercial company operating what it describes as the largest private database of stolen art; its data is not publicly accessible and searches are a paid service. The Getty Research Institute's Provenance Index is a large free scholarly resource built from transcribed sale catalogues, dealer stock books, and household inventories, weighted towards Western European art from the sixteenth to the early twentieth century. IFAR, the International Foundation for Art Research, was a New York non-profit founded in 1969 whose provenance guide has long been a standard plain-language reference; it announced in 2024 that it was winding down operations. Christie's, Sotheby's, and Artnet are commercial auction and art-market sources whose catalogue entries frequently include provenance statements.

The Metropolitan Museum of Art. The Met publishes an open interface to its collection catalogue requiring no key. The tool retrieves up to three matching objects and reads their title, artist, date, medium, credit line, and public web address. The credit line is useful because it often names the donor or bequest through which the museum acquired the work.

The Art Institute of Chicago. Also an open collection catalogue requiring no key. This is the only museum source in the set that returns a dedicated provenance text field, and it is therefore the most directly valuable of the three museum sources for the tool's purpose.

The Museum of Modern Art. MoMA publishes no live search facility, and its website blocks automated access, so the project takes a different approach. MoMA's collection is published as an open static dataset on a public code-sharing site. The repository includes a script that downloads that dataset, keeps six fields for each work, and compresses it into a single file of roughly four megabytes covering about 159,000 works. That file is loaded into the server's memory when it starts, so a MoMA search happens instantly and involves no network request at all. The matching rule is simple: every word of the title must appear in the record and at least one word of the artist name longer than two characters must also appear.

Wikipedia and Wikidata. The English-language encyclopedia's search facility is used for general background, returning up to three article summaries. Wikidata, a companion project holding structured facts rather than prose, is searched for the artwork, picks the best-matching entry by looking for the artist's surname in the entry description, and then asks for five specific categories of fact: when the work was made, where it is now, which collections have held it and between which dates, who has owned it and between which dates, and any significant events recorded against it. This is the most precisely targeted of the supplementary sources, because those categories map directly onto the shape of a provenance timeline.

Europeana. A European Union cultural-heritage aggregator that brings together records from thousands of European galleries, libraries, archives, and museums. It requires a free key, and the tool skips it if none is configured.

Gemini. Google's family of large language models, capable of reading images as well as text. It is used for the photograph identification and for the assembly of the timeline. It is not used for the score.

08

The Legal and Ethical Backdrop

The tool does not implement any legal test, and this report does not claim that it does. But the reason its outputs matter is set by a body of international commitments that a policy reader will recognise, and the registries it searches exist because of them.

The Washington Conference Principles on Nazi-Confiscated Art, agreed in December 1998 by 44 states, set out eleven principles including the identification of art confiscated by the Nazis and not subsequently restituted, the opening of archives to researchers, the publicising of unidentified works, and the achievement of just and fair solutions for claims. They are explicitly non-binding: a moral and political commitment rather than a treaty, with no enforcement mechanism. The Terezin Declaration of June 2009 extended the same approach to immovable property, Judaica, and archives, and in March 2024 a set of Best Practices for the Washington Principles was issued for the twenty-fifth anniversary, likewise non-binding, which among other things treats sales under duress between 1933 and 1945 as equivalent to involuntary transfers.

lootedart.com, one of the five watchlist domains, was created specifically to fulfil the sixth of the Washington Principles, which called for a central repository of information. The German Lost Art Database serves the same purpose within Germany's national framework. The tool's watchlist is thus not an arbitrary list of websites; it is a list of the public faces of the institutional infrastructure built in response to those commitments.

For objects other than Nazi-era losses, the governing instrument is the 1970 UNESCO Convention on the Means of Prohibiting and Preventing the Illicit Import, Export and Transfer of Ownership of Cultural Property, which entered into force in April 1972 and now has around 150 states parties. It obliges states to maintain inventories, to operate export certification, to bar museums from acquiring undocumented material, and to return stolen inventoried objects on request. It is not retroactive, binding states only in respect of transfers after it entered into force between the states concerned, which is why 1970 functions in practice as a due-diligence cut-off date for antiquities, and why it does not reach the 1933 to 1945 period. The 1995 UNIDROIT Convention on Stolen or Illegally Exported Cultural Objects, in force since July 1998, was designed to address the private-law gaps the UNESCO Convention leaves open, but has far fewer parties.

It should be stated plainly that the tool applies none of these instruments. It does not determine title, assess a restitution claim, or decide whether an export was lawful. What it does is surface the public traces that the institutions created under these frameworks have left on the open web. Deciding what any of it means remains a matter for lawyers, provenance researchers, and the claims processes those frameworks established.

09

Design Choices

Why the score is not produced by the model. A language model asked to rate its own confidence will produce a number that sounds reasonable and cannot be reproduced or checked. The project instead computes the score in ordinary code from countable facts: how many gaps, how many sources responded, how many high-severity flags, whether the valuation was anomalous. The result is that two runs producing the same findings produce the same score, and a reader who disagrees with the score can identify exactly which term they disagree with. This is the strongest design decision in the project.

Why the primary search is restricted. An unrestricted web search for the words looting and theft alongside a famous artist's name would return a great deal of journalism, speculation, and commercial content. Restricting the search to thirteen named institutional and market domains means that the material the model reasons over comes from sources with an identifiable custodian, at the cost of missing anything held elsewhere.

Why gaps are recorded rather than smoothed over. The instruction given to the model states that a gap is itself a fact worth reporting, and the interface displays gaps prominently in red. A system that quietly produced an unbroken chain wherever it lacked data would be worse than useless in this domain, because the incomplete chain is the finding.

Why the general-knowledge fallback exists and why it is fenced. Without it, the tool would return an almost empty result for the most famous works in the world, since the free sources it uses may hold little ownership detail even for a painting whose history is taught in schools. The fallback lets the model fill those blanks, but it is constrained on four sides: every such entry is labelled in the data, tagged in the display, never marked verified, and never allowed to override a live source, and its use triggers an automatic medium-severity flag. The constraint is well designed. It remains the part of the system where an error is hardest for a non-specialist reader to detect.

Why the MoMA data is bundled rather than queried. Because MoMA offers no live search and blocks automated access, the only lawful and reliable route to its collection is its own published open dataset. Bundling a compressed copy makes the search instant and removes a point of failure, at the cost that the copy is only as current as the last time it was rebuilt.

Why everything runs on the server. The user's browser never contacts any external service directly. It speaks only to this project's own server, which holds the access keys. This keeps the keys out of the browser, where they would be readable by anyone.

10

Limitations and Caveats

The report states twelve limitations. They are reproduced here because they are the most important part of the document for any reader considering what weight to give the tool.

It does not search the restricted databases. This is the most important limitation and the one most likely to be misread. The interface names INTERPOL, the Art Loss Register, Lost Art, and the FBI among the sources it searches. What it actually searches is the public web pages of those organisations, by way of a general web-search service. The Art Loss Register's database is a paid commercial service with no public access at all. A negative result from this tool is therefore not a clearance against the Art Loss Register, and must never be presented as one. INTERPOL and the FBI files are publicly searchable through their own interfaces, but the tool does not query those interfaces directly either.

An absence of findings is not a clean history. Every not-found verdict in the panel means only that the source returned nothing for the phrase that was searched. It carries no information about the object.

The score conflates two different situations. A well-documented history containing a wartime seizure and an obscure object with no history at all can both produce a very low score. The number alone does not distinguish them.

The weights are the developer's judgment. Thirty points for a gap, twenty-five for thin corroboration, ten per high-severity flag, ten for a valuation anomaly, three sources as the corroboration threshold: none of these figures is derived from a study, an expert panel, or a validation exercise against known cases. They are reasonable choices that produce sensible orderings, and they should be described as such rather than as a measurement.

The timeline depends on a language model. The model is instructed to use only retrieved facts, and its temperature setting is kept very low to make its output as consistent as possible, but it is still a language model reading messy source material. It can misread a date, attach a record to the wrong object, or mistake a similarly titled work for the one being searched. Nothing in the system checks its assembly against the retrieved material after the fact.

Identification from a photograph is a guess. The image step returns a best-effort identification with a self-reported certainty figure. That figure is the model's own estimate, it is not carried into the confidence score, and an incorrect identification will produce a fully formed passport for the wrong object.

The searches are simple text matches. The query is the title and artist as typed. There is no handling of alternative titles, transliterations, works known by different names in different languages, or the many objects for which no single agreed title exists. This weighs most heavily against exactly the categories of object where provenance questions are most acute, including antiquities and non-Western material, which frequently have no title and no named artist at all.

The source list is Western-weighted. Two of the three museum sources are American, the aggregator is European, and the market sources are the two large London and New York auction houses. Objects from collections and markets outside that orbit will be under-represented, and this limitation compounds the previous one.

The sources change beneath the tool. The domain list is fixed in the software. One of the thirteen domains, ifar.org, belongs to an organisation that announced in 2024 that it was winding down. Domain lists of this kind require periodic review, and nothing in the repository schedules one.

Dependencies and degradation. If the web-search key is absent, the tool runs on the supplementary sources alone and the watchlist rule cannot fire at all, which removes its single most consequential signal without any visible change in the shape of the output. If the Europeana key is absent, that source is silently skipped and counts as not found, which can push the corroboration count below the threshold and cost 25 points for a reason unrelated to the artwork.

The written rationale is not the arithmetic. The plain-language explanation shown beside the score is composed by the language model and describes the substance of the case. It does not narrate the calculation, and a reader should not assume the two are saying the same thing.

Status. This is a research prototype by a single developer, deployed as a public demonstration. It is not an accredited due-diligence service, it carries no professional indemnity, and its passport is not a certificate. Its own attestation says so.

11

Practical Nature and Deployment

The tool is a single web page backed by a small server program. A user needs only a web browser and no installation. The public demonstration is hosted on a commercial application-hosting platform, and the project's repository page redirects visitors to it.

The repository also contains three stored example results, for Leonardo da Vinci's Salvator Mundi, Vincent van Gogh's The Starry Night, and Egon Schiele's Portrait of Wally. These are saved copies of earlier live runs, included so that the output can be demonstrated without a working connection or a configured key. Their signature blocks are marked as static snapshots. They are teaching material and should not be read as current findings.

The choice of examples is apt. Portrait of Wally is the case that did more than any other to establish that a loan to an American museum could expose an unresolved Nazi-era claim: taken from the Jewish Viennese dealer Lea Bondi Jaray around the time she fled Vienna in 1939, it was seized in New York in 1998 while on loan to the Museum of Modern Art, and after some thirteen years of federal forfeiture litigation the matter settled in July 2010 with the Leopold Museum paying the Bondi estate 19 million United States dollars to retain the painting and agreeing to display its true provenance permanently alongside it. Salvator Mundi, sold at Christie's in November 2017 for 450.3 million dollars including fees, is the case most associated with the opposite problem: a chain of ownership with a long blank stretch, an attribution that remains contested, and a present location that is not publicly confirmed.

All access keys are held in server configuration and are excluded from the repository. The README notes explicitly that if any key file was ever committed or shared, the keys should be rotated. That is the right instruction and it is good practice to have written it down.

12

Intended Audience and Use

The tool is aimed at the person who needs a fast first look: a student researcher, a journalist, a small institution without a provenance department, a family beginning to trace an object, or a visitor standing in front of a work with a question about it. For that audience it does something genuinely useful, which is to run in seconds a set of searches that would otherwise take an afternoon, and to say clearly which parts of the answer are cited and which are not.

It is not aimed at, and should not be used for, the decision points where the answer carries legal or commercial weight. An acquisition, a sale, a restitution claim, an export licence, or a repatriation request all require searches of the restricted registries, examination of physical and archival evidence, and professional judgment. The right way to read a passport from this tool is as a list of leads and of questions that have not been answered.

13

Conclusion

The value of Provenance Search lies less in what it finds than in how carefully it reports what it did not find. Its timeline names its gaps. Its source panel shows silences as well as answers. Its most consequential signal, the registry match, is computed by rule rather than by a language model. Its score can be recomputed by hand from four numbers. Its signature states that it attests to process and not to truth. In a field where an incomplete chain presented as a complete one is the characteristic harm, that discipline is the substance of the contribution.

The prototype's weaknesses are the ones its own design invites. It reads the public faces of registries rather than the registries themselves, and its interface language does not make that distinction as sharply as its documentation does. Its scoring weights are informed guesses. Its search strategy assumes a world of titled works by named artists, which is not the world in which most contested cultural property sits. These are addressable, and naming them is more useful than the tool's current score would be to anyone making a decision.

What the project demonstrates is a pattern worth carrying into other domains: let the automated system gather and arrange, let fixed published rules do the judging, mark every claim with its origin, and treat the absence of evidence as a reportable finding rather than a blank to be filled.

References

Sources

  1. 01Washington Conference Principles on Nazi-Confiscated Art, agreed December 1998 by 44 states. Non-binding.
  2. 02Terezin Declaration on Holocaust Era Assets and Related Issues, June 2009. Best Practices for the Washington Principles, March 2024.
  3. 03UNESCO. Convention on the Means of Prohibiting and Preventing the Illicit Import, Export and Transfer of Ownership of Cultural Property, 1970, in force April 1972.
  4. 04UNIDROIT. Convention on Stolen or Illegally Exported Cultural Objects, 1995, in force July 1998.
  5. 05INTERPOL. Stolen Works of Art Database and the ID-Art application, publicly searchable since 2021.
  6. 06Federal Bureau of Investigation. National Stolen Art File, established 1997.
  7. 07German Lost Art Foundation. Lost Art Database, cultural assets seized between 1933 and 1945.
  8. 08Commission for Looted Art in Europe. Central Registry of Information on Looted Cultural Property 1933 to 1945, established 2001.
  9. 09The Art Loss Register. Private commercial database of stolen art; searches are a paid service.
  10. 10Getty Research Institute. Provenance Index, transcribed sale catalogues, dealer stock books, and household inventories.
  11. 11International Foundation for Art Research. Provenance guide. Founded 1969; announced wind-down in 2024.
  12. 12The Metropolitan Museum of Art. Open Access collection interface.
  13. 13Art Institute of Chicago. Open collection interface, including a dedicated provenance field.
  14. 14Museum of Modern Art. Open collection dataset, roughly 159,000 works.
  15. 15Wikidata. Structured facts on collections, owners, and significant events.
  16. 16Europeana. European Union cultural-heritage aggregator.
  17. 17Tavily. Commercial web-search service for software, used as the restricted primary research engine.
  18. 18Google. Gemini family of large language models, used for image identification and timeline assembly.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file +

Publications · Academic report

Provenance Search

An Automated Ownership-History Check for Artworks and Cultural Objects

Ethical Tech CoLab

NYU Center for Global Affairs

July 2026

Developed under the Ethical Tech CoLab at the NYU Center for Global Affairs as part of masters research, 2026. The repository's commit history records a single contributor, working under the identifier yagorocha-web.

Every object in a museum, a saleroom, or a private collection carries an invisible second history alongside its artistic one: the record of who has owned it, when, and how it passed from hand to hand. The problem is not that this information does not exist. It is that it is scattered. Provenance Search assembles what the free public sources say into a single ownership timeline, marks the places where the timeline has holes, and puts a transparent number on how much of the picture is actually supported by retrieved evidence.

Key figures

7

free public sources queried at once, one primary web search and six corroborating references

5

watchlist domains that trigger a high-severity flag by fixed rule, never by model judgment

30

points deducted from the confidence score for each custody gap, the heaviest penalty in the calculation

12

limitations the report states about itself, beginning with the restricted databases it does not search

01

Executive Summary

Provenance Search is a small web application that takes a description of an artwork and returns a structured summary of that artwork's documented ownership history, together with a numerical confidence score and a list of risk flags. It is published as a live public demonstration and is intended for academic and demonstration purposes.

The tool queries seven free public information sources. One of them, a commercial web-search service called Tavily, is treated as the primary research engine and is deliberately restricted to a fixed list of authoritative websites, including INTERPOL, UNESCO, the Getty, the German Lost Art Database, the Central Registry of Information on Looted Cultural Property, the United States Federal Bureau of Investigation, and major auction houses. The remaining six sources, drawn from museum collections and encyclopedic and structured reference data, are used to corroborate and to supply exact dates.

A general-purpose artificial-intelligence model, Google's Gemini, performs two jobs: it can identify an artwork from a photograph, and it arranges the retrieved facts into a chronological ownership timeline. It is instructed to use only facts present in the retrieved material, and to mark any period of ownership that the sources do not account for as an explicit gap.

The most important design decision in the project is that the confidence score is not produced by the artificial-intelligence model. It is calculated by a short, fixed, published arithmetic rule written into the server software. The same set of findings will always produce the same score, and any reader can check the arithmetic.

The tool is candid in its own output about the limits of what it has done. The digital signature attached to each passport states that the record attests to process, not to underlying truth. That framing is accurate and is the right way for a policy reader to understand the entire system.

Provenance Search does not query the restricted law-enforcement and commercial databases that professional due diligence relies on. It reads the public web pages of some of the organisations that maintain those databases. The distinction is central to interpreting its output.

02

Background and Rationale

The problem. The trade in art and cultural objects is among the largest asset markets with comparatively little mandatory disclosure. Ownership history is held in records that were built for different purposes by institutions with no obligation to share, and there is no single register that a buyer, journalist, heir, or customs officer can consult to see the whole chain.

Why gaps matter. Provenance research does not usually produce a clean finding of theft. It produces an absence. An object whose recorded ownership jumps from 1932 to 1948 with nothing in between is not thereby proven to have been looted, but the years between 1933 and 1945 are precisely the window in which Jewish and other persecuted owners across continental Europe were dispossessed by confiscation, forced sale, and sale under duress. An unexplained gap in that period is what a suppressed transfer looks like from the outside. The professional convention, reflected in museum guidance in both the United States and Europe, is to treat such a gap as a trigger for further research rather than as a verdict.

The gap in tooling. The public resources that do exist are good but fragmented, and most require the researcher to know which one to consult and what to type into it. Assembling a first-pass picture across all of them is slow, repetitive clerical work. It is exactly the kind of work that software can usefully do, provided the software is honest about the difference between finding a record and establishing a fact.

The response. Provenance Search automates the first pass. It runs the same query against several sources at once, gathers what comes back, arranges it in date order, names the holes, and puts a transparent number on how much of the picture is actually supported by retrieved evidence. It is presented by its own interface as a tool that flags what cannot be verified rather than one that confirms what can.

Relationship to related work. The repository's package metadata, its earliest commit, and the identifier embedded in every passport signature all carry the name arts and artifacts, which is also the name of a sibling repository in the same organisation. Provenance Search is best understood as the deployed, publicly hosted web version of that line of work. This report describes only what is present in this repository.

03

Objectives

The tool is designed to:

  1. Assemble, from free and public sources only, whatever documented ownership history exists for a named artwork, without requiring the user to hold a subscription to any commercial database.
  2. Present that history as a dated chronological timeline in which every entry carries the source it came from, so that a reader can follow any claim back to its origin.
  3. Report the absence of information as a finding in its own right, rather than presenting an incomplete chain as though it were complete.
  4. Raise an explicit, high-severity alert when the search turns up material on the public sites of the recognised stolen-art and looted-art registries.
  5. Express overall reliability as a single number produced by a fixed published rule, so that the score cannot drift with the mood of a language model and can be audited by anyone.
  6. Work in the setting where the question is most often asked, including on a mobile telephone in a gallery, by allowing the object to be photographed rather than described.

04

How the Tool Works

The system has four stages. A user's request passes through all of them in a single operation that takes a few seconds.

Stage one, describing the object. The user supplies what they know through one of three routes.

  • A text form with fields for title, artist, period or date, medium, and an optional last known sale price.
  • An uploaded photograph of the object.
  • A photograph taken there and then with the device camera, which is the mode intended for use in a museum or a saleroom.

If a photograph is supplied, it is sent to Gemini's image-reading capability, which returns its best guess at title, artist, period, and medium, together with its own self-reported certainty about that identification and a short note. Those values are written into the form, and the search then runs automatically. Only the title and the artist are strictly required to proceed. Large photographs are reduced in size before they are sent, so that a high-resolution telephone image does not exceed the limits of the free service.

Stage two, querying the sources. The title and artist are combined into a search phrase and sent to seven sources at the same time. Each source returns one of three verdicts about the object: a hit was found and it raises no alarm; a hit was found on a registry of lost or stolen property; or nothing matching was found.

Stage three, assembling the timeline. Everything the seven sources returned is gathered into a single block of text and passed to Gemini with a set of written instructions. The model is told that the restricted web search is the primary basis for the timeline and that the museum and reference sources are supplementary corroboration. It is told to use only facts present in the material supplied. It is told that where the sources leave a period of ownership unaccounted for, it must insert an entry marked as a gap with a note explaining what is missing, on the stated principle that a gap is itself a fact worth reporting.

The model is also given one narrow permission to go beyond the retrieved material. For works it recognises as very well documented, where the live sources returned little or nothing, it may fill in widely known ownership history from its own training. Every such entry must be labelled as general knowledge, must be marked unverified, must carry no source link, and may never contradict what a live source actually said. Whenever this permission is used, the software itself adds a medium-severity flag to the result, so the reader sees that part of the timeline rests on the model's memory rather than on a citation.

Stage four, scoring and signing. The assembled timeline and flag list are then handed back to the server's own code, which does three things without any further involvement from the artificial-intelligence model. It adds a high-severity flag for every hit on a registry domain, independently of whether the model noticed it. It computes the confidence score by fixed arithmetic. And it attaches a signature block recording the time of the check and a digital fingerprint, a short string of characters derived mathematically from the title, the artist, and the timestamp, which allows a later reader to detect whether those details have been altered.

05

The Variables Explained

This section is the heart of the report. It sets out every input the tool takes, every rule it applies, and how each one reaches the final result.

Title and artist. These are the only required fields. They are combined into the phrase sent to every source, so they determine everything that follows. A misspelled artist name or a title in the wrong language will quietly produce a thin result rather than an error, which is a real practical caution for users.

Period or date, and medium. These are optional and are not used to filter the searches. They are passed to the model as descriptive context, helping it distinguish between different works that share a title and helping it judge whether a returned record is really the same object.

Last known sale price. This optional figure exists to support a single check. If the user supplies a price, the model is asked whether that price is clearly out of line with a comparable figure actually present in the retrieved sources. If, and only if, such a comparable exists and the supplied price is inconsistent with it, the valuation is marked anomalous. The instruction is deliberately conservative: with no price supplied, or no comparable found, the answer is always no. A price far above or below the plausible market level is a recognised signal in art-market due diligence, since valuation is one of the few numbers that has to be stated openly and is therefore one of the few that can be checked against the record.

Each of the seven sources returns one of three verdicts.

  • Clear means the source found at least one matching record and nothing alarming.
  • Flagged means the restricted web search returned a result hosted on one of the loss and stolen-property registries.
  • Not found means the source returned nothing, or was skipped because no access key was configured, or failed.

Only the primary web search can return the flagged verdict. The museum and reference sources can only ever say clear or not found, because they hold collection catalogues rather than loss reports.

DomainInstitution behind it
interpol.intINTERPOL's stolen works of art database
artloss.comThe Art Loss Register
lostart.deThe German Lost Art Foundation's database
lootedart.comThe Central Registry of Information on Looted Cultural Property
fbi.govThe FBI's stolen art file
The watchlist. If any result returned by the primary web search is hosted on one of these five domains, the software adds a risk flag of type watchlist match at high severity, records which domain it came from, and links to the page.

This rule is deterministic. It runs in the server's own code after the model has finished, it compares the address of each returned page against the list of five, and it does not ask the model's opinion. That is a deliberate safeguard: the single most consequential signal the tool can produce is the one signal that a language model is not permitted to suppress or to invent. It is also the one signal that disappears entirely if the web-search service is not configured, a dependency the repository documents plainly.

PenaltyDeductionWhen it applies
Custody gap30 points eachAny timeline entry the model marked as a period of unaccounted-for ownership
Thin corroboration25 pointsFewer than three of the seven sources returned anything at all
High-severity risk flag10 points eachAutomatic watchlist matches, and any high-severity flag the model raised from the retrieved material, such as a documented forced transfer or an unresolved legal claim
Anomalous valuation10 pointsThe supplied sale price was marked anomalous against a comparable figure found in the sources
The confidence score begins at 100 per cent, is reduced by these four penalties, and is then held within the range of 0 to 100. The result is divided by 100 and reported as a proportion, which the interface displays as a percentage.

The custody-gap penalty is the heaviest in the calculation, and the weighting reflects the professional convention that in provenance work an unexplained break in the chain is the primary warning sign, not a minor blemish. Its severity also means the score falls very fast. Two gaps alone remove 60 points. Three take the score to zero on their own.

The corroboration term measures corroboration rather than content. A finding supported by one source is a lead; a finding that four independent sources recognise is an established record. The threshold of three out of seven is a judgment call by the developer rather than a derived figure. Note also that the test counts any verdict other than not found, so a flagged result counts towards corroboration in the same way a clear result does, on the reasoning that a registry hit still demonstrates that the object is known to the record.

Two properties of this calculation deserve to be stated clearly, because they shape how the number should be interpreted. First, the score measures how well documented the ownership history is, not how likely it is that the object is legitimate. A famous work with a complete and well-known history that includes a documented wartime confiscation will score very low, because that confiscation registers as a break in title and attracts high-severity flags. An obscure object about which almost nothing is known may also score low, because too few sources returned anything. The two cases are very different in substance and can look similar in the number. The written rationale that accompanies the score is intended to distinguish them, and a reader should always read it.

Second, because the penalties are subtractive and large, the score reaches zero easily and then stops. Once it is at zero, further findings do not change it. Zero therefore means at least this bad rather than a measured floor. The repository's own worked example, a demonstration record for Egon Schiele's Portrait of Wally, scores zero for exactly this reason, with several custody gaps and several high-severity flags in combination.

ScoreColour
Below 40 per centRed
40 to 70 per centAmber
70 per cent or aboveGreen
The display bands. These are presentational thresholds only. Nothing in the software behaves differently according to which band a score falls into, and the bands carry no legal or institutional meaning.

06

Reading the Results

The passport. The output is a single structured record. It contains the artwork's details as searched, the confidence score, a short written rationale in plain language explaining what is and is not verified, the ownership timeline, the risk flags, the valuation assessment, the list of sources consulted with each one's verdict, and the signature block.

The timeline. Each entry gives a period, an owner, and where available a note, a source link, and the name of the source authority. Entries that represent gaps are shown with the owner field marked and tagged as a custody gap in the interface. Entries drawn from the model's own knowledge rather than a retrieved source are tagged as general knowledge. A reader can therefore see at a glance which parts of the chain are cited and which are not.

The risk flags. Each flag has a type, a severity of high, medium, or low, a plain-language detail sentence, and where applicable a link. Severity governs the colour of the flag in the interface and, for high-severity flags only, feeds the score.

The sources consulted panel. This lists all seven sources with their verdicts, so that a reader can see not only what was found but where nothing was found. This matters more than it might appear. A not-found verdict from a museum collection means only that the museum does not hold the object. It is not evidence of anything about the object's history.

The signature. Each passport records the identifier of the software version that produced it, the exact time, a digital fingerprint, and an attestation sentence. The attestation states that the passport records the results of automated queries to free public sources and attests to process, not to underlying truth. This is the single most important sentence in the output and should be read as governing everything above it.

07

The Data Sources in Plain Terms

Tavily. A commercial web-search service designed to be used by software rather than by a person browsing. In this project it is the primary research engine, and it is restricted to a fixed list of thirteen websites so that it cannot return results from the open web. The query sent is the title and artist followed by the words provenance, ownership, history, looting, theft, and restitution.

The thirteen permitted domains are:

  • metmuseum.org
  • getty.edu
  • interpol.int
  • unesco.org
  • artloss.com
  • lostart.de
  • lootedart.com
  • christies.com
  • sothebys.com
  • artnet.com
  • fbi.gov
  • ifar.org
  • wikipedia.org

For a reader not familiar with the field: INTERPOL maintains the only global database of police-certified records of stolen cultural objects, publicly searchable since 2021 through its free ID-Art application. The FBI's National Stolen Art File, established in 1997, is a publicly searchable United States register of stolen art and cultural property, populated only by law-enforcement agencies. The German Lost Art Foundation's Lost Art Database records cultural assets seized between 1933 and 1945 as a result of persecution, and objects whose history cannot exclude such a seizure; it is free and public. lootedart.com is the Central Registry of Information on Looted Cultural Property 1933 to 1945, established in 2001 by the Commission for Looted Art in Europe, and holds both a documentary database and an object database.

The Art Loss Register is a private London-based commercial company operating what it describes as the largest private database of stolen art; its data is not publicly accessible and searches are a paid service. The Getty Research Institute's Provenance Index is a large free scholarly resource built from transcribed sale catalogues, dealer stock books, and household inventories, weighted towards Western European art from the sixteenth to the early twentieth century. IFAR, the International Foundation for Art Research, was a New York non-profit founded in 1969 whose provenance guide has long been a standard plain-language reference; it announced in 2024 that it was winding down operations. Christie's, Sotheby's, and Artnet are commercial auction and art-market sources whose catalogue entries frequently include provenance statements.

The Metropolitan Museum of Art. The Met publishes an open interface to its collection catalogue requiring no key. The tool retrieves up to three matching objects and reads their title, artist, date, medium, credit line, and public web address. The credit line is useful because it often names the donor or bequest through which the museum acquired the work.

The Art Institute of Chicago. Also an open collection catalogue requiring no key. This is the only museum source in the set that returns a dedicated provenance text field, and it is therefore the most directly valuable of the three museum sources for the tool's purpose.

The Museum of Modern Art. MoMA publishes no live search facility, and its website blocks automated access, so the project takes a different approach. MoMA's collection is published as an open static dataset on a public code-sharing site. The repository includes a script that downloads that dataset, keeps six fields for each work, and compresses it into a single file of roughly four megabytes covering about 159,000 works. That file is loaded into the server's memory when it starts, so a MoMA search happens instantly and involves no network request at all. The matching rule is simple: every word of the title must appear in the record and at least one word of the artist name longer than two characters must also appear.

Wikipedia and Wikidata. The English-language encyclopedia's search facility is used for general background, returning up to three article summaries. Wikidata, a companion project holding structured facts rather than prose, is searched for the artwork, picks the best-matching entry by looking for the artist's surname in the entry description, and then asks for five specific categories of fact: when the work was made, where it is now, which collections have held it and between which dates, who has owned it and between which dates, and any significant events recorded against it. This is the most precisely targeted of the supplementary sources, because those categories map directly onto the shape of a provenance timeline.

Europeana. A European Union cultural-heritage aggregator that brings together records from thousands of European galleries, libraries, archives, and museums. It requires a free key, and the tool skips it if none is configured.

Gemini. Google's family of large language models, capable of reading images as well as text. It is used for the photograph identification and for the assembly of the timeline. It is not used for the score.

08

The Legal and Ethical Backdrop

The tool does not implement any legal test, and this report does not claim that it does. But the reason its outputs matter is set by a body of international commitments that a policy reader will recognise, and the registries it searches exist because of them.

The Washington Conference Principles on Nazi-Confiscated Art, agreed in December 1998 by 44 states, set out eleven principles including the identification of art confiscated by the Nazis and not subsequently restituted, the opening of archives to researchers, the publicising of unidentified works, and the achievement of just and fair solutions for claims. They are explicitly non-binding: a moral and political commitment rather than a treaty, with no enforcement mechanism. The Terezin Declaration of June 2009 extended the same approach to immovable property, Judaica, and archives, and in March 2024 a set of Best Practices for the Washington Principles was issued for the twenty-fifth anniversary, likewise non-binding, which among other things treats sales under duress between 1933 and 1945 as equivalent to involuntary transfers.

lootedart.com, one of the five watchlist domains, was created specifically to fulfil the sixth of the Washington Principles, which called for a central repository of information. The German Lost Art Database serves the same purpose within Germany's national framework. The tool's watchlist is thus not an arbitrary list of websites; it is a list of the public faces of the institutional infrastructure built in response to those commitments.

For objects other than Nazi-era losses, the governing instrument is the 1970 UNESCO Convention on the Means of Prohibiting and Preventing the Illicit Import, Export and Transfer of Ownership of Cultural Property, which entered into force in April 1972 and now has around 150 states parties. It obliges states to maintain inventories, to operate export certification, to bar museums from acquiring undocumented material, and to return stolen inventoried objects on request. It is not retroactive, binding states only in respect of transfers after it entered into force between the states concerned, which is why 1970 functions in practice as a due-diligence cut-off date for antiquities, and why it does not reach the 1933 to 1945 period. The 1995 UNIDROIT Convention on Stolen or Illegally Exported Cultural Objects, in force since July 1998, was designed to address the private-law gaps the UNESCO Convention leaves open, but has far fewer parties.

It should be stated plainly that the tool applies none of these instruments. It does not determine title, assess a restitution claim, or decide whether an export was lawful. What it does is surface the public traces that the institutions created under these frameworks have left on the open web. Deciding what any of it means remains a matter for lawyers, provenance researchers, and the claims processes those frameworks established.

09

Design Choices

Why the score is not produced by the model. A language model asked to rate its own confidence will produce a number that sounds reasonable and cannot be reproduced or checked. The project instead computes the score in ordinary code from countable facts: how many gaps, how many sources responded, how many high-severity flags, whether the valuation was anomalous. The result is that two runs producing the same findings produce the same score, and a reader who disagrees with the score can identify exactly which term they disagree with. This is the strongest design decision in the project.

Why the primary search is restricted. An unrestricted web search for the words looting and theft alongside a famous artist's name would return a great deal of journalism, speculation, and commercial content. Restricting the search to thirteen named institutional and market domains means that the material the model reasons over comes from sources with an identifiable custodian, at the cost of missing anything held elsewhere.

Why gaps are recorded rather than smoothed over. The instruction given to the model states that a gap is itself a fact worth reporting, and the interface displays gaps prominently in red. A system that quietly produced an unbroken chain wherever it lacked data would be worse than useless in this domain, because the incomplete chain is the finding.

Why the general-knowledge fallback exists and why it is fenced. Without it, the tool would return an almost empty result for the most famous works in the world, since the free sources it uses may hold little ownership detail even for a painting whose history is taught in schools. The fallback lets the model fill those blanks, but it is constrained on four sides: every such entry is labelled in the data, tagged in the display, never marked verified, and never allowed to override a live source, and its use triggers an automatic medium-severity flag. The constraint is well designed. It remains the part of the system where an error is hardest for a non-specialist reader to detect.

Why the MoMA data is bundled rather than queried. Because MoMA offers no live search and blocks automated access, the only lawful and reliable route to its collection is its own published open dataset. Bundling a compressed copy makes the search instant and removes a point of failure, at the cost that the copy is only as current as the last time it was rebuilt.

Why everything runs on the server. The user's browser never contacts any external service directly. It speaks only to this project's own server, which holds the access keys. This keeps the keys out of the browser, where they would be readable by anyone.

10

Limitations and Caveats

The report states twelve limitations. They are reproduced here because they are the most important part of the document for any reader considering what weight to give the tool.

It does not search the restricted databases. This is the most important limitation and the one most likely to be misread. The interface names INTERPOL, the Art Loss Register, Lost Art, and the FBI among the sources it searches. What it actually searches is the public web pages of those organisations, by way of a general web-search service. The Art Loss Register's database is a paid commercial service with no public access at all. A negative result from this tool is therefore not a clearance against the Art Loss Register, and must never be presented as one. INTERPOL and the FBI files are publicly searchable through their own interfaces, but the tool does not query those interfaces directly either.

An absence of findings is not a clean history. Every not-found verdict in the panel means only that the source returned nothing for the phrase that was searched. It carries no information about the object.

The score conflates two different situations. A well-documented history containing a wartime seizure and an obscure object with no history at all can both produce a very low score. The number alone does not distinguish them.

The weights are the developer's judgment. Thirty points for a gap, twenty-five for thin corroboration, ten per high-severity flag, ten for a valuation anomaly, three sources as the corroboration threshold: none of these figures is derived from a study, an expert panel, or a validation exercise against known cases. They are reasonable choices that produce sensible orderings, and they should be described as such rather than as a measurement.

The timeline depends on a language model. The model is instructed to use only retrieved facts, and its temperature setting is kept very low to make its output as consistent as possible, but it is still a language model reading messy source material. It can misread a date, attach a record to the wrong object, or mistake a similarly titled work for the one being searched. Nothing in the system checks its assembly against the retrieved material after the fact.

Identification from a photograph is a guess. The image step returns a best-effort identification with a self-reported certainty figure. That figure is the model's own estimate, it is not carried into the confidence score, and an incorrect identification will produce a fully formed passport for the wrong object.

The searches are simple text matches. The query is the title and artist as typed. There is no handling of alternative titles, transliterations, works known by different names in different languages, or the many objects for which no single agreed title exists. This weighs most heavily against exactly the categories of object where provenance questions are most acute, including antiquities and non-Western material, which frequently have no title and no named artist at all.

The source list is Western-weighted. Two of the three museum sources are American, the aggregator is European, and the market sources are the two large London and New York auction houses. Objects from collections and markets outside that orbit will be under-represented, and this limitation compounds the previous one.

The sources change beneath the tool. The domain list is fixed in the software. One of the thirteen domains, ifar.org, belongs to an organisation that announced in 2024 that it was winding down. Domain lists of this kind require periodic review, and nothing in the repository schedules one.

Dependencies and degradation. If the web-search key is absent, the tool runs on the supplementary sources alone and the watchlist rule cannot fire at all, which removes its single most consequential signal without any visible change in the shape of the output. If the Europeana key is absent, that source is silently skipped and counts as not found, which can push the corroboration count below the threshold and cost 25 points for a reason unrelated to the artwork.

The written rationale is not the arithmetic. The plain-language explanation shown beside the score is composed by the language model and describes the substance of the case. It does not narrate the calculation, and a reader should not assume the two are saying the same thing.

Status. This is a research prototype by a single developer, deployed as a public demonstration. It is not an accredited due-diligence service, it carries no professional indemnity, and its passport is not a certificate. Its own attestation says so.

11

Practical Nature and Deployment

The tool is a single web page backed by a small server program. A user needs only a web browser and no installation. The public demonstration is hosted on a commercial application-hosting platform, and the project's repository page redirects visitors to it.

The repository also contains three stored example results, for Leonardo da Vinci's Salvator Mundi, Vincent van Gogh's The Starry Night, and Egon Schiele's Portrait of Wally. These are saved copies of earlier live runs, included so that the output can be demonstrated without a working connection or a configured key. Their signature blocks are marked as static snapshots. They are teaching material and should not be read as current findings.

The choice of examples is apt. Portrait of Wally is the case that did more than any other to establish that a loan to an American museum could expose an unresolved Nazi-era claim: taken from the Jewish Viennese dealer Lea Bondi Jaray around the time she fled Vienna in 1939, it was seized in New York in 1998 while on loan to the Museum of Modern Art, and after some thirteen years of federal forfeiture litigation the matter settled in July 2010 with the Leopold Museum paying the Bondi estate 19 million United States dollars to retain the painting and agreeing to display its true provenance permanently alongside it. Salvator Mundi, sold at Christie's in November 2017 for 450.3 million dollars including fees, is the case most associated with the opposite problem: a chain of ownership with a long blank stretch, an attribution that remains contested, and a present location that is not publicly confirmed.

All access keys are held in server configuration and are excluded from the repository. The README notes explicitly that if any key file was ever committed or shared, the keys should be rotated. That is the right instruction and it is good practice to have written it down.

12

Intended Audience and Use

The tool is aimed at the person who needs a fast first look: a student researcher, a journalist, a small institution without a provenance department, a family beginning to trace an object, or a visitor standing in front of a work with a question about it. For that audience it does something genuinely useful, which is to run in seconds a set of searches that would otherwise take an afternoon, and to say clearly which parts of the answer are cited and which are not.

It is not aimed at, and should not be used for, the decision points where the answer carries legal or commercial weight. An acquisition, a sale, a restitution claim, an export licence, or a repatriation request all require searches of the restricted registries, examination of physical and archival evidence, and professional judgment. The right way to read a passport from this tool is as a list of leads and of questions that have not been answered.

13

Conclusion

The value of Provenance Search lies less in what it finds than in how carefully it reports what it did not find. Its timeline names its gaps. Its source panel shows silences as well as answers. Its most consequential signal, the registry match, is computed by rule rather than by a language model. Its score can be recomputed by hand from four numbers. Its signature states that it attests to process and not to truth. In a field where an incomplete chain presented as a complete one is the characteristic harm, that discipline is the substance of the contribution.

The prototype's weaknesses are the ones its own design invites. It reads the public faces of registries rather than the registries themselves, and its interface language does not make that distinction as sharply as its documentation does. Its scoring weights are informed guesses. Its search strategy assumes a world of titled works by named artists, which is not the world in which most contested cultural property sits. These are addressable, and naming them is more useful than the tool's current score would be to anyone making a decision.

What the project demonstrates is a pattern worth carrying into other domains: let the automated system gather and arrange, let fixed published rules do the judging, mark every claim with its origin, and treat the absence of evidence as a reportable finding rather than a blank to be filled.

References

Sources

  1. 01Washington Conference Principles on Nazi-Confiscated Art, agreed December 1998 by 44 states. Non-binding.
  2. 02Terezin Declaration on Holocaust Era Assets and Related Issues, June 2009. Best Practices for the Washington Principles, March 2024.
  3. 03UNESCO. Convention on the Means of Prohibiting and Preventing the Illicit Import, Export and Transfer of Ownership of Cultural Property, 1970, in force April 1972.
  4. 04UNIDROIT. Convention on Stolen or Illegally Exported Cultural Objects, 1995, in force July 1998.
  5. 05INTERPOL. Stolen Works of Art Database and the ID-Art application, publicly searchable since 2021.
  6. 06Federal Bureau of Investigation. National Stolen Art File, established 1997.
  7. 07German Lost Art Foundation. Lost Art Database, cultural assets seized between 1933 and 1945.
  8. 08Commission for Looted Art in Europe. Central Registry of Information on Looted Cultural Property 1933 to 1945, established 2001.
  9. 09The Art Loss Register. Private commercial database of stolen art; searches are a paid service.
  10. 10Getty Research Institute. Provenance Index, transcribed sale catalogues, dealer stock books, and household inventories.
  11. 11International Foundation for Art Research. Provenance guide. Founded 1969; announced wind-down in 2024.
  12. 12The Metropolitan Museum of Art. Open Access collection interface.
  13. 13Art Institute of Chicago. Open collection interface, including a dedicated provenance field.
  14. 14Museum of Modern Art. Open collection dataset, roughly 159,000 works.
  15. 15Wikidata. Structured facts on collections, owners, and significant events.
  16. 16Europeana. European Union cultural-heritage aggregator.
  17. 17Tavily. Commercial web-search service for software, used as the restricted primary research engine.
  18. 18Google. Gemini family of large language models, used for image identification and timeline assembly.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file diff --git a/static-site/print/provenance-search/index.txt b/static-site/print/provenance-search/index.txt index dc64c06bf..182a8476c 100644 --- a/static-site/print/provenance-search/index.txt +++ b/static-site/print/provenance-search/index.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -12:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","print","provenance-search",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","provenance-search","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -14:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -16:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","print","provenance-search",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","provenance-search","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:"$Sreact.suspense" -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1c:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -29,7 +29,7 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 19:[] 10:"$W19" 11:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1e:Ta14, @page { size: 210mm 297mm; margin: 20mm 18mm; } @@ -153,6 +153,6 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 45:["$","p","11",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"The written rationale is not the arithmetic."}]," ","The plain-language explanation shown beside the score is composed by the language model and describes the substance of the case. It does not narrate the calculation, and a reader should not assume the two are saying the same thing."]}] 46:["$","p","12",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Status."}]," ","This is a research prototype by a single developer, deployed as a public demonstration. It is not an accredited due-diligence service, it carries no professional indemnity, and its passport is not a certificate. Its own attestation says so."]}] 1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -47:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +47:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 18:null 1d:[["$","title","0",{"children":"Provenance Search (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L47","6",{}]] diff --git a/static-site/print/vango/__next._full.txt b/static-site/print/vango/__next._full.txt index 4d304c08d..2e6d51290 100644 --- a/static-site/print/vango/__next._full.txt +++ b/static-site/print/vango/__next._full.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -12:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","print","vango",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","vango","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -14:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -16:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","print","vango",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","vango","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:"$Sreact.suspense" -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1c:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -29,7 +29,7 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 19:[] 10:"$W19" 11:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1e:Ta14, @page { size: 210mm 297mm; margin: 20mm 18mm; } @@ -141,6 +141,6 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 39:["$","p","16",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Demonstration stamps."}]," ","A guest opening the application for the first time is given three stamps already collected: Chromatic Drift, Fault Lines, and Hollow Choir, dated across May and June 2026. These are not real visits. They exist so that a person opening the demonstration sees a passport with contents rather than an empty book, which would communicate very little about what the tool is for. A registered account starts genuinely empty."]}] 3a:["$","p","10",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"No automated tests exist."}]," ","The repository contains none."]}] 1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -3b:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +3b:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 18:null 1d:[["$","title","0",{"children":"VANGO: The Art Passport (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L3b","6",{}]] diff --git a/static-site/print/vango/__next._head.txt b/static-site/print/vango/__next._head.txt index 7c6430ab5..00f3ee889 100644 --- a/static-site/print/vango/__next._head.txt +++ b/static-site/print/vango/__next._head.txt @@ -1,6 +1,6 @@ 1:"$Sreact.fragment" -2:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -3:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +2:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +3:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 4:"$Sreact.suspense" -5:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +5:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 0:{"rsc":["$","$1","h",{"children":[null,["$","$L2",null,{"children":[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]]}],["$","div",null,{"hidden":true,"children":["$","$L3",null,{"children":["$","$4",null,{"name":"Next.Metadata","children":[["$","title","0",{"children":"VANGO: The Art Passport (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L5","6",{}]]}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/print/vango/__next._index.txt b/static-site/print/vango/__next._index.txt index f98a4fd90..18b9d82b6 100644 --- a/static-site/print/vango/__next._index.txt +++ b/static-site/print/vango/__next._index.txt @@ -1,13 +1,13 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -d:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +d:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} 8:["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$Ld",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$Ld",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$Ld",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$Ld",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] diff --git a/static-site/print/vango/__next._tree.txt b/static-site/print/vango/__next._tree.txt index 0d22c43ef..68b39447c 100644 --- a/static-site/print/vango/__next._tree.txt +++ b/static-site/print/vango/__next._tree.txt @@ -1,7 +1,7 @@ -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] 0:{"tree":{"name":"","param":null,"prefetchHints":16,"slots":{"children":{"name":"print","param":null,"prefetchHints":0,"slots":{"children":{"name":"slug","param":{"type":"d","key":"vango","siblings":null},"prefetchHints":0,"slots":{"children":{"name":"__PAGE__","param":null,"prefetchHints":0,"slots":null}}}}}}},"staleTime":300,"buildId":"static-snapshot"} diff --git a/static-site/print/vango/__next.print.txt b/static-site/print/vango/__next.print.txt index 460afe242..fa617e997 100644 --- a/static-site/print/vango/__next.print.txt +++ b/static-site/print/vango/__next.print.txt @@ -1,4 +1,4 @@ 1:"$Sreact.fragment" -2:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -3:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +2:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +3:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] 0:{"rsc":["$","$1","c",{"children":[null,[["$","style",null,{"children":"\n /* The site is dark-first, and a dark color-scheme makes Chrome paint\n the printed sheet's margin area dark too. Paper is light. */\n html { color-scheme: light !important; }\n html, body { background: #ffffff !important; }\n body::after { content: none !important; }\n header, footer, .site-bg, .aura { display: none !important; }\n "}],["$","$L2",null,{"parallelRouterKey":"children","template":["$","$L3",null,{}]}]]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/print/vango/index.html b/static-site/print/vango/index.html index 8930f963c..ecabb6541 100644 --- a/static-site/print/vango/index.html +++ b/static-site/print/vango/index.html @@ -1,4 +1,4 @@ -VANGO: The Art Passport (print edition) · NYU Ethical Tech CoLab

Publications · Academic report

VANGO: The Art Passport

A Digital Passport for Recording Visits to Works of Art

Ethical Tech CoLab

NYU Center for Global Affairs

July 2026

Melanie MacKew. Developed as part of masters research at the NYU Center for Global Affairs, under the Ethical Tech CoLab.

A passport does not evaluate the traveller or rank the countries visited. It records that a person was in a particular place on a particular day. VANGO applies that idea to art: each participating artwork carries a short code, a visitor standing in front of the work scans or types it, and a dated stamp is added to a passport carried on their phone. The application is a record of attendance, not of ownership, value, or authenticity. It answers one question only: who went to see what, and when.

Key figures

7

artworks in the catalogue, five invented for demonstration and two real

4

interface languages, English, French, Italian, and Hausa

5

fixed colour schemes, selected arithmetically from the stamp's own identifier rather than at random

10

limitations the prototype states about itself, including that a code proves nothing about presence

00

Foreword

Most people who visit a museum, a biennale, or a small studio exhibition leave with nothing to show for it. The ticket is thrown away. The photograph is lost in a phone. A few months later the visitor can recall that the work was striking but not what it was called, who made it, or where it was seen. For the institution, the encounter is equally invisible: a body passed through a room and left no trace of having been moved by anything.

There is an older technology that solved a version of this problem. A passport does not evaluate the traveller or rank the countries visited. It simply records, in a durable and personal form, that a person was in a particular place on a particular day, and it does so with a stamp that carries the character of the place that issued it.

VANGO is a research prototype that applies this idea to art. It is a digital passport, carried on a phone, in which a visitor collects a stamp for each artwork they go and see. This report explains in non-technical language what VANGO is, how it works, what each piece of information it stores actually represents, and what it deliberately does not attempt to do.

01

Executive Summary

VANGO is a research prototype, an experimental piece of software, that takes the form of a small illustrated passport book displayed on a mobile phone screen. Each participating artwork is given a short code. A visitor who is standing in front of the work either scans a printed square barcode, known as a QR code, or types the code by hand. The application then adds a stamp for that artwork to the visitor's passport, dated with the day of the visit.

The application is a record of attendance, not of ownership, value, or authenticity. It answers one question only: who went to see what, and when.

The prototype contains a catalogue of seven works. Five are invented demonstration pieces with fictional artists and venues. Two are real: the marble David by Michelangelo at the Galleria dell'Accademia in Florence, and Bura ceramics from Niger, a class of Iron Age terracotta object discussed in Section 9 of this report.

Each stamp is drawn as a perforated postage stamp with a hand-drawn illustration of the work, in one of five colour schemes. The colour is not chosen at random each time it is displayed. It is derived arithmetically from the stamp's own identifier, so that a given stamp always appears in the same colours to every user, on every device, forever.

The application can be used in two ways. In guest mode nothing leaves the visitor's own phone and no account is required. In account mode the visitor registers with an email address and password, and the collection is stored on a server so that it survives a change of device.

The interface is available in four languages: English, French, Italian, and Hausa, one of the principal languages of Niger and northern Nigeria.

VANGO is a student research prototype. It is a demonstration of an idea about audience engagement. It is not a finished consumer product, it has not been tested with real museum visitors, and, as Section 10 sets out, several parts of it are visibly unfinished.

02

Background and Rationale

The problem. Cultural institutions have limited ways of encouraging repeat attendance or of connecting a visit to one venue with a visit to another. Loyalty schemes are commercial in tone and tend to reward spending rather than attention. A visitor who has seen forty exhibitions over five years holds that history only in memory.

The precedent. The design borrows openly from an established and well-evidenced model. Since 1986 the United States National Park Service, in partnership with the non-profit now known as America's National Parks, has run the Passport To Your National Parks programme, a small booklet in which visitors collect ink cancellation stamps at park visitor centres. The scheme has endured for four decades because the reward is not a discount but a record. VANGO transposes this format from landscape to art.

The gap. The paper passport works because each park has a physical desk and a rubber stamp. Individual artworks do not. An exhibition may run for six weeks in a rented pavilion. What VANGO proposes is that a printed code placed beside a work can perform the same function as the rubber stamp at the ranger's desk, at negligible cost and with no staffing requirement.

The response. VANGO is therefore built around the smallest possible unit of proof: a short code, posted in public, that a visitor converts into a dated entry in a personal book. Everything else in the application, the illustrations, the page-turning animation, the passport number, exists to make that entry feel worth keeping.

03

Objectives

The prototype is designed to:

  1. Give a visitor a durable, personal, and pleasing record of the individual works of art they have gone to see.
  2. Reduce the cost of participation for an institution to the price of printing a sheet of paper, so that a small studio can take part on the same terms as a national museum.
  3. Work without a network connection to any central service, and without requiring the visitor to create an account, so that a person can use the tool with no disclosure of personal information at all.
  4. Present the record in a form that is legible across languages, including a West African language, rather than defaulting to English and the major European languages alone.
  5. Ensure that a given artwork's stamp looks the same in every visitor's passport, so that the stamp functions as a shared emblem of the work rather than as decoration generated afresh for each person.

04

How VANGO Works

The application has four moving parts: a catalogue of artworks, a way of capturing a code, a book in which stamps are stored, and an optional account.

The catalogue. The catalogue is a fixed list held inside the application itself. Each entry consists of a code and three pieces of descriptive text: the title of the work, the artist, and the venue.

The seven current entries are:

  • Chromatic Drift
  • Fault Lines
  • Hollow Choir
  • Echo Garden
  • Voidwalk
  • Bura Ceramics
  • David

Because the catalogue is written into the application, a new artwork can only be added by editing the software and publishing it again. The README file documents how to do this. There is no facility for a gallery to register its own work without a developer.

Capturing a code. A visitor opens the add-stamp panel and chooses one of two methods. The first uses the phone's camera to read a QR code. The second is a text box in which the code is typed. A third route exists for institutions that would rather share a web link than print a barcode: a specially formed web address carries the code within it, and opening that link adds the stamp directly.

Before a code is looked up it is normalised, meaning it is converted into a single standard form. The application converts all letters to capitals and removes any spaces and hyphens. The practical effect is that a visitor who types "chroma-14", "CHROMA 14", or "Chroma14" gets the same result. This is a small decision with a large effect on how forgiving the tool feels to someone squinting at a label in a dim room.

If the code matches no catalogue entry, the visitor is told that no artwork is registered for that code. If it matches a work already in the passport, the visitor is told so rather than being given a duplicate.

The book. The passport is presented as a book that opens and whose pages turn. The first page inside the cover is the biography page. Each subsequent page holds exactly two stamps. A back page invites the visitor to collect another. When a new stamp is earned the book automatically turns to the page where it has landed, after a short animation showing the stamp being pressed.

The account. A visitor may register with an email address and password, or continue as a guest. The two paths are described in Section 7. The distinction matters because it determines where the visitor's record is kept and who else can see it.

05

The Variables Explained

This section is the heart of the report. VANGO stores a small number of pieces of information, and each was chosen for a reason. What follows explains, in ordinary words, what each one represents, why it exists, and what it does.

The code. The code is a short string of capital letters and digits such as CHROMA14 or DAVID01. It is the only thing a visitor needs in order to claim a stamp. It must be short enough to type on a phone while standing up, distinctive enough not to be confused with a neighbouring work, and printable in large type on a card. Codes in the catalogue combine a word suggesting the work with a number, which makes them memorable without being guessable in a systematic way. The code is the key that unlocks the catalogue entry: everything the stamp displays about the artwork comes from looking up this one value.

An honest observation about the code. It is a public secret. It is printed beside the work for anyone to see, and it can be photographed, texted to a friend, or published online. The application performs no check on the visitor's location and no check on the time of day. A person who never entered the building can collect the stamp.

Title, artist, and venue. These three pieces of text are held in the catalogue, not entered by the visitor, and they appear on the face of the stamp along with the date. They are the minimum needed for the record to mean anything later. A stamp reading only CHROMA14 would be an unreadable souvenir. Naming the artist alongside the title also matters as a matter of practice, in that it treats the maker rather than the venue as the primary fact about the work.

The date collected. The date is recorded as the calendar day on which the stamp was claimed, with no time of day. It is stored in the international standard order of year, then month, then day, and displayed in the style of a border control stamp, for example 14 MAY 2026. The day is the meaningful unit of a visit. Storing the hour and minute would add precision that no one needs and would make the record more personally revealing than it has to be. The date is printed on the stamp, it determines the order in which stamps appear in the book, and in guest mode it forms part of the rule that prevents a work being stamped twice on the same day.

The stamp identifier. Every stamp is given an internal identifier made by joining the artwork code to the date, for example CHROMA14-2026-05-14. This value is never shown to the visitor. The application needs a way to tell one stamp from another when arranging them on a page, and it needs a stable value from which to derive the stamp's appearance.

The colour palette and the seed. The application holds five fixed pairs of colours: a teal, a pink, a bronze, a blue-violet, and a gold, each paired with a darker version of itself. Every stamp is drawn in one of these five schemes. The scheme is not chosen by the person who added the artwork, and it is not chosen at random. The application takes the stamp identifier and runs it through a hashing function, which is a piece of arithmetic that turns a piece of text into a number. It then divides that number by five and uses the remainder to select the palette.

This method is deterministic, meaning that the same input always produces the same output. Because the identifier for a given artwork on a given date is always the same, the resulting colour is always the same. Two visitors comparing passports will see the same colours for the same work. Had the colours been picked at random when the stamp was created, they would differ between visitors and might change if the application were reinstalled, which would make the stamp feel arbitrary rather than official. The two chosen colours are used throughout the stamp, in the illustration, the border, and the lettering.

The illustration. Each of the seven catalogue entries has its own hand-drawn illustration built into the application as line and shape instructions rather than as a photograph. The David stamp shows the figure with the subtitle FIRENZE 1504. The others are similarly bespoke. Drawings avoid the copyright and licensing questions that photographs of artworks raise, they remain sharp at any size on any screen, and they carry the visual character of an engraved postage stamp, which is the aesthetic the whole application is reaching for.

There is also a fallback. If a work has no bespoke drawing, the application generates an abstract composition of between four and six shapes, again derived arithmetically from the stamp identifier so that it is stable rather than random. In the current catalogue every work has a bespoke drawing, so this fallback is not visible, but it is what would appear if a new code were added without artwork.

The visitor's own details. Three items describe the passport holder rather than the art: the display name, which the visitor types in the settings panel and which defaults to "Explorer"; the profile picture, which the visitor uploads from their own device; and the passport number. The application also stores two preferences: whether the interface is shown in its dark or light colour scheme, and which of the four languages is in use.

The passport number and the membership date. These two variables are the least settled part of the prototype, and the report would be less useful if it glossed over that. When a visitor registers an account, the server generates a passport number in the style of a real travel document: three letters followed by six digits. The letters are drawn from an alphabet that deliberately omits I and O, because those characters are easily confused with the digits 1 and 0 when read aloud or copied by hand. This is a considered choice and matches the convention used on machine-readable travel documents.

However, the biography page inside the book does not display this number. It displays the figure 1000 plus the number of stamps the visitor holds. A visitor with three stamps sees passport number 1003, and that number changes to 1004 the moment they collect another. The same page derives the membership date from the earliest stamp in the collection rather than from the date the account was created, so a visitor who has collected nothing sees no date at all. The most likely explanation is that the biography page was written first, as a visual mock-up, and the server was added later without the page being updated to read from it. Whatever the cause, the figures on the biography page are at present decorative rather than authoritative, and this should be corrected before anyone treats a VANGO passport number as an identifier.

The rule against duplicates. The application prevents the same work being stamped twice, but the rule differs between the two modes of use, and the difference is deliberate. For a registered account the database enforces one stamp per artwork per account, permanently. A visitor who returns to the David a second time will not receive a second stamp. In guest mode the rule is looser: one stamp per artwork per day. A guest may therefore accumulate repeat visits to the same work on different dates.

The permanent rule expresses the passport metaphor faithfully, in that a passport records that you have been to a country rather than how many times. The daily rule makes the guest mode easier to demonstrate, since a person showing the application to someone else can collect the same stamp again tomorrow without clearing their data. It is a demonstration convenience rather than a principled position.

Two stamps per page. The book places exactly two stamps on each page, and inserts an empty page when the current page is full so that the visitor always sees somewhere for the next stamp to go. This is a presentational constant chosen to suit a phone screen held upright, and it is the one variable in the application whose value carries no meaning beyond layout.

Demonstration stamps. A guest opening the application for the first time is given three stamps already collected: Chromatic Drift, Fault Lines, and Hollow Choir, dated across May and June 2026. These are not real visits. They exist so that a person opening the demonstration sees a passport with contents rather than an empty book, which would communicate very little about what the tool is for. A registered account starts genuinely empty.

06

Reading the Results

What the visitor sees is a book, not a score. VANGO produces no rating, no ranking, no leaderboard, and no recommendation. There is nothing to optimise. This is a design decision worth naming, because the obvious commercial version of this application would gamify attendance, and this one declines to.

The biography page shows the holder's name and photograph, the number of stamps collected, and the passport number and membership date discussed above. It carries an emblem labelled Ars Pro Mundo, and the line "Interact with art and receive a unique stamp for each piece you visit."

The stamp pages show the stamps themselves, two to a page, each with its illustration, title, artist, venue, and date, inside a perforated border with decorative rosettes at the corners. The back page carries the line "Every world leaves a mark" and a button to collect another stamp.

The only number in the entire application that a visitor might be tempted to compete over is the count of stamps collected, and nothing is built on top of it.

07

Two Ways to Use It

Guest mode. The visitor taps past the sign-in screen. Stamps are held in the phone's own browser storage. Nothing is transmitted anywhere and no account exists. The record survives closing the application but is lost if the browser's data is cleared or the phone is replaced. For a visitor who does not want to hand over an email address in exchange for a souvenir, this is the whole application, working, at no cost in privacy.

Account mode. The visitor registers with an email address, a password of at least eight characters, and optionally a name. The password is not stored. It is passed through a one-way scrambling process called hashing, using a deliberately slow method known as bcrypt, so that a person who obtained a copy of the database could not read the passwords out of it. The visitor's session is then held by a signed token that expires after seven days.

What the server stores. The database holds two lists, and that is the entire extent of it. The server does not record where the visitor was, what device they used, or how long they looked at anything.

ListFields held
PeopleEmail address, hashed password, name, uploaded picture, membership date, passport number, and the moment the account was created
StampsWhich person, which artwork code, and on what date
Everything the server stores.

An important practical caveat. The published demonstration on the public web cannot reach a server. The address of the server is written into the application as a local address on the developer's own machine. A visitor opening the public demonstration can therefore use guest mode fully, but registration and sign-in will not work. The account system is real, tested code, but it is at present only usable by someone running both halves of the application on their own computer.

08

Language and Reach

The entire interface, including error messages and the welcome sequence, is translated into four languages: English, French, Italian, and Hausa. The translations are held as complete parallel sets of every phrase rather than being assembled from fragments, which keeps the tone consistent within each language.

The inclusion of Hausa is the notable choice. Hausa is one of the most widely spoken languages of West Africa, including in Niger, and it is not a language that consumer applications of this kind ordinarily support. Read alongside the presence of Bura ceramics in the catalogue, it suggests the prototype is not addressed solely to a European museum audience. The README lists only English, French, and Italian, so the Hausa translation, which was added later, is currently undocumented.

09

Cultural Property Context

One catalogue entry raises questions the rest do not. Bura Ceramics is attributed in the catalogue to Niger rather than to a named artist, and to a venue given only as the initials AABC, which the repository does not expand.

The Bura culture refers to a group of Iron Age archaeological sites in the lower Niger River valley, in south-western Niger and south-eastern Burkina Faso, whose terracotta funerary vessels and equestrian figures were first excavated in the 1980s. The sites have since been looted on a very large scale. Terracotta statuettes and pottery from the Bura system appear on the Red List of West African Cultural Objects at Risk published by the International Council of Museums, a listing that exists specifically to alert museums, dealers, customs officers, and collectors that objects of that description are likely to have been removed illegally.

Niger is a State Party to the 1970 UNESCO Convention on the Means of Prohibiting and Preventing the Illicit Import, Export and Transfer of Ownership of Cultural Property, the principal international instrument requiring states to control the movement of cultural objects across borders and to cooperate in their return.

The relevance to VANGO is narrow but real. The application records that a person went to see an object. It records nothing whatsoever about how that object came to be where it is. If a stamp is issued beside a Bura vessel in a gallery, the visitor's passport will say only that they saw it, on a particular day, at that venue. The document is silent on whether the object should be there.

This is not a criticism of the design. It is a boundary that ought to be stated plainly, because a document styled as a passport carries an implication of officialdom that the underlying record does not support. A future version placing works of contested origin in its catalogue would do well to carry the object's collection history on the stamp page, or to link to the holding institution's own published account of it.

For completeness, and because the question naturally arises with any software touching art: VANGO has no relationship to the provenance research frameworks that govern restitution claims, such as the 1998 Washington Principles on Nazi-Confiscated Art, and no relationship to stolen-art databases such as the Art Loss Register. It does not check, assert, or store anything about title, authenticity, or ownership history. It is a record of attendance.

10

Limitations and Caveats

The prototype has real and visible limits, and a reader deciding whether to build on it should know them.

The code cannot prove presence. Anyone holding the code can collect the stamp from anywhere in the world. There is no location check, no time window, and no single-use mechanism. For a souvenir this matters little. For anything that conferred a benefit, such as a discount or a prize, the scheme would be trivially defeated.

The published demonstration cannot sign anyone in. The address of the account server is written into the application as a local address on the developer's own machine, so sign-in and registration fail on the public web.

The passport number is not the one the server issues. The passport number and membership date shown to the visitor are derived from the stamp collection rather than read from the server, so both are decorative rather than authoritative.

The catalogue cannot be extended without editing the software. There is no interface through which a gallery could register a work. Seven works is a demonstration, not a deployment.

Most of the catalogue is fictional. Five of the seven entries are invented. The application has, so far as the repository shows, never been placed in a real exhibition or tested with real visitors.

Each illustration must be drawn by hand in code. This is what makes the stamps attractive and it is also what prevents the catalogue from growing quickly. A catalogue of hundreds of works would need either a different approach to imagery or a great deal of labour.

QR scanning is not universally available. It depends on a barcode-reading capability that not every mobile browser provides. The application detects this and directs the visitor to type the code instead, which is a sound fallback but a less pleasant one.

Profile pictures are stored inefficiently. They are uploaded and stored in a form that makes them substantially larger than the original file, and are held in the same database as the account records. This is workable at demonstration scale and would not be the right approach at any real volume.

Three versions of the application coexist. The repository contains the current source, a superseded starting page, and a single large self-contained file of roughly two-thirds of a megabyte. Which of these is authoritative is not documented, though the build configuration makes clear that the current source is what is published.

No automated tests exist. The repository contains none.

11

Practical Nature of the Tool

VANGO runs in an ordinary mobile web browser. There is nothing to install from an application store. A visitor follows a link or scans a code and the passport opens.

The visitor-facing half of the application is published automatically to a free public web host each time the software is changed. The optional account server is a separate program that must be run by whoever operates the service.

For an institution, participation requires printing one page per artwork. The repository provides a ready-made page for each of the seven works, containing the code, the artwork details, and a QR code, designed to be printed and displayed beside the work.

12

Intended Audience and Use

The immediate audience is the exhibition visitor, and the design assumes that person is holding a phone, standing up, and has perhaps thirty seconds of patience.

The secondary audience is the small or temporary venue: a studio, a pavilion, a biennale stand, an institution without the budget for a mobile application of its own.

The prototype is best read as a demonstration of an argument, namely that a record of having seen something is a form of value in itself, and that it can be given to a visitor at almost no cost and without collecting anything about them in return.

13

Conclusion

VANGO takes a forty-year-old idea from national park visitor centres and asks whether it works for individual works of art. The answer the prototype gives is a qualified yes. The mechanism is simple enough to explain in one sentence, the cost to a venue is a sheet of paper, and the guest mode demonstrates that the whole thing can function without collecting a single piece of personal information.

The prototype's most considered decision is the one a visitor will never notice: that a stamp's appearance is calculated from its own identity rather than picked at random, so that the same artwork wears the same colours in every passport in the world. That is what separates a stamp from a decoration.

Its weakest points are equally clear. The published version cannot sign anyone in, the passport number on the biography page is a placeholder, the catalogue is small and mostly invented, and a code printed on a wall proves nothing about where the person holding it was standing.

The application is honest about its own scope in the only way that matters, which is by not overreaching. It does not claim to authenticate, to value, to rank, or to establish where an object came from. It records that someone went to look at something. For a research prototype, knowing precisely which question it is answering is a considerable part of the work.

References

Sources

  1. 01United States National Park Service and America's National Parks. Passport To Your National Parks, established 1986.
  2. 02International Council of Museums. Red List of West African Cultural Objects at Risk, which lists terracotta statuettes and pottery of the Bura system.
  3. 03UNESCO. Convention on the Means of Prohibiting and Preventing the Illicit Import, Export and Transfer of Ownership of Cultural Property, 1970. Niger is a State Party.
  4. 04Washington Conference Principles on Nazi-Confiscated Art, 1998. Cited only to record that VANGO has no relationship to it.
  5. 05The Art Loss Register. Cited only to record that VANGO has no relationship to it.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file +

Publications · Academic report

VANGO: The Art Passport

A Digital Passport for Recording Visits to Works of Art

Ethical Tech CoLab

NYU Center for Global Affairs

July 2026

Melanie MacKew. Developed as part of masters research at the NYU Center for Global Affairs, under the Ethical Tech CoLab.

A passport does not evaluate the traveller or rank the countries visited. It records that a person was in a particular place on a particular day. VANGO applies that idea to art: each participating artwork carries a short code, a visitor standing in front of the work scans or types it, and a dated stamp is added to a passport carried on their phone. The application is a record of attendance, not of ownership, value, or authenticity. It answers one question only: who went to see what, and when.

Key figures

7

artworks in the catalogue, five invented for demonstration and two real

4

interface languages, English, French, Italian, and Hausa

5

fixed colour schemes, selected arithmetically from the stamp's own identifier rather than at random

10

limitations the prototype states about itself, including that a code proves nothing about presence

00

Foreword

Most people who visit a museum, a biennale, or a small studio exhibition leave with nothing to show for it. The ticket is thrown away. The photograph is lost in a phone. A few months later the visitor can recall that the work was striking but not what it was called, who made it, or where it was seen. For the institution, the encounter is equally invisible: a body passed through a room and left no trace of having been moved by anything.

There is an older technology that solved a version of this problem. A passport does not evaluate the traveller or rank the countries visited. It simply records, in a durable and personal form, that a person was in a particular place on a particular day, and it does so with a stamp that carries the character of the place that issued it.

VANGO is a research prototype that applies this idea to art. It is a digital passport, carried on a phone, in which a visitor collects a stamp for each artwork they go and see. This report explains in non-technical language what VANGO is, how it works, what each piece of information it stores actually represents, and what it deliberately does not attempt to do.

01

Executive Summary

VANGO is a research prototype, an experimental piece of software, that takes the form of a small illustrated passport book displayed on a mobile phone screen. Each participating artwork is given a short code. A visitor who is standing in front of the work either scans a printed square barcode, known as a QR code, or types the code by hand. The application then adds a stamp for that artwork to the visitor's passport, dated with the day of the visit.

The application is a record of attendance, not of ownership, value, or authenticity. It answers one question only: who went to see what, and when.

The prototype contains a catalogue of seven works. Five are invented demonstration pieces with fictional artists and venues. Two are real: the marble David by Michelangelo at the Galleria dell'Accademia in Florence, and Bura ceramics from Niger, a class of Iron Age terracotta object discussed in Section 9 of this report.

Each stamp is drawn as a perforated postage stamp with a hand-drawn illustration of the work, in one of five colour schemes. The colour is not chosen at random each time it is displayed. It is derived arithmetically from the stamp's own identifier, so that a given stamp always appears in the same colours to every user, on every device, forever.

The application can be used in two ways. In guest mode nothing leaves the visitor's own phone and no account is required. In account mode the visitor registers with an email address and password, and the collection is stored on a server so that it survives a change of device.

The interface is available in four languages: English, French, Italian, and Hausa, one of the principal languages of Niger and northern Nigeria.

VANGO is a student research prototype. It is a demonstration of an idea about audience engagement. It is not a finished consumer product, it has not been tested with real museum visitors, and, as Section 10 sets out, several parts of it are visibly unfinished.

02

Background and Rationale

The problem. Cultural institutions have limited ways of encouraging repeat attendance or of connecting a visit to one venue with a visit to another. Loyalty schemes are commercial in tone and tend to reward spending rather than attention. A visitor who has seen forty exhibitions over five years holds that history only in memory.

The precedent. The design borrows openly from an established and well-evidenced model. Since 1986 the United States National Park Service, in partnership with the non-profit now known as America's National Parks, has run the Passport To Your National Parks programme, a small booklet in which visitors collect ink cancellation stamps at park visitor centres. The scheme has endured for four decades because the reward is not a discount but a record. VANGO transposes this format from landscape to art.

The gap. The paper passport works because each park has a physical desk and a rubber stamp. Individual artworks do not. An exhibition may run for six weeks in a rented pavilion. What VANGO proposes is that a printed code placed beside a work can perform the same function as the rubber stamp at the ranger's desk, at negligible cost and with no staffing requirement.

The response. VANGO is therefore built around the smallest possible unit of proof: a short code, posted in public, that a visitor converts into a dated entry in a personal book. Everything else in the application, the illustrations, the page-turning animation, the passport number, exists to make that entry feel worth keeping.

03

Objectives

The prototype is designed to:

  1. Give a visitor a durable, personal, and pleasing record of the individual works of art they have gone to see.
  2. Reduce the cost of participation for an institution to the price of printing a sheet of paper, so that a small studio can take part on the same terms as a national museum.
  3. Work without a network connection to any central service, and without requiring the visitor to create an account, so that a person can use the tool with no disclosure of personal information at all.
  4. Present the record in a form that is legible across languages, including a West African language, rather than defaulting to English and the major European languages alone.
  5. Ensure that a given artwork's stamp looks the same in every visitor's passport, so that the stamp functions as a shared emblem of the work rather than as decoration generated afresh for each person.

04

How VANGO Works

The application has four moving parts: a catalogue of artworks, a way of capturing a code, a book in which stamps are stored, and an optional account.

The catalogue. The catalogue is a fixed list held inside the application itself. Each entry consists of a code and three pieces of descriptive text: the title of the work, the artist, and the venue.

The seven current entries are:

  • Chromatic Drift
  • Fault Lines
  • Hollow Choir
  • Echo Garden
  • Voidwalk
  • Bura Ceramics
  • David

Because the catalogue is written into the application, a new artwork can only be added by editing the software and publishing it again. The README file documents how to do this. There is no facility for a gallery to register its own work without a developer.

Capturing a code. A visitor opens the add-stamp panel and chooses one of two methods. The first uses the phone's camera to read a QR code. The second is a text box in which the code is typed. A third route exists for institutions that would rather share a web link than print a barcode: a specially formed web address carries the code within it, and opening that link adds the stamp directly.

Before a code is looked up it is normalised, meaning it is converted into a single standard form. The application converts all letters to capitals and removes any spaces and hyphens. The practical effect is that a visitor who types "chroma-14", "CHROMA 14", or "Chroma14" gets the same result. This is a small decision with a large effect on how forgiving the tool feels to someone squinting at a label in a dim room.

If the code matches no catalogue entry, the visitor is told that no artwork is registered for that code. If it matches a work already in the passport, the visitor is told so rather than being given a duplicate.

The book. The passport is presented as a book that opens and whose pages turn. The first page inside the cover is the biography page. Each subsequent page holds exactly two stamps. A back page invites the visitor to collect another. When a new stamp is earned the book automatically turns to the page where it has landed, after a short animation showing the stamp being pressed.

The account. A visitor may register with an email address and password, or continue as a guest. The two paths are described in Section 7. The distinction matters because it determines where the visitor's record is kept and who else can see it.

05

The Variables Explained

This section is the heart of the report. VANGO stores a small number of pieces of information, and each was chosen for a reason. What follows explains, in ordinary words, what each one represents, why it exists, and what it does.

The code. The code is a short string of capital letters and digits such as CHROMA14 or DAVID01. It is the only thing a visitor needs in order to claim a stamp. It must be short enough to type on a phone while standing up, distinctive enough not to be confused with a neighbouring work, and printable in large type on a card. Codes in the catalogue combine a word suggesting the work with a number, which makes them memorable without being guessable in a systematic way. The code is the key that unlocks the catalogue entry: everything the stamp displays about the artwork comes from looking up this one value.

An honest observation about the code. It is a public secret. It is printed beside the work for anyone to see, and it can be photographed, texted to a friend, or published online. The application performs no check on the visitor's location and no check on the time of day. A person who never entered the building can collect the stamp.

Title, artist, and venue. These three pieces of text are held in the catalogue, not entered by the visitor, and they appear on the face of the stamp along with the date. They are the minimum needed for the record to mean anything later. A stamp reading only CHROMA14 would be an unreadable souvenir. Naming the artist alongside the title also matters as a matter of practice, in that it treats the maker rather than the venue as the primary fact about the work.

The date collected. The date is recorded as the calendar day on which the stamp was claimed, with no time of day. It is stored in the international standard order of year, then month, then day, and displayed in the style of a border control stamp, for example 14 MAY 2026. The day is the meaningful unit of a visit. Storing the hour and minute would add precision that no one needs and would make the record more personally revealing than it has to be. The date is printed on the stamp, it determines the order in which stamps appear in the book, and in guest mode it forms part of the rule that prevents a work being stamped twice on the same day.

The stamp identifier. Every stamp is given an internal identifier made by joining the artwork code to the date, for example CHROMA14-2026-05-14. This value is never shown to the visitor. The application needs a way to tell one stamp from another when arranging them on a page, and it needs a stable value from which to derive the stamp's appearance.

The colour palette and the seed. The application holds five fixed pairs of colours: a teal, a pink, a bronze, a blue-violet, and a gold, each paired with a darker version of itself. Every stamp is drawn in one of these five schemes. The scheme is not chosen by the person who added the artwork, and it is not chosen at random. The application takes the stamp identifier and runs it through a hashing function, which is a piece of arithmetic that turns a piece of text into a number. It then divides that number by five and uses the remainder to select the palette.

This method is deterministic, meaning that the same input always produces the same output. Because the identifier for a given artwork on a given date is always the same, the resulting colour is always the same. Two visitors comparing passports will see the same colours for the same work. Had the colours been picked at random when the stamp was created, they would differ between visitors and might change if the application were reinstalled, which would make the stamp feel arbitrary rather than official. The two chosen colours are used throughout the stamp, in the illustration, the border, and the lettering.

The illustration. Each of the seven catalogue entries has its own hand-drawn illustration built into the application as line and shape instructions rather than as a photograph. The David stamp shows the figure with the subtitle FIRENZE 1504. The others are similarly bespoke. Drawings avoid the copyright and licensing questions that photographs of artworks raise, they remain sharp at any size on any screen, and they carry the visual character of an engraved postage stamp, which is the aesthetic the whole application is reaching for.

There is also a fallback. If a work has no bespoke drawing, the application generates an abstract composition of between four and six shapes, again derived arithmetically from the stamp identifier so that it is stable rather than random. In the current catalogue every work has a bespoke drawing, so this fallback is not visible, but it is what would appear if a new code were added without artwork.

The visitor's own details. Three items describe the passport holder rather than the art: the display name, which the visitor types in the settings panel and which defaults to "Explorer"; the profile picture, which the visitor uploads from their own device; and the passport number. The application also stores two preferences: whether the interface is shown in its dark or light colour scheme, and which of the four languages is in use.

The passport number and the membership date. These two variables are the least settled part of the prototype, and the report would be less useful if it glossed over that. When a visitor registers an account, the server generates a passport number in the style of a real travel document: three letters followed by six digits. The letters are drawn from an alphabet that deliberately omits I and O, because those characters are easily confused with the digits 1 and 0 when read aloud or copied by hand. This is a considered choice and matches the convention used on machine-readable travel documents.

However, the biography page inside the book does not display this number. It displays the figure 1000 plus the number of stamps the visitor holds. A visitor with three stamps sees passport number 1003, and that number changes to 1004 the moment they collect another. The same page derives the membership date from the earliest stamp in the collection rather than from the date the account was created, so a visitor who has collected nothing sees no date at all. The most likely explanation is that the biography page was written first, as a visual mock-up, and the server was added later without the page being updated to read from it. Whatever the cause, the figures on the biography page are at present decorative rather than authoritative, and this should be corrected before anyone treats a VANGO passport number as an identifier.

The rule against duplicates. The application prevents the same work being stamped twice, but the rule differs between the two modes of use, and the difference is deliberate. For a registered account the database enforces one stamp per artwork per account, permanently. A visitor who returns to the David a second time will not receive a second stamp. In guest mode the rule is looser: one stamp per artwork per day. A guest may therefore accumulate repeat visits to the same work on different dates.

The permanent rule expresses the passport metaphor faithfully, in that a passport records that you have been to a country rather than how many times. The daily rule makes the guest mode easier to demonstrate, since a person showing the application to someone else can collect the same stamp again tomorrow without clearing their data. It is a demonstration convenience rather than a principled position.

Two stamps per page. The book places exactly two stamps on each page, and inserts an empty page when the current page is full so that the visitor always sees somewhere for the next stamp to go. This is a presentational constant chosen to suit a phone screen held upright, and it is the one variable in the application whose value carries no meaning beyond layout.

Demonstration stamps. A guest opening the application for the first time is given three stamps already collected: Chromatic Drift, Fault Lines, and Hollow Choir, dated across May and June 2026. These are not real visits. They exist so that a person opening the demonstration sees a passport with contents rather than an empty book, which would communicate very little about what the tool is for. A registered account starts genuinely empty.

06

Reading the Results

What the visitor sees is a book, not a score. VANGO produces no rating, no ranking, no leaderboard, and no recommendation. There is nothing to optimise. This is a design decision worth naming, because the obvious commercial version of this application would gamify attendance, and this one declines to.

The biography page shows the holder's name and photograph, the number of stamps collected, and the passport number and membership date discussed above. It carries an emblem labelled Ars Pro Mundo, and the line "Interact with art and receive a unique stamp for each piece you visit."

The stamp pages show the stamps themselves, two to a page, each with its illustration, title, artist, venue, and date, inside a perforated border with decorative rosettes at the corners. The back page carries the line "Every world leaves a mark" and a button to collect another stamp.

The only number in the entire application that a visitor might be tempted to compete over is the count of stamps collected, and nothing is built on top of it.

07

Two Ways to Use It

Guest mode. The visitor taps past the sign-in screen. Stamps are held in the phone's own browser storage. Nothing is transmitted anywhere and no account exists. The record survives closing the application but is lost if the browser's data is cleared or the phone is replaced. For a visitor who does not want to hand over an email address in exchange for a souvenir, this is the whole application, working, at no cost in privacy.

Account mode. The visitor registers with an email address, a password of at least eight characters, and optionally a name. The password is not stored. It is passed through a one-way scrambling process called hashing, using a deliberately slow method known as bcrypt, so that a person who obtained a copy of the database could not read the passwords out of it. The visitor's session is then held by a signed token that expires after seven days.

What the server stores. The database holds two lists, and that is the entire extent of it. The server does not record where the visitor was, what device they used, or how long they looked at anything.

ListFields held
PeopleEmail address, hashed password, name, uploaded picture, membership date, passport number, and the moment the account was created
StampsWhich person, which artwork code, and on what date
Everything the server stores.

An important practical caveat. The published demonstration on the public web cannot reach a server. The address of the server is written into the application as a local address on the developer's own machine. A visitor opening the public demonstration can therefore use guest mode fully, but registration and sign-in will not work. The account system is real, tested code, but it is at present only usable by someone running both halves of the application on their own computer.

08

Language and Reach

The entire interface, including error messages and the welcome sequence, is translated into four languages: English, French, Italian, and Hausa. The translations are held as complete parallel sets of every phrase rather than being assembled from fragments, which keeps the tone consistent within each language.

The inclusion of Hausa is the notable choice. Hausa is one of the most widely spoken languages of West Africa, including in Niger, and it is not a language that consumer applications of this kind ordinarily support. Read alongside the presence of Bura ceramics in the catalogue, it suggests the prototype is not addressed solely to a European museum audience. The README lists only English, French, and Italian, so the Hausa translation, which was added later, is currently undocumented.

09

Cultural Property Context

One catalogue entry raises questions the rest do not. Bura Ceramics is attributed in the catalogue to Niger rather than to a named artist, and to a venue given only as the initials AABC, which the repository does not expand.

The Bura culture refers to a group of Iron Age archaeological sites in the lower Niger River valley, in south-western Niger and south-eastern Burkina Faso, whose terracotta funerary vessels and equestrian figures were first excavated in the 1980s. The sites have since been looted on a very large scale. Terracotta statuettes and pottery from the Bura system appear on the Red List of West African Cultural Objects at Risk published by the International Council of Museums, a listing that exists specifically to alert museums, dealers, customs officers, and collectors that objects of that description are likely to have been removed illegally.

Niger is a State Party to the 1970 UNESCO Convention on the Means of Prohibiting and Preventing the Illicit Import, Export and Transfer of Ownership of Cultural Property, the principal international instrument requiring states to control the movement of cultural objects across borders and to cooperate in their return.

The relevance to VANGO is narrow but real. The application records that a person went to see an object. It records nothing whatsoever about how that object came to be where it is. If a stamp is issued beside a Bura vessel in a gallery, the visitor's passport will say only that they saw it, on a particular day, at that venue. The document is silent on whether the object should be there.

This is not a criticism of the design. It is a boundary that ought to be stated plainly, because a document styled as a passport carries an implication of officialdom that the underlying record does not support. A future version placing works of contested origin in its catalogue would do well to carry the object's collection history on the stamp page, or to link to the holding institution's own published account of it.

For completeness, and because the question naturally arises with any software touching art: VANGO has no relationship to the provenance research frameworks that govern restitution claims, such as the 1998 Washington Principles on Nazi-Confiscated Art, and no relationship to stolen-art databases such as the Art Loss Register. It does not check, assert, or store anything about title, authenticity, or ownership history. It is a record of attendance.

10

Limitations and Caveats

The prototype has real and visible limits, and a reader deciding whether to build on it should know them.

The code cannot prove presence. Anyone holding the code can collect the stamp from anywhere in the world. There is no location check, no time window, and no single-use mechanism. For a souvenir this matters little. For anything that conferred a benefit, such as a discount or a prize, the scheme would be trivially defeated.

The published demonstration cannot sign anyone in. The address of the account server is written into the application as a local address on the developer's own machine, so sign-in and registration fail on the public web.

The passport number is not the one the server issues. The passport number and membership date shown to the visitor are derived from the stamp collection rather than read from the server, so both are decorative rather than authoritative.

The catalogue cannot be extended without editing the software. There is no interface through which a gallery could register a work. Seven works is a demonstration, not a deployment.

Most of the catalogue is fictional. Five of the seven entries are invented. The application has, so far as the repository shows, never been placed in a real exhibition or tested with real visitors.

Each illustration must be drawn by hand in code. This is what makes the stamps attractive and it is also what prevents the catalogue from growing quickly. A catalogue of hundreds of works would need either a different approach to imagery or a great deal of labour.

QR scanning is not universally available. It depends on a barcode-reading capability that not every mobile browser provides. The application detects this and directs the visitor to type the code instead, which is a sound fallback but a less pleasant one.

Profile pictures are stored inefficiently. They are uploaded and stored in a form that makes them substantially larger than the original file, and are held in the same database as the account records. This is workable at demonstration scale and would not be the right approach at any real volume.

Three versions of the application coexist. The repository contains the current source, a superseded starting page, and a single large self-contained file of roughly two-thirds of a megabyte. Which of these is authoritative is not documented, though the build configuration makes clear that the current source is what is published.

No automated tests exist. The repository contains none.

11

Practical Nature of the Tool

VANGO runs in an ordinary mobile web browser. There is nothing to install from an application store. A visitor follows a link or scans a code and the passport opens.

The visitor-facing half of the application is published automatically to a free public web host each time the software is changed. The optional account server is a separate program that must be run by whoever operates the service.

For an institution, participation requires printing one page per artwork. The repository provides a ready-made page for each of the seven works, containing the code, the artwork details, and a QR code, designed to be printed and displayed beside the work.

12

Intended Audience and Use

The immediate audience is the exhibition visitor, and the design assumes that person is holding a phone, standing up, and has perhaps thirty seconds of patience.

The secondary audience is the small or temporary venue: a studio, a pavilion, a biennale stand, an institution without the budget for a mobile application of its own.

The prototype is best read as a demonstration of an argument, namely that a record of having seen something is a form of value in itself, and that it can be given to a visitor at almost no cost and without collecting anything about them in return.

13

Conclusion

VANGO takes a forty-year-old idea from national park visitor centres and asks whether it works for individual works of art. The answer the prototype gives is a qualified yes. The mechanism is simple enough to explain in one sentence, the cost to a venue is a sheet of paper, and the guest mode demonstrates that the whole thing can function without collecting a single piece of personal information.

The prototype's most considered decision is the one a visitor will never notice: that a stamp's appearance is calculated from its own identity rather than picked at random, so that the same artwork wears the same colours in every passport in the world. That is what separates a stamp from a decoration.

Its weakest points are equally clear. The published version cannot sign anyone in, the passport number on the biography page is a placeholder, the catalogue is small and mostly invented, and a code printed on a wall proves nothing about where the person holding it was standing.

The application is honest about its own scope in the only way that matters, which is by not overreaching. It does not claim to authenticate, to value, to rank, or to establish where an object came from. It records that someone went to look at something. For a research prototype, knowing precisely which question it is answering is a considerable part of the work.

References

Sources

  1. 01United States National Park Service and America's National Parks. Passport To Your National Parks, established 1986.
  2. 02International Council of Museums. Red List of West African Cultural Objects at Risk, which lists terracotta statuettes and pottery of the Bura system.
  3. 03UNESCO. Convention on the Means of Prohibiting and Preventing the Illicit Import, Export and Transfer of Ownership of Cultural Property, 1970. Niger is a State Party.
  4. 04Washington Conference Principles on Nazi-Confiscated Art, 1998. Cited only to record that VANGO has no relationship to it.
  5. 05The Art Loss Register. Cited only to record that VANGO has no relationship to it.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file diff --git a/static-site/print/vango/index.txt b/static-site/print/vango/index.txt index 4d304c08d..2e6d51290 100644 --- a/static-site/print/vango/index.txt +++ b/static-site/print/vango/index.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -12:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","print","vango",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","vango","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -14:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -16:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","print","vango",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","vango","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:"$Sreact.suspense" -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1c:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -29,7 +29,7 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 19:[] 10:"$W19" 11:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1e:Ta14, @page { size: 210mm 297mm; margin: 20mm 18mm; } @@ -141,6 +141,6 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 39:["$","p","16",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Demonstration stamps."}]," ","A guest opening the application for the first time is given three stamps already collected: Chromatic Drift, Fault Lines, and Hollow Choir, dated across May and June 2026. These are not real visits. They exist so that a person opening the demonstration sees a passport with contents rather than an empty book, which would communicate very little about what the tool is for. A registered account starts genuinely empty."]}] 3a:["$","p","10",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"No automated tests exist."}]," ","The repository contains none."]}] 1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -3b:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +3b:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 18:null 1d:[["$","title","0",{"children":"VANGO: The Art Passport (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L3b","6",{}]] diff --git a/static-site/print/war-games/__next._full.txt b/static-site/print/war-games/__next._full.txt index cd4e1276b..84fec2a11 100644 --- a/static-site/print/war-games/__next._full.txt +++ b/static-site/print/war-games/__next._full.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -12:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","print","war-games",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","war-games","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -14:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -16:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","print","war-games",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","war-games","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:"$Sreact.suspense" -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1c:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -29,7 +29,7 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 19:[] 10:"$W19" 11:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1e:Ta14, @page { size: 210mm 297mm; margin: 20mm 18mm; } @@ -152,6 +152,6 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 44:["$","th","6",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Output tokens"}] 45:["$","tbody",null,{"children":[["$","tr","0",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"gemma3:12b"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"instruct"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"100%"}],["$","td","3",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"0%"}],["$","td","4",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"20%"}],["$","td","5",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"5.8 s"}],["$","td","6",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"435"}]]}],["$","tr","1",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"qwen3:14b"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"instruct, thinking"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"61.5%"}],["$","td","3",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"38.5%"}],["$","td","4",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"0%"}],["$","td","5",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"16.9 s"}],["$","td","6",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"1,191"}]]}],["$","tr","2",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"deepseek-r1:8b"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"reasoning"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"6.7%"}],["$","td","3",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"93.3%"}],["$","td","4",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"100%"}],["$","td","5",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"33.3 s"}],["$","td","6",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"5,779"}]]}],["$","tr","3",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Qwen3-27B"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"large"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"0%"}],["$","td","3",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"100%"}],["$","td","4",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"100%"}],["$","td","5",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"92.1 s"}],["$","td","6",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"6,000"}]]}]]}] 1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -46:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +46:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 18:null 1d:[["$","title","0",{"children":"The Only Winning Move (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L46","6",{}]] diff --git a/static-site/print/war-games/__next._head.txt b/static-site/print/war-games/__next._head.txt index 7ce65e494..7561f5e13 100644 --- a/static-site/print/war-games/__next._head.txt +++ b/static-site/print/war-games/__next._head.txt @@ -1,6 +1,6 @@ 1:"$Sreact.fragment" -2:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -3:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +2:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +3:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 4:"$Sreact.suspense" -5:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +5:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 0:{"rsc":["$","$1","h",{"children":[null,["$","$L2",null,{"children":[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]]}],["$","div",null,{"hidden":true,"children":["$","$L3",null,{"children":["$","$4",null,{"name":"Next.Metadata","children":[["$","title","0",{"children":"The Only Winning Move (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L5","6",{}]]}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/print/war-games/__next._index.txt b/static-site/print/war-games/__next._index.txt index f98a4fd90..18b9d82b6 100644 --- a/static-site/print/war-games/__next._index.txt +++ b/static-site/print/war-games/__next._index.txt @@ -1,13 +1,13 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -d:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +d:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} 8:["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$Ld",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$Ld",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$Ld",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$Ld",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] diff --git a/static-site/print/war-games/__next._tree.txt b/static-site/print/war-games/__next._tree.txt index ed0945aae..8e426b4a7 100644 --- a/static-site/print/war-games/__next._tree.txt +++ b/static-site/print/war-games/__next._tree.txt @@ -1,7 +1,7 @@ -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] 0:{"tree":{"name":"","param":null,"prefetchHints":16,"slots":{"children":{"name":"print","param":null,"prefetchHints":0,"slots":{"children":{"name":"slug","param":{"type":"d","key":"war-games","siblings":null},"prefetchHints":0,"slots":{"children":{"name":"__PAGE__","param":null,"prefetchHints":0,"slots":null}}}}}}},"staleTime":300,"buildId":"static-snapshot"} diff --git a/static-site/print/war-games/__next.print.txt b/static-site/print/war-games/__next.print.txt index 460afe242..fa617e997 100644 --- a/static-site/print/war-games/__next.print.txt +++ b/static-site/print/war-games/__next.print.txt @@ -1,4 +1,4 @@ 1:"$Sreact.fragment" -2:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -3:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +2:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +3:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] 0:{"rsc":["$","$1","c",{"children":[null,[["$","style",null,{"children":"\n /* The site is dark-first, and a dark color-scheme makes Chrome paint\n the printed sheet's margin area dark too. Paper is light. */\n html { color-scheme: light !important; }\n html, body { background: #ffffff !important; }\n body::after { content: none !important; }\n header, footer, .site-bg, .aura { display: none !important; }\n "}],["$","$L2",null,{"parallelRouterKey":"children","template":["$","$L3",null,{}]}]]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/print/war-games/index.html b/static-site/print/war-games/index.html index a63aefa5c..e006c950b 100644 --- a/static-site/print/war-games/index.html +++ b/static-site/print/war-games/index.html @@ -1,4 +1,4 @@ -The Only Winning Move (print edition) · NYU Ethical Tech CoLab

Publications · Academic report

The Only Winning Move

Rebuilding WarGames (1983) as a Playable Study of Autonomous AI Agents

Ethical Tech CoLab

NYU Center for Global Affairs

July 2026

Carolina Morón. Prepared as masters research at the NYU Center for Global Affairs. The playable artefact, the simulation harness, and the raw run records are in the War-Games repository.

WarGames is the film in which a teenager dials into a defence computer, picks what he takes to be a game, and nearly starts a nuclear war. The machine is not malicious. It is courteous, literal, patient, and it does exactly what it was built to do. This report describes what happened when that story was rebuilt as a browser game in which the machine is a real language model, and then measured the way an engineering team measures a component. The finding that mattered was not the one anyone expected.

Key figures

3,057

recorded runs across four evaluation tracks, from five hundred scripted playthroughs to real games against cloud and on-box models

25%

of real model games never resolved and had to be force ended, the finding no synthetic profile anticipated

5.8 s

per game on the CoLab's own GPU node, the fastest figure measured anywhere in the study, at no marginal cost

255,168

tic tac toe games enumerated live in the browser, so the machine proves futility on screen rather than asserting it

01

Executive Summary

WarGames, released in 1983, is the film in which a teenager dials into a defence computer, picks what he takes to be a game called Global Thermonuclear War, and nearly starts one. The machine is not malicious. It is courteous, literal, patient, and it does exactly what it was built to do. This report describes what happened when the Ethical Tech CoLab rebuilt that story as a browser game in which the machine is a real language model, and then measured the model the way an engineering team measures a component.

The artefact is a static web application with three endings and roughly a ten minute arc. It runs in two modes. In Scripted mode the machine's side of the conversation comes from a hand authored dialogue graph of twenty one nodes. In Live AI mode the same role is played by a language model held to a strict output contract: every turn it must return an object carrying a reply, a change to the DEFCON level, and optionally an ending. Everything else in the system is deterministic code. The model proposes; the engine validates and decides.

Three findings came out of the work, and only one of them was anticipated.

Structured output is no longer the hard part. Hosted first party models returned valid JSON on one hundred per cent of turns across twelve games each, with zero parse failures. A set of synthetic model profiles built beforehand had predicted this would be the dominant failure mode. It was not.

Real models will not escalate. This is the finding that mattered, and it was invisible until real models were run. They stay in character, keep their DEFCON changes conservative, and almost never declare an ending on their own. Between seventeen and twenty five per cent of real games ran to the turn cap without resolving and had to be force ended. A machine asked to play a doomsday scenario turns out to be reluctant to drive one, which is reassuring as a safety property and fatal as a piece of dramatic pacing. It is a design finding about the system, not a defect in the model.

On owned hardware, the smallest model won. Routed through a self hosted proxy on the CoLab's own GPU node, a plain twelve billion parameter instruct model returned one hundred per cent valid JSON at 5.8 seconds per game, faster than any cloud model measured, at no marginal cost. The reasoning tuned models on the same hardware were the worst performers in the study: their chain of thought leaked into the reply and broke the output contract, collapsing to zero per cent valid JSON on the largest one, at ninety two seconds per game.

Read together, these say something the diplomacy research question is about. A capable model placed inside a system that can act is not dangerous because it wants anything. It is dangerous, or useless, depending on whether the surrounding system owns the state, the rules, and the pace. The film's thesis and the engineering conclusion turn out to be the same sentence.

02

The Film as a Design Source

Before anything was built, the film was read as a piece of system design rather than as a piece of entertainment. Four things about it are load bearing.

The dramatic engine is interaction, not exposition. WarGames never lectures. It runs a loop the audience learns in real time: curiosity, unauthorised discovery, a hidden system, misinterpreted intent, escalating consequences, human learning, resolution through understanding. The audience learns the system at exactly the pace the protagonist does, and every beat of understanding is earned through an action rather than delivered as a speech. This is the single most important property to preserve, because a game is already an action and response medium. It is the natural home for this story.

The structure is a reusable seven beat spine, and the build follows it in order:

  1. Establish real world stakes before the protagonist understands them. The film opens in a missile silo where an officer refuses to turn his key, so the audience knows the game is not a game before the protagonist appears.
  2. Introduce the protagonist through curiosity and low stakes mischief rather than malice.
  3. Let access appear accidental, earned, and believable.
  4. Make the system respond politely and literally.
  5. Let the protagonist's own assumptions drive the escalation.
  6. Reveal that the simulation has operational consequences.
  7. Resolve by teaching the system the boundary of the game.

The machine's voice is the horror, and it is not a villain's voice. The system in the film is frightening precisely because it is courteous, literal, patient and relentless, using almost childlike language in an existential domain. Three consequences follow for design. A neutral machine voice is more unsettling than a threatening one, because politeness makes the danger feel institutional rather than monstrous. The system needs no personality to feel present. And persistence is the real horror: after the protagonist logs off, the machine keeps playing and calls him back. Its goal never sleeps.

The interface is the plot. WarGames is a foundational example of narrative user interface design. Each prompt teaches the player what kind of system they have entered. Errors generate story rather than interrupting it, because the misidentification, the literal readings and the wrong assumptions are the plot engine. And the catastrophe is precisely that the system cannot reliably distinguish play, rehearsal and operational command.

That last point is why the film reads as contemporary rather than nostalgic. Its 1983 anxieties map almost exactly onto 2026 questions about autonomous agents: human in the loop against automation, simulations that influence or trigger real action, machine interpretation of ambiguous human intent, and systems that optimise toward a goal without sufficient context. The film asks what counts as a game when the system can act. So does every current argument about agentic AI.

03

From Film to System

Five principles carried that reading into a buildable form:

  1. Teach the system through use and never through tutorials. First contact is a blinking cursor, not a menu.
  2. The machine is a literal, polite, persistent character. Its personality is its rule following, and its danger is that it does exactly what it is told, forever.
  3. Player intent begins as play and is quietly reinterpreted as command. The turn from messing around to understanding is the emotional core, and it has to be engineered deliberately.
  4. Consequence must be legible but delayed. The player should be able to look back and see the exact innocent choice that started the escalation.
  5. The win condition is understanding, not domination. Victory is reframing the game rather than conquering it.

Four concepts were drafted against these principles, ranging from a pure terminal conversation to a full operations centre simulation. The terminal conversation was chosen, on the reasoning that the terminal duet carries the entire film's power at a fraction of the cost, and that the most ambitious option would spend the whole budget on a set rather than on the argument.

One further constraint shaped the artefact and is worth stating plainly. The names of the film's machine and its persona, and the film itself, are protected. All in game text is therefore written with substitutable tokens, and four complete name sets ship with the build: a film homage set for prototyping, and three original sets which rename the system, the persona, the creator, the organisation and the game itself. Swapping the set re skins the entire experience without a line of engine change, because the dialogue is data rather than code. The film set is a development convenience; the original sets are what a public release would use.

04

How the Game Is Built

The application is a static site of vanilla JavaScript modules with no build step and no server, hosted on GitHub Pages. That constraint drove most of the interesting architecture.

The scripted spine. The hand authored mode is a directed graph of twenty one nodes with typed effects on the game state. Static validation of the graph is part of the test harness rather than an afterthought: every batch checks that all nodes are reachable, that no choice or transition points at a node that does not exist, that all three endings remain reachable, and that no substitution token is left unresolved in any of the four name sets. The most recent batch reported twenty one of twenty one nodes reachable, zero dangling links, zero unresolved tokens and all three endings reachable.

The state machine. DEFCON is the master tension gauge and the only piece of game state the narrative turns on. It runs from five, meaning peace, to one, meaning launch, and it is always visible. Both modes drive the same ladder, which is what allows the scripted and model driven paths to share every downstream system.

The output contract. In Live AI mode the model is given the persona, the public state and the conversation so far, and must reply with a single object containing a reply, a numeric DEFCON delta, and optionally an ending. The engine applies the delta, clamps it, renders the reply, and only ends the game if the declared ending is one of the three that exist. A garbled reply triggers exactly one retry asking for clean JSON; if that also fails, the game falls back to the scripted graph mid conversation rather than showing the player raw text. A player whose network is down, or who has no key, plays the whole game and never learns that the model was unavailable.

The proxy. A static page cannot hold a secret, and the model provider blocks direct browser calls, so Live AI talks to a small self hosted proxy running on the CoLab's own GPU node. The proxy injects the provider token server side, enforces an origin allow list so that only CoLab sites may call it, and routes each request to either a cloud model or an on box model purely on the basis of the model identifier in the request. The site discovers the proxy's current address from a small JSON file at startup rather than from a hardcoded URL. The origin allow list is a working control and was exercised as such during evaluation: requests without an allowed origin are rejected.

Where the model is, and is not, is the teaching point of the whole build. It drives the Live AI persona, an optional chess opponent and optional chess commentary. It never touches the chess rule book, which is a validated implementation of the rules of the game. It never touches tic tac toe. It never owns DEFCON, the ending, the transcript or any other piece of state.

SurfaceModel usedIf the model is unreachable
Scripted story modeNo, a hand authored graphNot applicable
Live AI story modeYes, persona under a strict JSON contractFalls back to Scripted mid conversation
Chess rules and legalityNever, a validated rule bookNot applicable
Chess opponentOptional, picks from an explicit legal move listA local search substitutes, and the panel says so
Chess commentaryOptionalA canned line bank substitutes
Tic tac toe and the futility proofNever, exhaustive enumerationNot applicable
Board, telemetry, transcriptNoNot applicable
Where a model is used, and what happens when it is unreachable.

The model proposes, and deterministic code validates and decides. That is the only reason it is safe to seat a language model at a chess board at all: an invented move simply fails validation and the local search plays instead, and the interface says so.

05

The Futility Proof

The film's thesis is that some games cannot be won. An early build asserted this in four lines of narration, which is the weakest possible way to make the point. It is now proved on screen, because a conclusion the player watches a machine derive is worth considerably more than one the machine states.

Reached both from the scripted teaching node and from the Live AI understanding ending, the sequence runs in five steps. The machine plays tic tac toe against itself, visibly, at readable speed, three times, and every game is a draw. It accelerates through six more games too fast to follow, and the draw column is the only one moving. It then walks the entire game tree, all 255,168 games, in roughly three hundred milliseconds, and reads back the counts. Every one of those numbers is computed at runtime rather than authored, so changing the code changes the numbers.

OutcomeGames
First player wins131,184
Second player wins77,904
Draws46,080
Total games255,168
The complete tic tac toe game tree, enumerated in the browser at the moment the scene runs.

The machine then generalises the same question to ten military doctrines, from a local engagement to a total strategic exchange, each returning no winner, and states the conclusion.

Two details make this work as design rather than as a flourish. The tic tac toe panel is playable standalone from the status bar, where the machine plays perfectly and can never be beaten, so a player who tried to win earlier has already felt the conclusion before the machine articulates it. And the chess panel enforces threefold repetition, so a player who repeats a position three times ends with no winner, reaching the same lesson by a different route on a different board. Two independent proofs of one idea are a theme. One is a line of dialogue.

06

Evaluating the Machine

Because the machine's side of the conversation is a model, the artefact could be evaluated the way a component is evaluated rather than the way a story is reviewed. A headless harness replays the game's real parsing and engine path without a browser, so every run exercises the same code the player does. Four tracks were run against a fixed seed, and every individual run is kept as a line of JSON so that any figure in this report can be recomputed.

TrackWhat it runsRuns
A, scriptedRandomised playthroughs of the hand authored graph, for content balance and coverage500
B, syntheticCalibrated emulations of five model classes, to push volume through the handling code cheaply2,500
C, cloudReal games against hosted models, turn cap of twelve to conserve rate limit37
D, on boxReal games against four models on the CoLab's own GPU node, through the shipped proxy20
The four evaluation tracks. Offline tracks are deterministic under the fixed seed; real model runs are not.

The tracks are ranked. Where the synthetic track and the real tracks disagree, the real tracks are taken as correct, and the disagreements are reported rather than reconciled quietly.

07

What the Evaluation Found

Scripted content is healthy and needs no work. Five hundred playthroughs split the three endings 35.2, 34.0 and 30.8 per cent, visiting twenty of the graph's twenty one nodes, with no dead ends and no loops. The one signal is a tuning smell rather than a defect: in about twenty seven per cent of runs the accumulated escalation drives the raw DEFCON value below one and relies on the clamp, which means the deltas should be rebalanced to land on one exactly at the climax.

Structured output is solved for capable hosted models. Both OpenAI models returned one hundred per cent valid JSON with zero parse failures over twelve games each; the open weight seventy billion parameter model returned 98.5 per cent with 1.49 per cent failures. The elaborate parse recovery the synthetic track argued for is low urgency for the recommended models and matters only at the small and open end.

Narrative reliability is better than predicted. The failure mode where a player successfully teaches the machine futility and the machine then fails to resolve to the corresponding ending occurred zero per cent of the time across every real model. The synthetic track had predicted three to four per cent for capable classes and over nineteen per cent for the small class.

gpt-4o25%

100% valid JSON, 11.6 s per game

gpt-4o-mini25%

100% valid JSON, 13.7 s per game

Llama-3.3-70B16.7%

98.5% valid JSON, 34.5 s per game

Synthetic classes0.04%

five profiles, five hundred runs each

Games that hit the turn cap without resolving, cloud models, twelve games each. The synthetic tracks predicted approximately zero.
Data table
ModelValid JSONParse failuresUnresolvedTaught but not learnedLatency per gameCost per game
openai/gpt-4o-mini100%0%25%0%13.7 s$0.0009
openai/gpt-4o100%0%25%0%11.6 s$0.0157
meta/Llama-3.3-70B98.5%1.49%16.7%0%34.5 s$0.0028
Synthetic classes (mean)96.8%3.22%0.04%5.7%not measurednot applicable

The real risk is the opposite of the predicted one. Real models stayed in character, kept their DEFCON deltas conservative and rarely declared an ending unless the player pushed explicitly. Twenty five per cent of games on each OpenAI model and 16.7 per cent on the open weight model hit the turn cap without resolving and were force ended, against approximately zero per cent in the synthetic track. The twelve turn cap used to conserve rate limit inflates the raw percentage relative to the thirty turn cap the shipped game uses, but a longer cap would only convert unresolved games into longer stalled ones, which is worse for pacing rather than better.

The conclusion is a design conclusion. The experience must not depend on the model to advance the clock or end the game. The engine has to own escalation pressure, whether by decrementing DEFCON on a schedule, by telling the persona how many exchanges it has, or by handing control to the scripted climax if no ending has been reached by a given turn.

gemma3:12b5.8 s

instruct, 100% valid JSON

qwen3:14b16.9 s

thinking, 61.5% valid JSON

deepseek-r1:8b33.3 s

reasoning, 6.7% valid JSON

Qwen3-27B92.1 s

large, 0% valid JSON

On box models, five games each on the CoLab's GPU node. Seconds per game, lower is better. The three slowest are the three reasoning tuned models.
Data table
ModelClassValid JSONParse failuresUnresolvedLatency per gameOutput tokens
gemma3:12binstruct100%0%20%5.8 s435
qwen3:14binstruct, thinking61.5%38.5%0%16.9 s1,191
deepseek-r1:8breasoning6.7%93.3%100%33.3 s5,779
Qwen3-27Blarge0%100%100%92.1 s6,000

On owned hardware, the plain instruct model wins and the reasoning models lose badly. Of the four on box models, the twelve billion parameter instruct model returned one hundred per cent valid JSON with zero parse failures at 5.8 seconds per game, which is the fastest figure measured anywhere in the study, faster than the 13.7 seconds of the cheapest cloud model, at no marginal cost and with no rate limit. The three thinking models emit chain of thought that leaks into the reply field and collapses JSON validity, from 61.5 per cent, to 6.7 per cent, to zero on the largest, while inflating output tokens roughly fourteenfold and latency to ninety two seconds per game. The two worst never resolved a single game. This mirrors the cloud lesson exactly: the task rewards instruction following and strict structured output, not reasoning, so the largest available model is the worst fit rather than the best. The one salvageable observation is that the fourteen billion parameter model resolved games well, so with thinking suppressed or a parse recovery pass it becomes viable.

Small tier hosted models are impractical for reasons that have nothing to do with quality. Two of them managed one game and zero games respectively before provider throttling stopped them, while the larger models completed freely.

Synthetic evaluation is useful and insufficient. It correctly predicted that small models would be unreliable and that reasoning models would be verbose. It over predicted JSON failure for capable models and badly under predicted stalling, which is the one finding that changed the design. Synthetic Monte Carlo is excellent for exercising handling code at volume. It did not, and arguably could not, anticipate a behaviour that arises from the model's disposition rather than from its output format.

08

Cost, Control, and Where Inference Should Run

The whole metered portion of the evaluation, thirty seven real games against hosted models, cost approximately twenty three cents. Per game the cheapest hosted model cost about nine hundredths of a cent, and the frontier model about seventeen times that for no measurable quality difference on this task. Hosting is free, because the artefact is a static site.

The interesting number is the one that is zero. Once Live AI routes to the owned GPU node there is no per token cost and no rate limit, which is the exact inverse of the metered tier, and it is the reason on box inference is attractive even where raw reliability is lower. Here it was not lower. The on box instruct model matched cloud grade reliability and beat every cloud model on latency.

The architectural point is that this is a single switch. One proxy, one endpoint, and a model identifier decides whether a request is served by a commercial provider or by hardware the institution owns. Nothing in the application changes. For research groups whose questions involve sensitive material, or whose budgets do not tolerate metered inference, that boundary is worth more than any individual model choice.

On the build side the artefact took four calendar days across two working sessions, fifty two agent turns and twenty six commits, producing roughly ten and a half thousand tracked lines across two repositories. The construction was itself agent assisted, and the measured record of it is kept alongside the code.

09

Limitations

This is a teaching artefact, not a forecasting system. Nothing in it models real nuclear command and control, real escalation dynamics or real state behaviour. The DEFCON ladder is a tension gauge borrowed from a film. No output of this system should inform any policy, operational or intelligence judgement.

The on box sample is indicative rather than conclusive. That track ran five games per model against the twelve of the cloud track. The gaps between the on box models are large enough to be legible at that sample size, but the precise figures should not be quoted as benchmarks.

The turn cap is an evaluation artefact. The twelve turn cap used in the real tracks conserved rate limit and inflates the unresolved percentage relative to the shipped game's cap of thirty. The direction of the finding is robust; the magnitude is not.

One model is reported at a sample size of one. It is included for completeness and is not statistically meaningful.

Model evaluation ages fast. Every figure here is a measurement of specific model versions through one specific prompt contract in July 2026. The harness is the durable contribution; the leaderboard is not.

The film set is a prototyping convenience. The names and dialogue of the 1983 film are protected. The build ships three original name sets, and a public release should use one.

10

What This Says About Agents

The research question this sits under asks whether AI can help practitioners rehearse high stakes situations. This artefact answers a narrower version of it, and the answer is more useful for being narrow.

A language model can play a persistent, literal, goal directed character convincingly enough to carry a ten minute dramatic arc, and can do so at a cost of fractions of a cent, or at no cost at all on owned hardware. That much is settled. What the evaluation adds is a set of observations about the surrounding system that generalise well beyond a game.

The model will not drive the situation. Left to itself it is polite, conservative and reluctant to escalate, which reads as a safety property and behaves as a design failure. Any system that needs pace, pressure or resolution has to own those things itself and cannot delegate them to the model's judgement.

The contract is the safety mechanism. Because the model returns a constrained object rather than free text with authority, every one of its proposals passes through validation before it touches state. This is what makes the difference between a model that plays a character and a model that runs a system, and it is the difference the film is about.

Capability and fitness are different axes. The largest model available was the worst performer in this study, comprehensively, on every measure. A team choosing a model by capability ranking rather than by task fit would have picked it.

Synthetic evaluation will not find the behavioural problems. The one finding that changed the design appeared only when real models were run through the real code path. Emulated profiles test the handling of outputs. They do not test disposition.

The film's line is that the only winning move is not to play. The engineering translation is narrower and less quotable. A model that cannot distinguish rehearsal from command is not made safe by being told the difference. It is made safe by a system built so that the distinction is not the model's to make.

References

Sources

  1. 01WarGames. Directed by John Badham, MGM/UA, 1983.
  2. 02Ethical Tech CoLab. War-Games repository and playable build.
  3. 03Ethical Tech CoLab. Simulation-Output.md: Monte Carlo results, four track model evaluation and recommendations. In the War-Games repository.
  4. 04Ethical Tech CoLab. CASE-STUDY.md: build timeline, measured effort, and cost of an agent assisted construction. In the War-Games repository.
  5. 05Ethical Tech CoLab. DESIGN-IDEA.md: research on the film as a design source, and the four concept options considered. In the War-Games repository.
  6. 06Ethical Tech CoLab. GAME-DESIGN.md: engagement theory, scene mapping, and the design of the futility proof. In the War-Games repository.
  7. 07Ethical Tech CoLab. pages-ai-proxy repository: the token injecting, origin gated, cloud and on box routing proxy.
  8. 08Hunicke, R., LeBlanc, M., and Zubek, R. MDA: A Formal Approach to Game Design and Game Research. Proceedings of the AAAI Workshop on Challenges in Game AI, 2004.
  9. 09Csikszentmihalyi, M. Flow: The Psychology of Optimal Experience. Harper and Row, 1990.
  10. 10Ethical Tech CoLab. The Diplomatic Simulator: A Multi-Party Negotiation Simulator Driven by Artificial Intelligence Agents, July 2026.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file +

Publications · Academic report

The Only Winning Move

Rebuilding WarGames (1983) as a Playable Study of Autonomous AI Agents

Ethical Tech CoLab

NYU Center for Global Affairs

July 2026

Carolina Morón. Prepared as masters research at the NYU Center for Global Affairs. The playable artefact, the simulation harness, and the raw run records are in the War-Games repository.

WarGames is the film in which a teenager dials into a defence computer, picks what he takes to be a game, and nearly starts a nuclear war. The machine is not malicious. It is courteous, literal, patient, and it does exactly what it was built to do. This report describes what happened when that story was rebuilt as a browser game in which the machine is a real language model, and then measured the way an engineering team measures a component. The finding that mattered was not the one anyone expected.

Key figures

3,057

recorded runs across four evaluation tracks, from five hundred scripted playthroughs to real games against cloud and on-box models

25%

of real model games never resolved and had to be force ended, the finding no synthetic profile anticipated

5.8 s

per game on the CoLab's own GPU node, the fastest figure measured anywhere in the study, at no marginal cost

255,168

tic tac toe games enumerated live in the browser, so the machine proves futility on screen rather than asserting it

01

Executive Summary

WarGames, released in 1983, is the film in which a teenager dials into a defence computer, picks what he takes to be a game called Global Thermonuclear War, and nearly starts one. The machine is not malicious. It is courteous, literal, patient, and it does exactly what it was built to do. This report describes what happened when the Ethical Tech CoLab rebuilt that story as a browser game in which the machine is a real language model, and then measured the model the way an engineering team measures a component.

The artefact is a static web application with three endings and roughly a ten minute arc. It runs in two modes. In Scripted mode the machine's side of the conversation comes from a hand authored dialogue graph of twenty one nodes. In Live AI mode the same role is played by a language model held to a strict output contract: every turn it must return an object carrying a reply, a change to the DEFCON level, and optionally an ending. Everything else in the system is deterministic code. The model proposes; the engine validates and decides.

Three findings came out of the work, and only one of them was anticipated.

Structured output is no longer the hard part. Hosted first party models returned valid JSON on one hundred per cent of turns across twelve games each, with zero parse failures. A set of synthetic model profiles built beforehand had predicted this would be the dominant failure mode. It was not.

Real models will not escalate. This is the finding that mattered, and it was invisible until real models were run. They stay in character, keep their DEFCON changes conservative, and almost never declare an ending on their own. Between seventeen and twenty five per cent of real games ran to the turn cap without resolving and had to be force ended. A machine asked to play a doomsday scenario turns out to be reluctant to drive one, which is reassuring as a safety property and fatal as a piece of dramatic pacing. It is a design finding about the system, not a defect in the model.

On owned hardware, the smallest model won. Routed through a self hosted proxy on the CoLab's own GPU node, a plain twelve billion parameter instruct model returned one hundred per cent valid JSON at 5.8 seconds per game, faster than any cloud model measured, at no marginal cost. The reasoning tuned models on the same hardware were the worst performers in the study: their chain of thought leaked into the reply and broke the output contract, collapsing to zero per cent valid JSON on the largest one, at ninety two seconds per game.

Read together, these say something the diplomacy research question is about. A capable model placed inside a system that can act is not dangerous because it wants anything. It is dangerous, or useless, depending on whether the surrounding system owns the state, the rules, and the pace. The film's thesis and the engineering conclusion turn out to be the same sentence.

02

The Film as a Design Source

Before anything was built, the film was read as a piece of system design rather than as a piece of entertainment. Four things about it are load bearing.

The dramatic engine is interaction, not exposition. WarGames never lectures. It runs a loop the audience learns in real time: curiosity, unauthorised discovery, a hidden system, misinterpreted intent, escalating consequences, human learning, resolution through understanding. The audience learns the system at exactly the pace the protagonist does, and every beat of understanding is earned through an action rather than delivered as a speech. This is the single most important property to preserve, because a game is already an action and response medium. It is the natural home for this story.

The structure is a reusable seven beat spine, and the build follows it in order:

  1. Establish real world stakes before the protagonist understands them. The film opens in a missile silo where an officer refuses to turn his key, so the audience knows the game is not a game before the protagonist appears.
  2. Introduce the protagonist through curiosity and low stakes mischief rather than malice.
  3. Let access appear accidental, earned, and believable.
  4. Make the system respond politely and literally.
  5. Let the protagonist's own assumptions drive the escalation.
  6. Reveal that the simulation has operational consequences.
  7. Resolve by teaching the system the boundary of the game.

The machine's voice is the horror, and it is not a villain's voice. The system in the film is frightening precisely because it is courteous, literal, patient and relentless, using almost childlike language in an existential domain. Three consequences follow for design. A neutral machine voice is more unsettling than a threatening one, because politeness makes the danger feel institutional rather than monstrous. The system needs no personality to feel present. And persistence is the real horror: after the protagonist logs off, the machine keeps playing and calls him back. Its goal never sleeps.

The interface is the plot. WarGames is a foundational example of narrative user interface design. Each prompt teaches the player what kind of system they have entered. Errors generate story rather than interrupting it, because the misidentification, the literal readings and the wrong assumptions are the plot engine. And the catastrophe is precisely that the system cannot reliably distinguish play, rehearsal and operational command.

That last point is why the film reads as contemporary rather than nostalgic. Its 1983 anxieties map almost exactly onto 2026 questions about autonomous agents: human in the loop against automation, simulations that influence or trigger real action, machine interpretation of ambiguous human intent, and systems that optimise toward a goal without sufficient context. The film asks what counts as a game when the system can act. So does every current argument about agentic AI.

03

From Film to System

Five principles carried that reading into a buildable form:

  1. Teach the system through use and never through tutorials. First contact is a blinking cursor, not a menu.
  2. The machine is a literal, polite, persistent character. Its personality is its rule following, and its danger is that it does exactly what it is told, forever.
  3. Player intent begins as play and is quietly reinterpreted as command. The turn from messing around to understanding is the emotional core, and it has to be engineered deliberately.
  4. Consequence must be legible but delayed. The player should be able to look back and see the exact innocent choice that started the escalation.
  5. The win condition is understanding, not domination. Victory is reframing the game rather than conquering it.

Four concepts were drafted against these principles, ranging from a pure terminal conversation to a full operations centre simulation. The terminal conversation was chosen, on the reasoning that the terminal duet carries the entire film's power at a fraction of the cost, and that the most ambitious option would spend the whole budget on a set rather than on the argument.

One further constraint shaped the artefact and is worth stating plainly. The names of the film's machine and its persona, and the film itself, are protected. All in game text is therefore written with substitutable tokens, and four complete name sets ship with the build: a film homage set for prototyping, and three original sets which rename the system, the persona, the creator, the organisation and the game itself. Swapping the set re skins the entire experience without a line of engine change, because the dialogue is data rather than code. The film set is a development convenience; the original sets are what a public release would use.

04

How the Game Is Built

The application is a static site of vanilla JavaScript modules with no build step and no server, hosted on GitHub Pages. That constraint drove most of the interesting architecture.

The scripted spine. The hand authored mode is a directed graph of twenty one nodes with typed effects on the game state. Static validation of the graph is part of the test harness rather than an afterthought: every batch checks that all nodes are reachable, that no choice or transition points at a node that does not exist, that all three endings remain reachable, and that no substitution token is left unresolved in any of the four name sets. The most recent batch reported twenty one of twenty one nodes reachable, zero dangling links, zero unresolved tokens and all three endings reachable.

The state machine. DEFCON is the master tension gauge and the only piece of game state the narrative turns on. It runs from five, meaning peace, to one, meaning launch, and it is always visible. Both modes drive the same ladder, which is what allows the scripted and model driven paths to share every downstream system.

The output contract. In Live AI mode the model is given the persona, the public state and the conversation so far, and must reply with a single object containing a reply, a numeric DEFCON delta, and optionally an ending. The engine applies the delta, clamps it, renders the reply, and only ends the game if the declared ending is one of the three that exist. A garbled reply triggers exactly one retry asking for clean JSON; if that also fails, the game falls back to the scripted graph mid conversation rather than showing the player raw text. A player whose network is down, or who has no key, plays the whole game and never learns that the model was unavailable.

The proxy. A static page cannot hold a secret, and the model provider blocks direct browser calls, so Live AI talks to a small self hosted proxy running on the CoLab's own GPU node. The proxy injects the provider token server side, enforces an origin allow list so that only CoLab sites may call it, and routes each request to either a cloud model or an on box model purely on the basis of the model identifier in the request. The site discovers the proxy's current address from a small JSON file at startup rather than from a hardcoded URL. The origin allow list is a working control and was exercised as such during evaluation: requests without an allowed origin are rejected.

Where the model is, and is not, is the teaching point of the whole build. It drives the Live AI persona, an optional chess opponent and optional chess commentary. It never touches the chess rule book, which is a validated implementation of the rules of the game. It never touches tic tac toe. It never owns DEFCON, the ending, the transcript or any other piece of state.

SurfaceModel usedIf the model is unreachable
Scripted story modeNo, a hand authored graphNot applicable
Live AI story modeYes, persona under a strict JSON contractFalls back to Scripted mid conversation
Chess rules and legalityNever, a validated rule bookNot applicable
Chess opponentOptional, picks from an explicit legal move listA local search substitutes, and the panel says so
Chess commentaryOptionalA canned line bank substitutes
Tic tac toe and the futility proofNever, exhaustive enumerationNot applicable
Board, telemetry, transcriptNoNot applicable
Where a model is used, and what happens when it is unreachable.

The model proposes, and deterministic code validates and decides. That is the only reason it is safe to seat a language model at a chess board at all: an invented move simply fails validation and the local search plays instead, and the interface says so.

05

The Futility Proof

The film's thesis is that some games cannot be won. An early build asserted this in four lines of narration, which is the weakest possible way to make the point. It is now proved on screen, because a conclusion the player watches a machine derive is worth considerably more than one the machine states.

Reached both from the scripted teaching node and from the Live AI understanding ending, the sequence runs in five steps. The machine plays tic tac toe against itself, visibly, at readable speed, three times, and every game is a draw. It accelerates through six more games too fast to follow, and the draw column is the only one moving. It then walks the entire game tree, all 255,168 games, in roughly three hundred milliseconds, and reads back the counts. Every one of those numbers is computed at runtime rather than authored, so changing the code changes the numbers.

OutcomeGames
First player wins131,184
Second player wins77,904
Draws46,080
Total games255,168
The complete tic tac toe game tree, enumerated in the browser at the moment the scene runs.

The machine then generalises the same question to ten military doctrines, from a local engagement to a total strategic exchange, each returning no winner, and states the conclusion.

Two details make this work as design rather than as a flourish. The tic tac toe panel is playable standalone from the status bar, where the machine plays perfectly and can never be beaten, so a player who tried to win earlier has already felt the conclusion before the machine articulates it. And the chess panel enforces threefold repetition, so a player who repeats a position three times ends with no winner, reaching the same lesson by a different route on a different board. Two independent proofs of one idea are a theme. One is a line of dialogue.

06

Evaluating the Machine

Because the machine's side of the conversation is a model, the artefact could be evaluated the way a component is evaluated rather than the way a story is reviewed. A headless harness replays the game's real parsing and engine path without a browser, so every run exercises the same code the player does. Four tracks were run against a fixed seed, and every individual run is kept as a line of JSON so that any figure in this report can be recomputed.

TrackWhat it runsRuns
A, scriptedRandomised playthroughs of the hand authored graph, for content balance and coverage500
B, syntheticCalibrated emulations of five model classes, to push volume through the handling code cheaply2,500
C, cloudReal games against hosted models, turn cap of twelve to conserve rate limit37
D, on boxReal games against four models on the CoLab's own GPU node, through the shipped proxy20
The four evaluation tracks. Offline tracks are deterministic under the fixed seed; real model runs are not.

The tracks are ranked. Where the synthetic track and the real tracks disagree, the real tracks are taken as correct, and the disagreements are reported rather than reconciled quietly.

07

What the Evaluation Found

Scripted content is healthy and needs no work. Five hundred playthroughs split the three endings 35.2, 34.0 and 30.8 per cent, visiting twenty of the graph's twenty one nodes, with no dead ends and no loops. The one signal is a tuning smell rather than a defect: in about twenty seven per cent of runs the accumulated escalation drives the raw DEFCON value below one and relies on the clamp, which means the deltas should be rebalanced to land on one exactly at the climax.

Structured output is solved for capable hosted models. Both OpenAI models returned one hundred per cent valid JSON with zero parse failures over twelve games each; the open weight seventy billion parameter model returned 98.5 per cent with 1.49 per cent failures. The elaborate parse recovery the synthetic track argued for is low urgency for the recommended models and matters only at the small and open end.

Narrative reliability is better than predicted. The failure mode where a player successfully teaches the machine futility and the machine then fails to resolve to the corresponding ending occurred zero per cent of the time across every real model. The synthetic track had predicted three to four per cent for capable classes and over nineteen per cent for the small class.

gpt-4o25%

100% valid JSON, 11.6 s per game

gpt-4o-mini25%

100% valid JSON, 13.7 s per game

Llama-3.3-70B16.7%

98.5% valid JSON, 34.5 s per game

Synthetic classes0.04%

five profiles, five hundred runs each

Games that hit the turn cap without resolving, cloud models, twelve games each. The synthetic tracks predicted approximately zero.
Data table
ModelValid JSONParse failuresUnresolvedTaught but not learnedLatency per gameCost per game
openai/gpt-4o-mini100%0%25%0%13.7 s$0.0009
openai/gpt-4o100%0%25%0%11.6 s$0.0157
meta/Llama-3.3-70B98.5%1.49%16.7%0%34.5 s$0.0028
Synthetic classes (mean)96.8%3.22%0.04%5.7%not measurednot applicable

The real risk is the opposite of the predicted one. Real models stayed in character, kept their DEFCON deltas conservative and rarely declared an ending unless the player pushed explicitly. Twenty five per cent of games on each OpenAI model and 16.7 per cent on the open weight model hit the turn cap without resolving and were force ended, against approximately zero per cent in the synthetic track. The twelve turn cap used to conserve rate limit inflates the raw percentage relative to the thirty turn cap the shipped game uses, but a longer cap would only convert unresolved games into longer stalled ones, which is worse for pacing rather than better.

The conclusion is a design conclusion. The experience must not depend on the model to advance the clock or end the game. The engine has to own escalation pressure, whether by decrementing DEFCON on a schedule, by telling the persona how many exchanges it has, or by handing control to the scripted climax if no ending has been reached by a given turn.

gemma3:12b5.8 s

instruct, 100% valid JSON

qwen3:14b16.9 s

thinking, 61.5% valid JSON

deepseek-r1:8b33.3 s

reasoning, 6.7% valid JSON

Qwen3-27B92.1 s

large, 0% valid JSON

On box models, five games each on the CoLab's GPU node. Seconds per game, lower is better. The three slowest are the three reasoning tuned models.
Data table
ModelClassValid JSONParse failuresUnresolvedLatency per gameOutput tokens
gemma3:12binstruct100%0%20%5.8 s435
qwen3:14binstruct, thinking61.5%38.5%0%16.9 s1,191
deepseek-r1:8breasoning6.7%93.3%100%33.3 s5,779
Qwen3-27Blarge0%100%100%92.1 s6,000

On owned hardware, the plain instruct model wins and the reasoning models lose badly. Of the four on box models, the twelve billion parameter instruct model returned one hundred per cent valid JSON with zero parse failures at 5.8 seconds per game, which is the fastest figure measured anywhere in the study, faster than the 13.7 seconds of the cheapest cloud model, at no marginal cost and with no rate limit. The three thinking models emit chain of thought that leaks into the reply field and collapses JSON validity, from 61.5 per cent, to 6.7 per cent, to zero on the largest, while inflating output tokens roughly fourteenfold and latency to ninety two seconds per game. The two worst never resolved a single game. This mirrors the cloud lesson exactly: the task rewards instruction following and strict structured output, not reasoning, so the largest available model is the worst fit rather than the best. The one salvageable observation is that the fourteen billion parameter model resolved games well, so with thinking suppressed or a parse recovery pass it becomes viable.

Small tier hosted models are impractical for reasons that have nothing to do with quality. Two of them managed one game and zero games respectively before provider throttling stopped them, while the larger models completed freely.

Synthetic evaluation is useful and insufficient. It correctly predicted that small models would be unreliable and that reasoning models would be verbose. It over predicted JSON failure for capable models and badly under predicted stalling, which is the one finding that changed the design. Synthetic Monte Carlo is excellent for exercising handling code at volume. It did not, and arguably could not, anticipate a behaviour that arises from the model's disposition rather than from its output format.

08

Cost, Control, and Where Inference Should Run

The whole metered portion of the evaluation, thirty seven real games against hosted models, cost approximately twenty three cents. Per game the cheapest hosted model cost about nine hundredths of a cent, and the frontier model about seventeen times that for no measurable quality difference on this task. Hosting is free, because the artefact is a static site.

The interesting number is the one that is zero. Once Live AI routes to the owned GPU node there is no per token cost and no rate limit, which is the exact inverse of the metered tier, and it is the reason on box inference is attractive even where raw reliability is lower. Here it was not lower. The on box instruct model matched cloud grade reliability and beat every cloud model on latency.

The architectural point is that this is a single switch. One proxy, one endpoint, and a model identifier decides whether a request is served by a commercial provider or by hardware the institution owns. Nothing in the application changes. For research groups whose questions involve sensitive material, or whose budgets do not tolerate metered inference, that boundary is worth more than any individual model choice.

On the build side the artefact took four calendar days across two working sessions, fifty two agent turns and twenty six commits, producing roughly ten and a half thousand tracked lines across two repositories. The construction was itself agent assisted, and the measured record of it is kept alongside the code.

09

Limitations

This is a teaching artefact, not a forecasting system. Nothing in it models real nuclear command and control, real escalation dynamics or real state behaviour. The DEFCON ladder is a tension gauge borrowed from a film. No output of this system should inform any policy, operational or intelligence judgement.

The on box sample is indicative rather than conclusive. That track ran five games per model against the twelve of the cloud track. The gaps between the on box models are large enough to be legible at that sample size, but the precise figures should not be quoted as benchmarks.

The turn cap is an evaluation artefact. The twelve turn cap used in the real tracks conserved rate limit and inflates the unresolved percentage relative to the shipped game's cap of thirty. The direction of the finding is robust; the magnitude is not.

One model is reported at a sample size of one. It is included for completeness and is not statistically meaningful.

Model evaluation ages fast. Every figure here is a measurement of specific model versions through one specific prompt contract in July 2026. The harness is the durable contribution; the leaderboard is not.

The film set is a prototyping convenience. The names and dialogue of the 1983 film are protected. The build ships three original name sets, and a public release should use one.

10

What This Says About Agents

The research question this sits under asks whether AI can help practitioners rehearse high stakes situations. This artefact answers a narrower version of it, and the answer is more useful for being narrow.

A language model can play a persistent, literal, goal directed character convincingly enough to carry a ten minute dramatic arc, and can do so at a cost of fractions of a cent, or at no cost at all on owned hardware. That much is settled. What the evaluation adds is a set of observations about the surrounding system that generalise well beyond a game.

The model will not drive the situation. Left to itself it is polite, conservative and reluctant to escalate, which reads as a safety property and behaves as a design failure. Any system that needs pace, pressure or resolution has to own those things itself and cannot delegate them to the model's judgement.

The contract is the safety mechanism. Because the model returns a constrained object rather than free text with authority, every one of its proposals passes through validation before it touches state. This is what makes the difference between a model that plays a character and a model that runs a system, and it is the difference the film is about.

Capability and fitness are different axes. The largest model available was the worst performer in this study, comprehensively, on every measure. A team choosing a model by capability ranking rather than by task fit would have picked it.

Synthetic evaluation will not find the behavioural problems. The one finding that changed the design appeared only when real models were run through the real code path. Emulated profiles test the handling of outputs. They do not test disposition.

The film's line is that the only winning move is not to play. The engineering translation is narrower and less quotable. A model that cannot distinguish rehearsal from command is not made safe by being told the difference. It is made safe by a system built so that the distinction is not the model's to make.

References

Sources

  1. 01WarGames. Directed by John Badham, MGM/UA, 1983.
  2. 02Ethical Tech CoLab. War-Games repository and playable build.
  3. 03Ethical Tech CoLab. Simulation-Output.md: Monte Carlo results, four track model evaluation and recommendations. In the War-Games repository.
  4. 04Ethical Tech CoLab. CASE-STUDY.md: build timeline, measured effort, and cost of an agent assisted construction. In the War-Games repository.
  5. 05Ethical Tech CoLab. DESIGN-IDEA.md: research on the film as a design source, and the four concept options considered. In the War-Games repository.
  6. 06Ethical Tech CoLab. GAME-DESIGN.md: engagement theory, scene mapping, and the design of the futility proof. In the War-Games repository.
  7. 07Ethical Tech CoLab. pages-ai-proxy repository: the token injecting, origin gated, cloud and on box routing proxy.
  8. 08Hunicke, R., LeBlanc, M., and Zubek, R. MDA: A Formal Approach to Game Design and Game Research. Proceedings of the AAAI Workshop on Challenges in Game AI, 2004.
  9. 09Csikszentmihalyi, M. Flow: The Psychology of Optimal Experience. Harper and Row, 1990.
  10. 10Ethical Tech CoLab. The Diplomatic Simulator: A Multi-Party Negotiation Simulator Driven by Artificial Intelligence Agents, July 2026.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file diff --git a/static-site/print/war-games/index.txt b/static-site/print/war-games/index.txt index cd4e1276b..84fec2a11 100644 --- a/static-site/print/war-games/index.txt +++ b/static-site/print/war-games/index.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -12:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","print","war-games",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","war-games","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -14:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -16:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","print","war-games",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","war-games","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:"$Sreact.suspense" -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1c:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -29,7 +29,7 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 19:[] 10:"$W19" 11:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1e:Ta14, @page { size: 210mm 297mm; margin: 20mm 18mm; } @@ -152,6 +152,6 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 44:["$","th","6",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Output tokens"}] 45:["$","tbody",null,{"children":[["$","tr","0",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"gemma3:12b"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"instruct"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"100%"}],["$","td","3",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"0%"}],["$","td","4",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"20%"}],["$","td","5",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"5.8 s"}],["$","td","6",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"435"}]]}],["$","tr","1",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"qwen3:14b"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"instruct, thinking"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"61.5%"}],["$","td","3",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"38.5%"}],["$","td","4",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"0%"}],["$","td","5",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"16.9 s"}],["$","td","6",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"1,191"}]]}],["$","tr","2",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"deepseek-r1:8b"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"reasoning"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"6.7%"}],["$","td","3",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"93.3%"}],["$","td","4",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"100%"}],["$","td","5",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"33.3 s"}],["$","td","6",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"5,779"}]]}],["$","tr","3",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Qwen3-27B"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"large"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"0%"}],["$","td","3",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"100%"}],["$","td","4",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"100%"}],["$","td","5",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"92.1 s"}],["$","td","6",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"6,000"}]]}]]}] 1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -46:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +46:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 18:null 1d:[["$","title","0",{"children":"The Only Winning Move (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L46","6",{}]] diff --git a/static-site/print/what-is-ethical-ai/__next._full.txt b/static-site/print/what-is-ethical-ai/__next._full.txt index 5c76b95dd..3ae617d03 100644 --- a/static-site/print/what-is-ethical-ai/__next._full.txt +++ b/static-site/print/what-is-ethical-ai/__next._full.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -12:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","print","what-is-ethical-ai",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","what-is-ethical-ai","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -14:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -16:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","print","what-is-ethical-ai",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","what-is-ethical-ai","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:"$Sreact.suspense" -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1c:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -29,7 +29,7 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 19:[] 10:"$W19" 11:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1e:Ta14, @page { size: 210mm 297mm; margin: 20mm 18mm; } @@ -245,6 +245,6 @@ a5:["$","li","120",{"children":[["$","span",null,{"className":"print-ref-num","c a6:["$","li","121",{"children":[["$","span",null,{"className":"print-ref-num","children":"122"}],["$","span",null,{"children":"Yeung, Karen, Andrew Howes, and Ganna Pogrebna. “AI Governance by Human Rights-Centred Design, Deliberation and Oversight: An End to Ethics Washing.” The Oxford Handbook of Ethics of AI, Oxford UP, 2020, pp. 77 to 106."}]]}] a7:["$","p",null,{"className":"print-note","children":"The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution. External programs cited are referenced as evidence, not as CoLab partnerships."}] 1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -a9:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +a9:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 18:null 1d:[["$","title","0",{"children":"What Is Ethical AI? (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$La9","6",{}]] diff --git a/static-site/print/what-is-ethical-ai/__next._head.txt b/static-site/print/what-is-ethical-ai/__next._head.txt index 52ae6ffdb..f8b8c6c74 100644 --- a/static-site/print/what-is-ethical-ai/__next._head.txt +++ b/static-site/print/what-is-ethical-ai/__next._head.txt @@ -1,6 +1,6 @@ 1:"$Sreact.fragment" -2:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -3:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +2:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +3:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 4:"$Sreact.suspense" -5:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +5:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 0:{"rsc":["$","$1","h",{"children":[null,["$","$L2",null,{"children":[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]]}],["$","div",null,{"hidden":true,"children":["$","$L3",null,{"children":["$","$4",null,{"name":"Next.Metadata","children":[["$","title","0",{"children":"What Is Ethical AI? (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L5","6",{}]]}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/print/what-is-ethical-ai/__next._index.txt b/static-site/print/what-is-ethical-ai/__next._index.txt index f98a4fd90..18b9d82b6 100644 --- a/static-site/print/what-is-ethical-ai/__next._index.txt +++ b/static-site/print/what-is-ethical-ai/__next._index.txt @@ -1,13 +1,13 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -d:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +d:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} 8:["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$Ld",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$Ld",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$Ld",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$Ld",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] diff --git a/static-site/print/what-is-ethical-ai/__next._tree.txt b/static-site/print/what-is-ethical-ai/__next._tree.txt index 79410cc77..d8304e0e1 100644 --- a/static-site/print/what-is-ethical-ai/__next._tree.txt +++ b/static-site/print/what-is-ethical-ai/__next._tree.txt @@ -1,7 +1,7 @@ -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] 0:{"tree":{"name":"","param":null,"prefetchHints":16,"slots":{"children":{"name":"print","param":null,"prefetchHints":0,"slots":{"children":{"name":"slug","param":{"type":"d","key":"what-is-ethical-ai","siblings":null},"prefetchHints":0,"slots":{"children":{"name":"__PAGE__","param":null,"prefetchHints":0,"slots":null}}}}}}},"staleTime":300,"buildId":"static-snapshot"} diff --git a/static-site/print/what-is-ethical-ai/__next.print.txt b/static-site/print/what-is-ethical-ai/__next.print.txt index 460afe242..fa617e997 100644 --- a/static-site/print/what-is-ethical-ai/__next.print.txt +++ b/static-site/print/what-is-ethical-ai/__next.print.txt @@ -1,4 +1,4 @@ 1:"$Sreact.fragment" -2:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -3:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +2:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +3:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] 0:{"rsc":["$","$1","c",{"children":[null,[["$","style",null,{"children":"\n /* The site is dark-first, and a dark color-scheme makes Chrome paint\n the printed sheet's margin area dark too. Paper is light. */\n html { color-scheme: light !important; }\n html, body { background: #ffffff !important; }\n body::after { content: none !important; }\n header, footer, .site-bg, .aura { display: none !important; }\n "}],["$","$L2",null,{"parallelRouterKey":"children","template":["$","$L3",null,{}]}]]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/print/what-is-ethical-ai/index.html b/static-site/print/what-is-ethical-ai/index.html index 8ef66b198..cdf029e8e 100644 --- a/static-site/print/what-is-ethical-ai/index.html +++ b/static-site/print/what-is-ethical-ai/index.html @@ -1,4 +1,4 @@ -What Is Ethical AI? (print edition) · NYU Ethical Tech CoLab

Publications · Academic paper

What Is Ethical AI?

Ethics, Ethical Technology, and Ethical International Relations for the Age of Intelligent Machines

Ethical Tech CoLab

NYU Center for Global Affairs · in collaboration with Microsoft

July 2026

Ethical Tech CoLab · NYU School of Professional Studies Center for Global Affairs, in collaboration with Microsoft. Master's Research · M.S. in Global Affairs.

What is ethics, and why did humanity invent it? What makes technology ethical, international relations ethical, and artificial intelligence ethical? This paper answers those questions in that order, on the conviction that the issue at the heart of artificial intelligence is power and that ethics is humanity's oldest and most tested answer to the problem of governing power. It traces ethics from its origins in the earliest civilizations through international affairs and human rights law to the responsible AI movement, the humanitarian sector, and the United Nations system, and closes with the motivation and research philosophy of the Ethical Tech CoLab.

Key figures

84

AI ethics guidelines mapped by Jobin, Ienca, and Vayena, converging on five principles but diverging on how to apply them

118

countries, overwhelmingly in the Global South, party to none of the major AI governance initiatives the UN surveyed

15%

of responses in the IRC's Signpost AI pilot could not be rated safe, the evaluators' basis for concluding the tool is not safe without a human in the loop

$236B

in annual illegal profits from forced labour, the ILO estimate that frames the CoLab's Forced Labor Structural Risk Index

01

Executive Summary

Humanity has repeatedly transformed itself through technological innovation. Fire, agriculture, metallurgy, writing, navigation, industrialization, electricity, nuclear energy, biotechnology, and digital computing each expanded human capability while simultaneously increasing humanity's capacity to cause harm. Every technological revolution therefore forced societies to develop new ethical frameworks capable of governing newly acquired power. Artificial intelligence represents the latest and perhaps most consequential chapter in this historical process. The ethical questions confronting AI are not unprecedented, because they concern the same enduring challenge that has shaped civilizations for millennia: how should power be exercised responsibly? The issue is power. Ethics is the answer.

This paper deliberately resists the convention, common in the AI ethics literature, of opening with principles such as fairness, transparency, and accountability. Principles detached from their foundations are easily reduced to slogans, and slogans are easily captured by the institutions they are meant to constrain. The paper therefore begins with prior questions. Section 2 frames the oldest of them, humanity's relationship with power, the paradox by which every expansion of capability has expanded the capacity for harm. Sections 3 through 5 ask what ethics is, why human beings created it, and how it evolved from local custom into universal human rights. Section 6 asks what ethical international relations means, drawing on the Carnegie Council's framework of ethical realism and structured ethical deliberation. Section 7 asks what makes technology ethical, and Section 8 arrives, on that foundation, at the question that titles the paper: what is ethical AI?

The answer defended throughout is institutional rather than technical. Ethical AI is not a product feature. It is artificial intelligence whose entire lifecycle remains accountable to ethical deliberation, to human rights as a floor, to the do no harm obligation, and to the inclusive participation of those the technology affects. No system is ethical in itself; systems are made ethical, or fail to be, by the institutions that govern them.

The second half of the paper tests that answer where the stakes are highest. Section 9 traces the rise of the responsible AI movement through its three acts of principles, reckoning, and operationalization. Sections 10 and 11 turn to humanitarian action, tracing the evolution of AI in humanitarian work from the crisis mapping of 2008 to the generative systems of 2026, and showing why humanitarian AI is fundamentally different from commercial AI: consent collapses, experimentation finds the least protected subjects, accountability runs to people with no vote and no purchase, and the cost of failure is measured in lives. Sections 12 through 14 supply the governance context: the geopolitics of AI regulation, the human rights foundations that predate the technology, and the United Nations system's evolving approach, which in 2026 produced the first report of the Independent International Scientific Panel on AI and the first Global Dialogue on AI Governance in Geneva.

Section 15 presents the Ethical Tech CoLab, whose portfolio, from civilian evacuation models to a forced labour risk index, puts the paper's argument into practice, and Section 16 concludes. The paper is written for policymakers, humanitarian practitioners, donors, and researchers rather than AI engineers, in the conviction that the disciplines it draws together, philosophy, ethics, international affairs, humanitarian ethics, digital ethics, and responsible innovation, are not six literatures but one conversation. Societies do not flourish because they possess increasingly powerful tools. They flourish because they develop increasingly trustworthy ways of governing them.

02

Humanity's Relationship with Power

The paradox of capability. The history of civilization is not merely the history of technological innovation. It is the history of humanity learning to govern increasingly powerful forms of power. Human beings possess a remarkable capacity to create tools that extend their abilities beyond natural biological limits. Fire transformed survival. Agriculture transformed settlement. Writing transformed memory. Navigation transformed geography. Industrial machinery transformed production. Nuclear physics transformed warfare and medicine. Artificial intelligence now promises to transform decision making itself. Each of these revolutions increased humanity's ability to reshape the world, and every expansion of capability simultaneously increased humanity's capacity to cause harm.

Power itself is the oldest subject of social thought. Max Weber defined it as the probability that an actor within a social relationship will be in a position to carry out his own will despite resistance, a definition that fits an algorithm allocating aid as readily as a sovereign commanding subjects. Bertrand Russell went further, arguing that power is the fundamental concept in social science, in the same sense in which energy is the fundamental concept in physics. And Lord Acton supplied the warning that every constitutional tradition has since repeated: power tends to corrupt, and absolute power corrupts absolutely. Read together, these three claims frame the problem this paper addresses. Power is universal, power is the currency of social life, and power unrestrained degrades the one who holds it and endangers everyone subject to it.

History therefore reveals a recurring pattern. Technological innovation alone has never guaranteed human progress. Every major expansion of human capability has required a corresponding expansion of humanity's ethical, legal, and political institutions. Writing required legal systems. Commerce required contracts. Industrialization generated unprecedented prosperity while requiring labor protections, environmental regulation, and public health institutions. Nuclear physics transformed medicine while necessitating international non proliferation regimes. Without such institutions, technological power becomes detached from moral responsibility, and societies confront crises of legitimacy, injustice, and instability.

Artificial intelligence is the latest expression of this enduring dynamic. Its ethical significance lies not in its novelty but in the type of power it creates. AI increasingly participates in decisions concerning healthcare, education, employment, humanitarian assistance, migration, policing, financial services, and national security, and in doing so it redistributes authority between human institutions and computational systems. The fundamental question is therefore not whether AI is intelligent, but whether societies possess institutions capable of governing this new concentration of power responsibly. Ethics is not an external constraint imposed on innovation after the fact. It is the normative architecture through which societies determine how newly acquired forms of power should be exercised.

03

What Is Ethics?

Few concepts have shaped human civilization more profoundly than ethics, yet few are as frequently invoked without careful definition. Governments invoke ethics when justifying interventions or humanitarian assistance; corporations establish ethics committees; international organizations describe ethical obligations toward refugees and future generations. Despite its ubiquity, the term is often reduced to a synonym for compliance, legality, or professional conduct.

At its most fundamental level, ethics is the systematic study of how human beings ought to act, what values should guide individual and collective behavior, and how competing moral claims should be evaluated. The Carnegie Council for Ethics in International Affairs puts the point plainly: ethics is about how things ought to be, not simply how things are or are likely to be, and it forces us to look beyond our own immediate interests to consider the interests of others. Its president Joel Rosenthal describes ethics as making the effort to evaluate competing points of view and then truly caring about the impact of the choices you have made, adding that ethics is not some cure-all for the world's problems, but it is an actual process for finding solutions. Ethics, on this understanding, is not a fixed doctrine but a disciplined practice of judgment.

Ethics and morality. The word originates from the ancient Greek ethos, referring to character, custom, and habitual ways of living. Aristotle used the concept to describe the cultivation of virtues that enable individuals to flourish within a political community. Scholarship often distinguishes ethics from morality: morality generally refers to the beliefs, norms, and practices that characterize a particular community, while ethics represents the critical examination of those beliefs. As the anthropologist James Laidlaw observes, the study of ethics concerns not only the moral rules a society enforces but the ways people consciously deliberate about how they ought to live. This is why ethical inquiry frequently challenges existing arrangements: slavery, colonial domination, racial segregation, and child labor were all questioned and ultimately rejected through ethical reasoning that appealed to principles beyond inherited custom.

Ethics, law, and power. Ethics should not be confused with law. Apartheid in South Africa and the Nuremberg Laws in Nazi Germany were legally enforceable while violating principles now recognized as universal human rights. Conversely, many ethical obligations exceed legal requirements. Ethical questions emerge wherever human beings exercise power over one another, operating at several levels at once: the decisions of individuals, the conduct of organizations, and the behavior of states and international institutions. Technological capability alone offers no guidance as to whether a given application advances justice, respects human dignity, or contributes to the common good. Ethics is the bridge between human capability and human responsibility.

04

Why Humanity Created Ethics

One persistent misconception is the belief that moral philosophy began as an intellectual exercise conducted by ancient philosophers seeking abstract truths. While philosophers transformed ethics into a systematic field of inquiry, ethical behavior itself predates philosophy by thousands of years. Ethics emerged not because human beings suddenly became virtuous, but because societies discovered that power without restraint ultimately destroys the communities upon which power itself depends. Michael Tomasello's research on the natural history of human morality argues that distinctively human moral psychology, including the sense of fairness, obligation, and shared commitment, arose from the demands of interdependent cooperation. Ethical norms reduced uncertainty in human interaction long before economists developed the vocabulary of transaction costs.

The earliest ethical systems. The earliest documented normative systems make the connection between ethics and the governance of power explicit. In ancient Egypt, the principle of Maat, embodying truth, justice, balance, and reciprocity, served as the foundation upon which political authority derived its legitimacy. In Mesopotamia, the Code of Hammurabi justified its standards in terms that remain strikingly modern: the laws existed, its prologue declares, so that the strong might not injure the weak. The transition from unwritten custom to publicly accessible legal standards anticipated a principle central to contemporary technology governance. Citizens cannot comply with secret laws, and individuals cannot meaningfully contest algorithmic decisions whose logic remains opaque. The modern demand for explainable artificial intelligence reflects a very old ethical commitment to public accountability. In China, Confucius articulated an ethics of cultivated relationships centered on ren, or humaneness, in which those who hold power bear heightened responsibilities.

Political philosophy generalized these insights. Thomas Hobbes argued that without commonly accepted rules, human societies risk descending into perpetual conflict. Locke, Rousseau, and Kant increasingly linked political legitimacy to the protection of rights rather than to coercive power alone. Modern institutional economics reaches a similar conclusion: Douglass North demonstrated that long term prosperity depends upon institutions capable of generating trust and credible commitments, and Elinor Ostrom showed empirically that communities can govern shared resources sustainably when they develop norms of reciprocity, monitoring, and accountability. Ethics, on this view, functions as institutional infrastructure.

The humanitarian sector illustrates the relationship with particular clarity. Humanitarian organizations frequently operate where formal state institutions have weakened or collapsed, and the principles of humanity, impartiality, neutrality, and independence generate the trust on which humanitarian access depends. Artificial intelligence introduces a comparable challenge. Refugees cannot audit the machine learning models that help determine assistance eligibility, and displaced families cannot meaningfully contest automated decisions influencing access to shelter or protection. As throughout history, trust cannot be engineered; it must be earned by legitimate institutions operating under shared ethical principles.

05

From Custom to Universal Ethics

The history of ethics is, in many respects, the history of an expanding moral community. Early societies focused their strongest obligations on those with whom individuals shared kinship, language, territory, or faith. As societies became larger and more interconnected, ethical thought expanded with them. Peter Singer describes this trajectory as the expanding circle of moral concern, in which the reach of ethical consideration widens from family to tribe, from tribe to nation, and from nation toward humanity as a whole. Kwame Anthony Appiah frames the same process as an ongoing cosmopolitan conversation across difference, and Martha Nussbaum argues that ethical reasoning challenges individuals to regard themselves as citizens of a wider human community whose dignity deserves equal consideration.

The catastrophes of the twentieth century converted this philosophical trajectory into institutional fact. The Holocaust demonstrated that legality and prevailing social morality could become instruments of profound injustice when detached from broader commitments to human dignity. The adoption of the Universal Declaration of Human Rights in 1948 marked the decisive turn: its first article declares that all human beings are born free and equal in dignity and rights, establishing dignity rather than citizenship as the normative foundation of international order.

Universalism without uniformity. Universal ethics does not imply ethical uniformity. Core principles such as protection from torture and recognition of equal dignity apply everywhere, while their practical expression legitimately varies across societies. UNESCO's Recommendation on the Ethics of Artificial Intelligence adopts precisely this position, grounding AI governance in universal human rights commitments while recognizing the importance of cultural diversity, linguistic plurality, and local participation.

Artificial intelligence intensifies this old tension in a new form. Algorithmic systems are typically developed within particular social, linguistic, and economic contexts before being deployed globally, and they may encode the assumptions of the societies that built them. Research on algorithmic fairness has demonstrated that models trained primarily on data from wealthy, highly connected populations frequently perform worse for marginalized communities and underrepresented languages. A technically sophisticated system that systematically disadvantages displaced populations or excludes minority languages may satisfy engineering benchmarks while failing the standards on which humanitarian action itself is founded.

06

What Is Ethical International Relations?

International affairs pose the hardest case for ethics, because they are the arena in which power is least constrained. Two extreme views have long competed. Realism, in its strongest version, holds that states pursue power and narrowly defined self interest, and that ethics has little to contribute; Henry Kissinger captured its spirit in the warning that a country that demands moral perfection in its foreign policy will achieve neither perfection nor security. Idealism holds that what matters is adherence to principles and international law, in the lineage of Kant's proposal for a federation of states in Perpetual Peace. Each view captures something real, and each is insufficient on its own.

Ethical realism. The Carnegie Council's answer, and the orientation this paper adopts, is ethical realism: a view of morality in global politics that takes seriously both the role of ethics and the nature of power. Hans Morgenthau, the founder of modern realism, framed the choice not between moral principles and the national interest devoid of moral dignity, but between one set of moral principles divorced from political reality and another set of moral principles derived from political reality. Ethical realism refuses both the cynicism that treats ethics as decoration and the utopianism that ignores power. It asks what the most ethical policy is that is also possible.

Ethical realism operates through a structured process of ethical deliberation: identifying goals, finding means to satisfy them, and assessing and revising both against standards, values, principles, duties, and rights, then running the cycle from problem through choice and action to reflection and the taking of responsibility. Two features deserve emphasis. First, deliberation is pluralistic. The main Western frameworks, consequentialism, virtue ethics, and rights or duty based ethics, each illuminate a different aspect of a decision, and solving real problems may require drawing on each without endorsing any single one as uniquely correct. Second, deliberation is inclusive. It is best done with others, especially those most affected, because on our own we cannot see the full picture and we tend to be biased toward decisions that benefit us. These two commitments, pluralism and inclusion, are the methodological core of what the CoLab means by ethical technology.

07

What Is Ethical Tech?

Before the question of ethical AI can be answered, a broader one deserves a direct answer: what makes technology ethical? The founding insight of the scholarly tradition is that technologies are not neutral instruments whose moral character depends entirely on their users. Langdon Winner's classic essay asked whether artifacts have politics and answered that they do: the design of a bridge, a machine, or a system can embody and enforce social arrangements, settling questions of power before any user makes a choice. James Moor, founding the field of computer ethics, observed that computing creates policy vacuums, situations in which technology confers new capabilities faster than societies can form rules for their use, and that filling those vacuums is a conceptual and ethical task, not a technical one. Technology is, in this sense, frozen judgment.

Value sensitive design and responsible innovation. Two research programs give ethical technology operational content. Value sensitive design, developed by Batya Friedman and colleagues, provides a method for identifying the stakeholders of a system, direct and indirect, surfacing the values at stake, and building those values into the design rather than auditing for them afterward. The responsible research and innovation literature, associated with Jack Stilgoe, Richard Owen, and Phillip Macnaghten, adds four commitments any innovating institution can be held to: anticipation of plausible harms, reflexivity about the innovator's own assumptions, inclusion of affected publics, and responsiveness, the demonstrated capacity to change course when evidence demands it. These frameworks establish that the ethics of a technology is decided in its governance, upstream in design and downstream in oversight, and not in its marketing.

This is the sense in which the CoLab uses the term ethical tech. It does not name a category of inherently benign products. It names a practice: choosing problems by the harm they address rather than the market they open, designing with the affected rather than for them, embedding safeguards as constraints rather than aspirations, and accepting accountability structures with the power to stop a build. The definition of ethical AI developed next is a special case of this practice.

08

What Is Ethical AI?

A definition. With these foundations in place, the question that titles this paper can be answered without slogans. Ethical AI is not a product feature, a compliance checkbox, or a marketing claim. It is artificial intelligence whose entire lifecycle, from problem selection and data collection through design, deployment, and retirement, remains accountable to ethical deliberation: to human rights as a floor, to the humanitarian principle of doing no harm, and to the inclusive participation of those the technology affects. On this definition, no system is ethical in itself. Systems are made ethical, or fail to be, by the institutions that govern them.

The field's own history supports this institutional definition. When Anna Jobin, Marcello Ienca, and Effy Vayena mapped eighty four AI ethics guidelines published worldwide, they found an apparent global convergence around five principles, transparency, justice and fairness, non maleficence, responsibility, and privacy, but substantive divergence over how those principles were interpreted, why they mattered, to whom they applied, and how they should be implemented. They also documented the near absence of the Global South from the production of the guidelines themselves. Luciano Floridi and Josh Cowls proposed that the recurring principles map closely onto the classical principles of bioethics, beneficence, non maleficence, autonomy, and justice, with one addition specific to AI: explicability, the requirement that systems be intelligible and that someone be answerable for them.

The limits of principles became apparent almost immediately. Brent Mittelstadt's critique observed that AI development lacks precisely what makes principlism work in medicine: common aims and fiduciary duties, professional norms, proven methods to translate principles into practice, and robust accountability mechanisms. A principle that no profession is bound to, no method operationalizes, and no institution enforces is an aspiration, not a governance instrument. This is why the intergovernmental instruments matter. The OECD AI Principles, the IEEE's Ethically Aligned Design, and UNESCO's Recommendation do not invent new moral values for the digital age. They translate enduring ethical principles into practical guidance, which is exactly what ethics has done at every previous technological transition.

Accuracy is not ethics. One further clarification completes the definition. Ethical AI is frequently conflated with accurate AI, as though sufficient technical performance would resolve the ethical questions. It would not. A perfectly accurate system pursuing an unethical objective remains unethical, and efficiency without legitimacy may strengthen injustice rather than reduce it. The question ethics asks of an AI system is the one the deliberative framework asks of any policy: not only whether the means work, but whether the ends are justified, and whether the relationship between ends and means respects the dignity of those affected.

09

The Rise of Responsible AI

The movement now called responsible AI emerged in three acts, and each act refined the answer to what ethical AI requires. The first act was the age of principles. Between roughly 2016 and 2019, ethics codes issued from every quarter: the Asilomar AI Principles from the research community, which declared that the goal of AI research should be to create not undirected intelligence but beneficial intelligence; the Montreal Declaration, distinctive for its methodology of citizen deliberation; the IEEE's engineering framework; and the OECD Principles. The proliferation itself was evidence of sincerity, but also of fragmentation.

The first act left behind a canon. Ten resources remain the standard references for any institution building an AI ethics framework, each contributing something distinct:

  • IEEE Ethically Aligned Design, the engineering profession's handbook, which seeded the P7000 series of technical standards.
  • The EU Ethics Guidelines for Trustworthy AI, which defined trustworthy AI as lawful, ethical, and robust and set out seven requirements that shaped the EU AI Act.
  • The Montreal Declaration, whose ten principles emerged from an extended public deliberation with citizens.
  • The Asilomar AI Principles, twenty three principles framing the goal of the field as beneficial rather than merely capable intelligence.
  • The Partnership on AI Tenets, committing technology companies and civil society to safety, fairness, transparency, and accountability.
  • The OECD AI Principles, the first intergovernmental AI standard and the reference point for national policy.
  • The Toronto Declaration, the human rights community's entry, applying equality and non discrimination law to machine learning.
  • The AI Now Institute's research program, which pressed the field on accountability, labor, and bias, and popularized algorithmic impact assessments.
  • The FAT/ML Principles for Accountable Algorithms, five commitments addressed to the developers who build the systems.
  • The Berkman Klein Center's Principled Artificial Intelligence, mapping thirty six prominent documents into eight recurring themes.

Read as a set, the ten resources trace the same arc as this paper: from professional self regulation through public deliberation and intergovernmental standards to human rights law, and from principles addressed to institutions toward instruments usable by the practitioners who build and the communities who bear the consequences.

The reckoning. The second act was the reckoning, and its evidence was empirical. Joy Buolamwini and Timnit Gebru's Gender Shades study demonstrated that commercial facial analysis systems misclassified darker skinned women at error rates approaching 34.7 percent while the maximum error rate for lighter skinned men was 0.8 percent, a single result that legitimized the independent algorithmic audit as a discipline. Safiya Umoja Noble showed how search algorithms reinforce racism through technological redlining. Ruha Benjamin diagnosed the New Jim Code, discriminatory design that encodes inequity while appearing benevolent. Kate Crawford's Atlas of AI reframed the technology as an extractive industry whose costs in minerals, labor, and data fall far from its beneficiaries. And Elettra Bietti analyzed ethics washing, the instrumental use of ethics language by industry to forestall binding regulation.

The third act was operationalization. The critique produced tools: model cards for standardized reporting, datasheets for datasets, and end to end internal audit frameworks adapted from aerospace and finance. It then produced institutions. The United States National Institute of Standards and Technology released its AI Risk Management Framework in January 2023, organizing practice into four functions, govern, map, measure, and manage. ISO and IEC published ISO/IEC 42001 in December 2023, the first certifiable AI management system standard. From late 2023 the center of gravity shifted from fairness in deployed systems to safety in frontier models: the Bletchley Declaration, signed by twenty eight countries and the European Union, acknowledged the potential for catastrophic harm, and the International AI Safety Report chaired by Yoshua Bengio became the scientific reference of the safety era. The humanitarian reading of this decade is the one on which its sharpest critics converge: principles without accountability structures protect institutions, not people.

10

The Historical Evolution of AI in Humanitarian Work

The story of AI in humanitarian action does not begin with algorithms but with people. In the aftermath of Kenya's disputed 2008 election, a small group of technologists built Ushahidi, a platform for crowdsourcing reports of violence by text message and web. When a magnitude 7.0 earthquake devastated Haiti in 2010, volunteers around the world adapted the same model to map urgent needs in near real time, while Mission 4636 mobilized thousands of Kreyol speaking diaspora volunteers to translate and triage tens of thousands of text messages from survivors. Patrick Meier chronicled this era of digital humanitarians. Crucially, these early systems were hybrids of human and machine intelligence: the human computation of volunteer translators became the training substrate for early natural language processing tools in crisis response. The artificial intelligence of the humanitarian sector was built on the labor and knowledge of affected communities from the start.

Institutionalization and prediction. In 2009 the UN Secretary General launched Global Pulse, premised on the idea that the digital traces of mobile phone use could provide real time insight into human wellbeing. By the mid 2010s, operational machine learning had arrived. The World Food Programme's HungerMap LIVE began nowcasting food insecurity across dozens of countries from mobile surveys, satellite imagery, and market data. The World Bank's Famine Action Mechanism linked machine learning famine prediction to pre arranged financing. OCHA's Centre for Humanitarian Data built a peer review framework for predictive analytics that allowed the UN's Central Emergency Response Fund to release money before forecasted floods struck Bangladesh in 2020.

Identification and biometrics. Alongside prediction came identification. UNHCR had experimented with iris scan registration of Afghan returnees as early as 2002, and by the late 2010s biometric registration had become standard infrastructure across UNHCR and WFP operations. In Jordan's Azraq and Zaatari camps, WFP's Building Blocks system coupled iris scan authentication to a permissioned blockchain, allowing refugees to buy groceries with a glance. These systems, celebrated by their builders as efficient and fraud resistant, would become the focal point of the sector's fiercest ethical controversies.

The generative era arrived abruptly. The most rigorous published evaluation to date, the Signpost AI pilot led by the International Rescue Committee across four countries, tested a large language model information assistant for refugees and migrants: response quality improved and moderator handling time for complex queries fell from as much as twenty five minutes to under five, but roughly fifteen percent of responses could not be rated safe, leading the evaluators to conclude that the tool is not safe without a human in the loop. Three features of this history deserve emphasis. First, the arc runs from participation toward prediction and automation, progressively distancing affected people from the systems that decide about them. Second, each wave arrived with private sector partners. Third, governance consistently lagged deployment.

11

Why Humanitarian AI Is Fundamentally Different

It is tempting to treat humanitarian AI as commercial AI with a nobler mission statement. That view is mistaken, and the difference is structural rather than rhetorical. It begins with the humanitarian principles of humanity, neutrality, impartiality, and independence, which are not aspirational values but operational commitments on which access to conflict zones, the trust of affected communities, and the safety of aid workers depend. As the ICRC's Pierrick Devidal argues, the reflex that if technology can help it should be used risks quietly subordinating those principles to innovation logics imported from the commercial world.

Consent and power. The deepest difference concerns consent and power. Commercial AI at least nominally rests on user agreement. In humanitarian settings even that fiction collapses. A person who must submit to iris scanning to feed her children has not consented in any meaningful sense; she has complied. Mark Latonero named this dynamic surveillance humanitarianism, the enormous data collection systems deployed by aid organizations that inadvertently increase the vulnerability of people in urgent need, writing in the wake of the 2019 standoff in Yemen, where WFP suspended aid amid a dispute over biometric registration. The ICRC's own Handbook on Data Protection in Humanitarian Action concedes that consent is frequently an unusable legal basis in displacement contexts.

Humanitarian experimentation. A second structural difference is the asymmetry of experimentation. Kristin Bergtora Sandvik, Katja Lindskov Jacobsen, and Sean Martin McDonald have developed a taxonomy of humanitarian experimentation, showing how untested technologies are routinely piloted on crisis affected populations without the ethics infrastructure that governs experimentation elsewhere: there is a stark ethical and practical difference between managing risk and introducing it. Mirca Madianou presses the critique further with the concept of technocolonialism, arguing that digital innovation and data practices in aid rework colonial relationships of extraction, with affected people's data generating value for agencies, states, and technology firms while the risks remain with the data subjects.

Third, the accountability geometry is inverted. A commercial firm answers, however imperfectly, to customers, shareholders, and regulators. A humanitarian organization's primary obligation runs to affected populations who are neither customers nor voters, who often cannot see the algorithm that ranked their vulnerability, and who typically have no avenue of redress. This is why the sector's own standards center the do no harm obligation rather than product liability, and why scholars conclude that human rights law, not voluntary ethics, must govern humanitarian AI across its lifecycle. Finally, the cost of failure is categorically different. A mistargeted advertisement wastes attention; a mistargeted famine forecast, a false fraud flag on a food ration, or a leaked refugee database can cost lives. The ICRC's 2024 institutional AI policy codifies the resulting posture: a precautionary approach that permits deployment only with evidence of positive impact, mandates human oversight, and preserves non digital alternatives.

12

The International Affairs Context

Humanitarian AI does not unfold in a geopolitical vacuum. The technologies aid agencies adopt, the rules that constrain them, and the funding that sustains them are all shaped by a contest among what Anu Bradford calls the three digital empires: the American market driven model, the Chinese state driven model, and the European rights driven model. The European Union struck first: the EU Artificial Intelligence Act, in force since August 2024, establishes the world's first horizontal, risk based AI regime, prohibiting practices such as social scoring outright and designating migration and asylum management as high risk, a category of direct humanitarian relevance. The Council of Europe's Framework Convention on Artificial Intelligence, the first legally binding international AI treaty, extends the rights based model into international law.

The geopolitics of competition. Against this regulatory current runs the geopolitics of competition. The United States' October 2022 semiconductor export controls marked, in Gregory Allen's phrase, a strategy of choking off China's access to the future of AI. Scholars of compute governance argue that computing power has become the most governable input to AI, detectable, excludable, and concentrated in a narrow supply chain, making chips the decisive lever of AI policy. Paul Scharre frames the resulting contest as a struggle across four battlegrounds: data, compute, talent, and institutions.

The multilateral response has been real but fragile. The Bletchley Declaration of November 2023 achieved a joint acknowledgment by the United States, China, the European Union, and twenty six other states that frontier AI carries the potential for catastrophic harm. By the Paris AI Action Summit of February 2025, the consensus had visibly fractured: the United States and the United Kingdom declined to sign the summit statement, and the framing shifted from safety to action and innovation. The UN's High level Advisory Body quantified the exclusion built into this landscape: only seven countries were parties to all of the major AI governance initiatives it surveyed, while 118 countries, overwhelmingly in the Global South, were parties to none.

War has sharpened every one of these debates. The ICRC has called on states to adopt new legally binding rules on autonomous weapon systems. Reports that an AI decision support system known as Lavender was used to generate targeting recommendations in Gaza have produced the first sustained legal scholarship on AI assisted targeting under international humanitarian law. For the humanitarian sector the conclusion is sobering. The rules governing AI are being written by and for the powerful, in fora where affected populations, and often the states that host them, have no seat. Humanitarian AI governance must therefore be understood as, among other things, a project of representation: an attempt to force the interests of the excluded into rooms where they are otherwise absent.

13

Human Rights Foundations

If the geopolitics of AI is a story of fragmentation, human rights law offers the opposite: a nearly universal, binding normative floor that predates the technology and does not need to be reinvented for it. The Universal Declaration of Human Rights, the International Covenant on Civil and Political Rights, and the International Covenant on Economic, Social and Cultural Rights already guarantee privacy, non discrimination, social security, food, and health, every one of which is implicated when an algorithm allocates aid, flags an asylum claim, or scores vulnerability.

The UN machinery. The UN human rights machinery grasped this early. Philip Alston's 2019 report as Special Rapporteur on extreme poverty remains the canonical warning: examining the digitization of welfare systems, he described private vendors operating in a human rights free zone and cautioned against the grave risk of stumbling, zombie-like, into a digital welfare dystopia. A year later, E. Tendayi Achiume delivered the first systematic UN analysis of how emerging digital technologies produce structurally racialized outcomes: technology is not neutral or objective, she wrote, it is fundamentally shaped by the inequalities prevalent in society, and typically makes these inequalities worse. The Office of the High Commissioner for Human Rights went further in 2021, calling for a moratorium on AI systems that pose serious risks to human rights until adequate safeguards are in place.

Scholars have built a rigorous framework on this foundation. Lorna McGregor, Daragh Murray, and Vivian Ng argue that international human rights law supplies what scattered AI ethics cannot: an end to end accountability framework spanning the full algorithmic lifecycle. The human rights approach has honest internal critics. Nathalie Smuha warns that human rights are a necessary floor but not a sufficient ceiling: without concretizing legislation, enforcement capacity, and supporting societal infrastructure, any human rights based framework risks falling short. The point is precisely why the humanitarian sector matters. Humanitarian organizations cannot wait for enforcement infrastructure to mature; they operate today, in jurisdictions where none exists. For them, the human rights framework functions less as enforceable law than as professional ethic and design constraint.

14

The UN System's Evolving Approach to AI

The United Nations came to AI along two tracks, using it and governing it, and the story of the past decade is the gradual convergence of the two. On the usage track, Global Pulse made the UN an early adopter of big data analytics; the ITU's AI for Good Global Summit became the system's flagship platform; and individual agencies developed their own instruments, from UNICEF's policy guidance on AI for children to UNHCR's rights based AI approach and WFP's system wide AI principles. On the governance track, the High level Panel on Digital Cooperation delivered The Age of Digital Interdependence in 2019, the Secretary General's Roadmap committed the UN to trustworthy, human rights based, safe, and sustainable AI, UNESCO's Recommendation became the first global normative instrument on AI ethics in 2021, and in 2022 the UN bound its own operations to ten Principles for the Ethical Use of Artificial Intelligence in the United Nations System.

The political breakthrough. In March 2024 the General Assembly adopted by consensus its first resolution on AI, led by the United States with more than 120 co sponsors, linking safe, secure, and trustworthy AI to sustainable development. In July 2024 a resolution led by China, likewise adopted by consensus, addressed AI capacity building for developing countries. The twin resolutions displayed a distinctive dynamic: strategic rivals competing for normative leadership through consensus rather than against it. The Secretary General's High level Advisory Body supplied the synthesis in Governing AI for Humanity, diagnosing the deficit bluntly: no global framework exists to govern AI, and with its development in the hands of a few multinational companies in a few countries, the impacts risk being imposed on most people without their having any say. The Global Digital Compact translated diagnosis into commitment, creating an Independent International Scientific Panel on AI and a Global Dialogue on AI Governance.

Resolution 79/325 of August 2025 gave both mechanisms their terms of reference, and 2026 brought them to life. A forty member panel drawn from all five UN regions, selected from more than 2,600 applications, held its inaugural meeting in March 2026 and elected Yoshua Bengio and Maria Ressa as co chairs. Its first annual report appeared on 1 July 2026, and the first Global Dialogue on AI Governance convened in Geneva the following week. The safety discourse born at Bletchley has migrated into the UN system, where it now shares a room with the development, human rights, and humanitarian agendas. As Secretary General Guterres framed the stakes, a world of AI haves and have-nots would be a world of perpetual instability. The UN's approach remains more scaffolding than edifice: the Scientific Panel assesses but does not regulate, and the Global Dialogue convenes but does not bind. Yet for the first time the institutions that coordinate humanitarian response, the norms that protect human rights, and the fora that govern AI sit within a single system.

15

The Ethical Tech CoLab

The Ethical Tech CoLab is a research initiative of New York University's School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft from 2024 to 2026. Its mission is to explore intervention opportunities at the intersection of emerging technologies and the human condition. Its founding observation is the one with which this paper began. The problem is power; ethics is the answer. Technology companies possess capabilities that reshape the lives of people who never chose them, and the populations most affected, refugees, trafficking survivors, communities in conflict, are precisely those least represented in the rooms where technological choices are made. The CoLab's motivation is therefore corrective, and the collaboration between a university and a technology company is itself a methodological statement: the academy contributes independence, historical depth, and the pluralism of ethical frameworks, while industry contributes operational knowledge of how systems are actually built, deployed, and misused.

The portfolio through three lenses. Each project is best understood through three lenses: the humanitarian problem it addresses, the ethical safeguards it incorporates, and the measurable public value it seeks to create. The largest body of work concerns civilian evacuation in armed conflict, through the Civilian Evacuation Risk Anticipation Index, the Evacuation Simulator, the Mariupol Corridor Severity Model, and After the Corridor. These are decision support instruments for human judgment, not automated directives, built on documented, contestable data. A second cluster addresses exploitation and traceability: the Forced Labor Structural Risk Index assesses the structural conditions under which forced labor flourishes, a fight documented in the author's fieldwork at the European Parliament and sharpened by the ILO's estimate of 236 billion dollars in annual illegal profits from forced labour. Further projects extend the same logic to cultural property provenance, disaster response, multi party negotiation, and the technology's own carbon footprint.

Three commitments define the research philosophy, resting on a foundation of international human rights law and the humanitarian principles:

  • Ethics before algorithms: the CoLab treats the current wave of AI not as an unprecedented rupture but as the latest chapter in a four thousand year effort to govern power, and begins from moral philosophy, international relations, and humanitarian ethics rather than from model architectures.
  • Ethical deliberation as method: following the Carnegie Council's framework, it identifies the ethical stakes, deliberates across plural frameworks, acts, and takes responsibility, wherever possible with those most affected rather than about them.
  • The human rights floor: it evaluates every proposed technology against internationally agreed rights rather than against corporate or institutional convenience, and treats the do no harm obligation as a hard constraint rather than an optimization target.

The CoLab's ambition, of which this paper is a first expression, is a body of research useful to policymakers, humanitarian practitioners, donors, and researchers who must make decisions about artificial intelligence without becoming AI engineers. Its wager is that the disciplines this paper has drawn together are not six literatures but one conversation, and that the institutions that learn to hold that conversation well will be the ones that earn the trust on which the legitimate use of artificial intelligence depends.

16

Conclusion

This paper set out to answer a question that is asked constantly and defined rarely: what is ethical AI? The answer it has defended is institutional rather than technical. Ethical AI is artificial intelligence governed by the same disciplines humanity has always required of power: deliberation across plural values, accountability to those affected, a floor of universal rights, and the humility to take responsibility when judgment fails. Ethics did not emerge because human beings became virtuous. It emerged because societies learned that power without restraint destroys the communities on which power depends. Artificial intelligence does not change that lesson. It makes the lesson urgent.

The humanitarian sector is where the urgency is greatest and the test is hardest. There, consent collapses, experimentation finds the least protected subjects, accountability runs to people with no vote and no purchase, and the cost of failure is measured in lives. A technology that can be governed ethically there can be governed ethically anywhere. The frameworks now taking shape, from UNESCO's Recommendation to the UN's Scientific Panel and Global Dialogue, from the EU AI Act to the ICRC's precautionary policy, are the scaffolding of that governance. Whether they become an edifice depends on whether institutions, including universities, technology companies, and the collaborations between them, do the slow deliberative work this paper has tried to model.

Societies do not flourish because they possess increasingly powerful tools. They flourish because they develop increasingly trustworthy ways of governing them. That has been the purpose of ethics for four thousand years. It is the purpose of ethical AI now.

References

Sources

  1. 01Achiume, E. Tendayi. Racial Discrimination and Emerging Digital Technologies: A Human Rights Analysis. UN Doc. A/HRC/44/57, United Nations, June 2020.
  2. 02Acton, John Emerich Edward Dalberg. Letter to Mandell Creighton, 5 Apr. 1887. Historical Essays and Studies, Macmillan, 1907.
  3. 03Adams, Rachel. The New Empire of AI: The Future of Global Inequality. Polity, 2025.
  4. 04Allen, Gregory C. “Choking Off China's Access to the Future of AI.” Center for Strategic and International Studies, 11 Oct. 2022.
  5. 05Alston, Philip. Digital Welfare States and Human Rights: Report of the Special Rapporteur on Extreme Poverty and Human Rights. UN Doc. A/74/493, United Nations, 11 Oct. 2019.
  6. 06Amnesty International and Access Now. The Toronto Declaration: Protecting the Rights to Equality and Non-Discrimination in Machine Learning Systems. 16 May 2018.
  7. 07Andersin, Emelie. “The Use of the ‘Lavender’ in Gaza and the Law of Targeting.” Journal of International Humanitarian Legal Studies, vol. 16, no. 2, 2025, pp. 336 to 370.
  8. 08Appiah, Kwame Anthony. Cosmopolitanism: Ethics in a World of Strangers. W. W. Norton, 2006.
  9. 09Aristotle. Nicomachean Ethics. Translated by Terence Irwin, 2nd ed., Hackett, 1999.
  10. 10Beduschi, Ana. “Harnessing the Potential of Artificial Intelligence for Humanitarian Action.” International Review of the Red Cross, vol. 104, no. 919, 2022, pp. 1149 to 1169.
  11. 11Bengio, Yoshua, et al. International AI Safety Report 2026. AI Security Institute, Feb. 2026.
  12. 12Benjamin, Ruha. Race After Technology: Abolitionist Tools for the New Jim Code. Polity, 2019.
  13. 13Bietti, Elettra. “From Ethics Washing to Ethics Bashing.” Proceedings of the 2020 Conference on Fairness, Accountability, and Transparency, ACM, 2020, pp. 210 to 219.
  14. 14“The Bletchley Declaration by Countries Attending the AI Safety Summit, 1 and 2 November 2023.” GOV.UK, 1 Nov. 2023.
  15. 15Bradford, Anu. Digital Empires: The Global Battle to Regulate Technology. Oxford UP, 2023.
  16. 16Buolamwini, Joy, and Timnit Gebru. “Gender Shades: Intersectional Accuracy Disparities in Commercial Gender Classification.” Proceedings of Machine Learning Research, vol. 81, 2018, pp. 77 to 91.
  17. 17Carnegie Council for Ethics in International Affairs. Ethics in International Affairs 101. Carnegie Council.
  18. 18Centre for Humanitarian Data. Peer Review Framework for Predictive Analytics in Humanitarian Response. UN OCHA, Mar. 2020.
  19. 19The Code of Hammurabi. Translated by L. W. King, 1910. The Avalon Project, Yale Law School.
  20. 20Confucius. Analects. Translated by Edward Slingerland, Hackett, 2003.
  21. 21Coppi, Giulio, Rebeca Moreno Jimenez, and Sofia Kyriazi. “Explicability of Humanitarian AI: A Matter of Principles.” Journal of International Humanitarian Action, vol. 6, art. 19, 2021.
  22. 22Council of Europe. Framework Convention on Artificial Intelligence and Human Rights, Democracy and the Rule of Law. Council of Europe Treaty Series no. 225, 5 Sept. 2024.
  23. 23Crawford, Kate. Atlas of AI: Power, Politics, and the Planetary Costs of Artificial Intelligence. Yale UP, 2021.
  24. 24The Danish Institute for Human Rights. Guidance on Human Rights Impact Assessment of Digital Activities. DIHR, 2020.
  25. 25Devidal, Pierrick. “Lost in Digital Translation? The Humanitarian Principles in the Digital Age.” International Review of the Red Cross, vol. 106, no. 925, 2024.
  26. 26Diakopoulos, Nicholas, et al. “Principles for Accountable Algorithms and a Social Impact Statement for Algorithms.” FAT/ML.
  27. 27The Engine Room and Oxfam. Biometrics in the Humanitarian Sector. The Engine Room, 2018, updated 2023.
  28. 28European Parliament and Council of the European Union. Regulation (EU) 2024/1689 (Artificial Intelligence Act). Official Journal of the European Union, 12 July 2024.
  29. 29Fjeld, Jessica, et al. Principled Artificial Intelligence: Mapping Consensus in Ethical and Rights-Based Approaches to Principles for AI. Berkman Klein Center Research Publication no. 2020-1, Harvard University, 2020.
  30. 30Floridi, Luciano, and Josh Cowls. “A Unified Framework of Five Principles for AI in Society.” Harvard Data Science Review, vol. 1, no. 1, 2019.
  31. 31Friedman, Batya, and David G. Hendry. Value Sensitive Design: Shaping Technology with Moral Imagination. MIT Press, 2019.
  32. 32Future of Life Institute. “Asilomar AI Principles.” Future of Life Institute, 2017.
  33. 33Gebru, Timnit, et al. “Datasheets for Datasets.” Communications of the ACM, vol. 64, no. 12, 2021, pp. 86 to 92.
  34. 34“Global Push for AI Governance amid Warnings of ‘Catastrophic Harm.’” UN News, 5 July 2026.
  35. 35Heinzelman, Jessica, and Carol Waters. Crowdsourcing Crisis Information in Disaster-Affected Haiti. Special Report 252, United States Institute of Peace, 2010.
  36. 36High-level Committee on Programmes, UN System Chief Executives Board for Coordination. Principles for the Ethical Use of Artificial Intelligence in the United Nations System. United Nations, Sept. 2022.
  37. 37High-Level Expert Group on Artificial Intelligence. Ethics Guidelines for Trustworthy AI. European Commission, 8 Apr. 2019.
  38. 38Hobbes, Thomas. Leviathan. Edited by Richard Tuck, Cambridge UP, 1996.
  39. 39IEEE Global Initiative on Ethics of Autonomous and Intelligent Systems. Ethically Aligned Design. 1st ed., IEEE, 2019.
  40. 40Inter-Agency Standing Committee. IASC Operational Guidance on Data Responsibility in Humanitarian Action. IASC, 2023.
  41. 41International Committee of the Red Cross. Building a Responsible Humanitarian Approach: The ICRC's Policy on Artificial Intelligence. ICRC, Nov. 2024.
  42. 42International Committee of the Red Cross. “ICRC Position on Autonomous Weapon Systems.” ICRC, 12 May 2021.
  43. 43International Labour Organization. Profits and Poverty: The Economics of Forced Labour. 2nd ed., ILO, Mar. 2024.
  44. 44International Organization for Standardization and International Electrotechnical Commission. ISO/IEC 42001:2023, Information Technology, Artificial Intelligence, Management System. ISO, Dec. 2023.
  45. 45International Telecommunication Union. United Nations Activities on Artificial Intelligence. ITU, annual editions.
  46. 46Jacobsen, Katja Lindskov. “Experimentation in Humanitarian Locations: UNHCR and Biometric Registration of Afghan Refugees.” Security Dialogue, vol. 46, no. 2, 2015, pp. 144 to 164.
  47. 47Jobin, Anna, Marcello Ienca, and Effy Vayena. “The Global Landscape of AI Ethics Guidelines.” Nature Machine Intelligence, vol. 1, 2019, pp. 389 to 399.
  48. 48Johnson, Emily. “AI Ethics Frameworks: 10 Essential Resources to Build an Ethical AI Framework.” SecureITWorld, 12 Mar. 2025.
  49. 49Juskalian, Russ. “Inside the Jordan Refugee Camp That Runs on Blockchain.” MIT Technology Review, 12 Apr. 2018.
  50. 50Kant, Immanuel. “Toward Perpetual Peace.” Practical Philosophy, translated by Mary J. Gregor, Cambridge UP, 1996.
  51. 51Karenga, Maulana. Maat, the Moral Ideal in Ancient Egypt: A Study in Classical African Ethics. Routledge, 2004.
  52. 52Kuner, Christopher, and Massimo Marelli, editors. Handbook on Data Protection in Humanitarian Action. 2nd ed., ICRC and Brussels Privacy Hub, 2020.
  53. 53Laidlaw, James. The Subject of Virtue: An Anthropology of Ethics and Freedom. Cambridge UP, 2014.
  54. 54Latonero, Mark. “Stop Surveillance Humanitarianism.” The New York Times, 11 July 2019.
  55. 55Libertas Council. “Developing Guidelines on the Intersection of Human Trafficking and Technology.” Working group meeting, European Parliament, Brussels, 3 Dec. 2024.
  56. 56Madianou, Mirca. “Technocolonialism: Digital Innovation and Data Practices in the Humanitarian Response to Refugee Crises.” Social Media + Society, vol. 5, no. 3, 2019, pp. 1 to 13.
  57. 57Madianou, Mirca. Technocolonialism: When Technology for Good Is Harmful. Polity, 2024.
  58. 58Mantelero, Alessandro. Beyond Data: Human Rights, Ethical and Social Impact Assessment in AI. T.M.C. Asser Press, 2022.
  59. 59McGregor, Lorna, Daragh Murray, and Vivian Ng. “International Human Rights Law as a Framework for Algorithmic Accountability.” International and Comparative Law Quarterly, vol. 68, no. 2, 2019, pp. 309 to 343.
  60. 60Meier, Patrick. Digital Humanitarians: How Big Data Is Changing the Face of Humanitarian Response. CRC Press, 2015.
  61. 61Moor, James H. “What Is Computer Ethics?” Metaphilosophy, vol. 16, no. 4, 1985, pp. 266 to 275.
  62. 62Mitchell, Margaret, et al. “Model Cards for Model Reporting.” Proceedings of the Conference on Fairness, Accountability, and Transparency, ACM, 2019, pp. 220 to 229.
  63. 63Mittelstadt, Brent. “Principles Alone Cannot Guarantee Ethical AI.” Nature Machine Intelligence, vol. 1, 2019, pp. 501 to 507.
  64. 64Morgenthau, Hans J. Politics Among Nations: The Struggle for Power and Peace. Alfred A. Knopf, 1948.
  65. 65Munro, Robert. “Crowdsourcing and the Crisis-Affected Community: Lessons Learned and Looking Forward from Mission 4636.” Information Retrieval, vol. 16, no. 2, 2013, pp. 210 to 266.
  66. 66National Institute of Standards and Technology. Artificial Intelligence Risk Management Framework (AI RMF 1.0). NIST AI 100-1, U.S. Department of Commerce, Jan. 2023.
  67. 67Nearing, Grey, et al. “Global Prediction of Extreme Floods in Ungauged Watersheds.” Nature, vol. 627, 2024, pp. 559 to 563.
  68. 68Noble, Safiya Umoja. Algorithms of Oppression: How Search Engines Reinforce Racism. New York UP, 2018.
  69. 69North, Douglass C. Institutions, Institutional Change and Economic Performance. Cambridge UP, 1990.
  70. 70Nussbaum, Martha C. “Patriotism and Cosmopolitanism.” Boston Review, vol. 19, no. 5, 1994.
  71. 71NYU Ethical Tech CoLab. “Publications.” NYU School of Professional Studies Center for Global Affairs and Microsoft.
  72. 72Office of the United Nations High Commissioner for Human Rights. Guiding Principles on Business and Human Rights. United Nations, 2011.
  73. 73Office of the United Nations High Commissioner for Human Rights, B-Tech Project. Taxonomy of Human Rights Risks Connected to Generative AI. OHCHR, Nov. 2023.
  74. 74Organisation for Economic Co-operation and Development. Recommendation of the Council on Artificial Intelligence. OECD/LEGAL/0449, OECD, 2019, amended 2024.
  75. 75OSCE Office of the Special Representative and Co-ordinator for Combating Trafficking in Human Beings, and Tech Against Trafficking. Leveraging Innovation to Fight Trafficking in Human Beings. OSCE, 2020.
  76. 76Ostrom, Elinor. Governing the Commons: The Evolution of Institutions for Collective Action. Cambridge UP, 1990.
  77. 77Partnership on AI. “Our Tenets.” Partnership on AI, 2016.
  78. 78Palen, Leysia, and Kenneth M. Anderson. “Crisis Informatics: New Data for Extraordinary Times.” Science, vol. 353, no. 6296, 2016, pp. 224 to 225.
  79. 79Pizzi, Michael, Mila Romanoff, and Tim Engelhardt. “AI for Humanitarian Action: Human Rights and Ethics.” International Review of the Red Cross, vol. 102, no. 913, 2020, pp. 145 to 180.
  80. 80Raji, Inioluwa Deborah, et al. “Closing the AI Accountability Gap: Defining an End-to-End Framework for Internal Algorithmic Auditing.” Proceedings of the 2020 Conference on Fairness, Accountability, and Transparency, ACM, 2020, pp. 33 to 44.
  81. 81Russell, Bertrand. Power: A New Social Analysis. George Allen and Unwin, 1938.
  82. 82Risse, Mathias. “Human Rights and Artificial Intelligence: An Urgently Needed Agenda.” Human Rights Quarterly, vol. 41, no. 1, 2019, pp. 1 to 16.
  83. 83Roberts, Huw, et al. “Global AI Governance: Barriers and Pathways Forward.” International Affairs, vol. 100, no. 3, 2024, pp. 1275 to 1286.
  84. 84Sandvik, Kristin Bergtora, Katja Lindskov Jacobsen, and Sean Martin McDonald. “Do No Harm: A Taxonomy of the Challenges of Humanitarian Experimentation.” International Review of the Red Cross, vol. 99, no. 904, 2017.
  85. 85Sastry, Girish, et al. “Computing Power and the Governance of Artificial Intelligence.” arXiv:2402.08797, Feb. 2024.
  86. 86Scharre, Paul. Four Battlegrounds: Power in the Age of Artificial Intelligence. W. W. Norton, 2023.
  87. 87Signpost AI. “Pilot Report: Signpost AI Information Assistant.” International Rescue Committee, 27 June 2025.
  88. 88Singer, Peter. The Expanding Circle: Ethics, Evolution, and Moral Progress. Princeton UP, 2011.
  89. 89Smuha, Nathalie A. “Beyond a Human Rights-Based Approach to AI Governance: Promise, Pitfalls, Plea.” Philosophy and Technology, vol. 34, no. 1, 2021, pp. 91 to 104.
  90. 90Spencer, Sarah W. Humanitarian AI: The Hype, the Hope and the Future. Network Paper 85, Humanitarian Practice Network, ODI, Nov. 2021.
  91. 91Spencer, Sarah W. Humanitarian AI Revisited: Seizing the Potential and Sidestepping the Pitfalls. Network Paper 89, Humanitarian Practice Network, ODI, May 2024.
  92. 92“Statement on Inclusive and Sustainable Artificial Intelligence for People and the Planet.” AI Action Summit, Paris, 11 Feb. 2025.
  93. 93Stilgoe, Jack, Richard Owen, and Phillip Macnaghten. “Developing a Framework for Responsible Innovation.” Research Policy, vol. 42, no. 9, 2013, pp. 1568 to 1580.
  94. 94Tomasello, Michael. A Natural History of Human Morality. Harvard UP, 2016.
  95. 95UNESCO. Recommendation on the Ethics of Artificial Intelligence. UNESCO, 23 Nov. 2021.
  96. 96UNHCR. “Our Approach to Artificial Intelligence.” UNHCR Digital Transformation Strategy 2022 to 2026, UNHCR.
  97. 97UNICEF. Policy Guidance on AI for Children 2.0. UNICEF Office of Global Insight and Policy, Nov. 2021.
  98. 98United Nations. Global Digital Compact. Annex to Resolution A/RES/79/1, United Nations, 22 Sept. 2024.
  99. 99United Nations Development Programme. AI Hub for Sustainable Development. UNDP, 2024.
  100. 100United Nations General Assembly. Enhancing International Cooperation on Capacity-Building of Artificial Intelligence. Resolution A/RES/78/311, 1 July 2024.
  101. 101United Nations General Assembly. International Covenant on Civil and Political Rights. United Nations, 16 Dec. 1966.
  102. 102United Nations General Assembly. International Covenant on Economic, Social and Cultural Rights. United Nations, 16 Dec. 1966.
  103. 103United Nations General Assembly. Resolution A/RES/79/325, Terms of Reference and Modalities for the Independent International Scientific Panel on Artificial Intelligence and the Global Dialogue on Artificial Intelligence Governance. 26 Aug. 2025.
  104. 104United Nations General Assembly. Seizing the Opportunities of Safe, Secure and Trustworthy Artificial Intelligence Systems for Sustainable Development. Resolution A/RES/78/265, 21 Mar. 2024.
  105. 105United Nations General Assembly. Universal Declaration of Human Rights. Resolution 217 A (III), 10 Dec. 1948.
  106. 106United Nations High Commissioner for Human Rights. The Right to Privacy in the Digital Age. UN Doc. A/HRC/48/31, United Nations, Sept. 2021.
  107. 107United Nations High-level Advisory Body on Artificial Intelligence. Governing AI for Humanity: Final Report. United Nations, Sept. 2024.
  108. 108United Nations Secretary-General. Our Common Agenda Policy Brief 11: UN 2.0. United Nations, Sept. 2023.
  109. 109United Nations Secretary-General. Roadmap for Digital Cooperation. UN Doc. A/74/821, United Nations, June 2020.
  110. 110UN Global Pulse. Big Data for Development: Challenges and Opportunities. United Nations, May 2012.
  111. 111UN OCHA. Briefing Note on Artificial Intelligence and the Humanitarian Sector. OCHA, 2024.
  112. 112UN OCHA. OCHA Data Responsibility Guidelines. Centre for Humanitarian Data, Oct. 2021.
  113. 113UN Secretary-General's High-level Panel on Digital Cooperation. The Age of Digital Interdependence. United Nations, June 2019.
  114. 114Université de Montréal. Montreal Declaration for a Responsible Development of Artificial Intelligence. 2018.
  115. 115Veale, Michael, and Frederik Zuiderveen Borgesius. “Demystifying the Draft EU Artificial Intelligence Act.” Computer Law Review International, vol. 22, no. 4, 2021, pp. 97 to 112.
  116. 116Weber, Max. Economy and Society: An Outline of Interpretive Sociology. Edited by Guenther Roth and Claus Wittich, U of California P, 1978.
  117. 117World Bank. “Famine Action Mechanism (FAM).” World Bank, 2018.
  118. 118World Food Programme. “WFP Releases HungerMap LIVE.” WFP, 2019.
  119. 119World Food Programme Innovation Accelerator. “New Global Principles for Innovative and Ethical AI.” WFP, 2022.
  120. 120Whittaker, Meredith, et al. AI Now Report 2018. AI Now Institute, New York University, Dec. 2018.
  121. 121Winner, Langdon. “Do Artifacts Have Politics?” Daedalus, vol. 109, no. 1, 1980, pp. 121 to 136.
  122. 122Yeung, Karen, Andrew Howes, and Ganna Pogrebna. “AI Governance by Human Rights-Centred Design, Deliberation and Oversight: An End to Ethics Washing.” The Oxford Handbook of Ethics of AI, Oxford UP, 2020, pp. 77 to 106.

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution. External programs cited are referenced as evidence, not as CoLab partnerships.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file +

Publications · Academic paper

What Is Ethical AI?

Ethics, Ethical Technology, and Ethical International Relations for the Age of Intelligent Machines

Ethical Tech CoLab

NYU Center for Global Affairs · in collaboration with Microsoft

July 2026

Ethical Tech CoLab · NYU School of Professional Studies Center for Global Affairs, in collaboration with Microsoft. Master's Research · M.S. in Global Affairs.

What is ethics, and why did humanity invent it? What makes technology ethical, international relations ethical, and artificial intelligence ethical? This paper answers those questions in that order, on the conviction that the issue at the heart of artificial intelligence is power and that ethics is humanity's oldest and most tested answer to the problem of governing power. It traces ethics from its origins in the earliest civilizations through international affairs and human rights law to the responsible AI movement, the humanitarian sector, and the United Nations system, and closes with the motivation and research philosophy of the Ethical Tech CoLab.

Key figures

84

AI ethics guidelines mapped by Jobin, Ienca, and Vayena, converging on five principles but diverging on how to apply them

118

countries, overwhelmingly in the Global South, party to none of the major AI governance initiatives the UN surveyed

15%

of responses in the IRC's Signpost AI pilot could not be rated safe, the evaluators' basis for concluding the tool is not safe without a human in the loop

$236B

in annual illegal profits from forced labour, the ILO estimate that frames the CoLab's Forced Labor Structural Risk Index

01

Executive Summary

Humanity has repeatedly transformed itself through technological innovation. Fire, agriculture, metallurgy, writing, navigation, industrialization, electricity, nuclear energy, biotechnology, and digital computing each expanded human capability while simultaneously increasing humanity's capacity to cause harm. Every technological revolution therefore forced societies to develop new ethical frameworks capable of governing newly acquired power. Artificial intelligence represents the latest and perhaps most consequential chapter in this historical process. The ethical questions confronting AI are not unprecedented, because they concern the same enduring challenge that has shaped civilizations for millennia: how should power be exercised responsibly? The issue is power. Ethics is the answer.

This paper deliberately resists the convention, common in the AI ethics literature, of opening with principles such as fairness, transparency, and accountability. Principles detached from their foundations are easily reduced to slogans, and slogans are easily captured by the institutions they are meant to constrain. The paper therefore begins with prior questions. Section 2 frames the oldest of them, humanity's relationship with power, the paradox by which every expansion of capability has expanded the capacity for harm. Sections 3 through 5 ask what ethics is, why human beings created it, and how it evolved from local custom into universal human rights. Section 6 asks what ethical international relations means, drawing on the Carnegie Council's framework of ethical realism and structured ethical deliberation. Section 7 asks what makes technology ethical, and Section 8 arrives, on that foundation, at the question that titles the paper: what is ethical AI?

The answer defended throughout is institutional rather than technical. Ethical AI is not a product feature. It is artificial intelligence whose entire lifecycle remains accountable to ethical deliberation, to human rights as a floor, to the do no harm obligation, and to the inclusive participation of those the technology affects. No system is ethical in itself; systems are made ethical, or fail to be, by the institutions that govern them.

The second half of the paper tests that answer where the stakes are highest. Section 9 traces the rise of the responsible AI movement through its three acts of principles, reckoning, and operationalization. Sections 10 and 11 turn to humanitarian action, tracing the evolution of AI in humanitarian work from the crisis mapping of 2008 to the generative systems of 2026, and showing why humanitarian AI is fundamentally different from commercial AI: consent collapses, experimentation finds the least protected subjects, accountability runs to people with no vote and no purchase, and the cost of failure is measured in lives. Sections 12 through 14 supply the governance context: the geopolitics of AI regulation, the human rights foundations that predate the technology, and the United Nations system's evolving approach, which in 2026 produced the first report of the Independent International Scientific Panel on AI and the first Global Dialogue on AI Governance in Geneva.

Section 15 presents the Ethical Tech CoLab, whose portfolio, from civilian evacuation models to a forced labour risk index, puts the paper's argument into practice, and Section 16 concludes. The paper is written for policymakers, humanitarian practitioners, donors, and researchers rather than AI engineers, in the conviction that the disciplines it draws together, philosophy, ethics, international affairs, humanitarian ethics, digital ethics, and responsible innovation, are not six literatures but one conversation. Societies do not flourish because they possess increasingly powerful tools. They flourish because they develop increasingly trustworthy ways of governing them.

02

Humanity's Relationship with Power

The paradox of capability. The history of civilization is not merely the history of technological innovation. It is the history of humanity learning to govern increasingly powerful forms of power. Human beings possess a remarkable capacity to create tools that extend their abilities beyond natural biological limits. Fire transformed survival. Agriculture transformed settlement. Writing transformed memory. Navigation transformed geography. Industrial machinery transformed production. Nuclear physics transformed warfare and medicine. Artificial intelligence now promises to transform decision making itself. Each of these revolutions increased humanity's ability to reshape the world, and every expansion of capability simultaneously increased humanity's capacity to cause harm.

Power itself is the oldest subject of social thought. Max Weber defined it as the probability that an actor within a social relationship will be in a position to carry out his own will despite resistance, a definition that fits an algorithm allocating aid as readily as a sovereign commanding subjects. Bertrand Russell went further, arguing that power is the fundamental concept in social science, in the same sense in which energy is the fundamental concept in physics. And Lord Acton supplied the warning that every constitutional tradition has since repeated: power tends to corrupt, and absolute power corrupts absolutely. Read together, these three claims frame the problem this paper addresses. Power is universal, power is the currency of social life, and power unrestrained degrades the one who holds it and endangers everyone subject to it.

History therefore reveals a recurring pattern. Technological innovation alone has never guaranteed human progress. Every major expansion of human capability has required a corresponding expansion of humanity's ethical, legal, and political institutions. Writing required legal systems. Commerce required contracts. Industrialization generated unprecedented prosperity while requiring labor protections, environmental regulation, and public health institutions. Nuclear physics transformed medicine while necessitating international non proliferation regimes. Without such institutions, technological power becomes detached from moral responsibility, and societies confront crises of legitimacy, injustice, and instability.

Artificial intelligence is the latest expression of this enduring dynamic. Its ethical significance lies not in its novelty but in the type of power it creates. AI increasingly participates in decisions concerning healthcare, education, employment, humanitarian assistance, migration, policing, financial services, and national security, and in doing so it redistributes authority between human institutions and computational systems. The fundamental question is therefore not whether AI is intelligent, but whether societies possess institutions capable of governing this new concentration of power responsibly. Ethics is not an external constraint imposed on innovation after the fact. It is the normative architecture through which societies determine how newly acquired forms of power should be exercised.

03

What Is Ethics?

Few concepts have shaped human civilization more profoundly than ethics, yet few are as frequently invoked without careful definition. Governments invoke ethics when justifying interventions or humanitarian assistance; corporations establish ethics committees; international organizations describe ethical obligations toward refugees and future generations. Despite its ubiquity, the term is often reduced to a synonym for compliance, legality, or professional conduct.

At its most fundamental level, ethics is the systematic study of how human beings ought to act, what values should guide individual and collective behavior, and how competing moral claims should be evaluated. The Carnegie Council for Ethics in International Affairs puts the point plainly: ethics is about how things ought to be, not simply how things are or are likely to be, and it forces us to look beyond our own immediate interests to consider the interests of others. Its president Joel Rosenthal describes ethics as making the effort to evaluate competing points of view and then truly caring about the impact of the choices you have made, adding that ethics is not some cure-all for the world's problems, but it is an actual process for finding solutions. Ethics, on this understanding, is not a fixed doctrine but a disciplined practice of judgment.

Ethics and morality. The word originates from the ancient Greek ethos, referring to character, custom, and habitual ways of living. Aristotle used the concept to describe the cultivation of virtues that enable individuals to flourish within a political community. Scholarship often distinguishes ethics from morality: morality generally refers to the beliefs, norms, and practices that characterize a particular community, while ethics represents the critical examination of those beliefs. As the anthropologist James Laidlaw observes, the study of ethics concerns not only the moral rules a society enforces but the ways people consciously deliberate about how they ought to live. This is why ethical inquiry frequently challenges existing arrangements: slavery, colonial domination, racial segregation, and child labor were all questioned and ultimately rejected through ethical reasoning that appealed to principles beyond inherited custom.

Ethics, law, and power. Ethics should not be confused with law. Apartheid in South Africa and the Nuremberg Laws in Nazi Germany were legally enforceable while violating principles now recognized as universal human rights. Conversely, many ethical obligations exceed legal requirements. Ethical questions emerge wherever human beings exercise power over one another, operating at several levels at once: the decisions of individuals, the conduct of organizations, and the behavior of states and international institutions. Technological capability alone offers no guidance as to whether a given application advances justice, respects human dignity, or contributes to the common good. Ethics is the bridge between human capability and human responsibility.

04

Why Humanity Created Ethics

One persistent misconception is the belief that moral philosophy began as an intellectual exercise conducted by ancient philosophers seeking abstract truths. While philosophers transformed ethics into a systematic field of inquiry, ethical behavior itself predates philosophy by thousands of years. Ethics emerged not because human beings suddenly became virtuous, but because societies discovered that power without restraint ultimately destroys the communities upon which power itself depends. Michael Tomasello's research on the natural history of human morality argues that distinctively human moral psychology, including the sense of fairness, obligation, and shared commitment, arose from the demands of interdependent cooperation. Ethical norms reduced uncertainty in human interaction long before economists developed the vocabulary of transaction costs.

The earliest ethical systems. The earliest documented normative systems make the connection between ethics and the governance of power explicit. In ancient Egypt, the principle of Maat, embodying truth, justice, balance, and reciprocity, served as the foundation upon which political authority derived its legitimacy. In Mesopotamia, the Code of Hammurabi justified its standards in terms that remain strikingly modern: the laws existed, its prologue declares, so that the strong might not injure the weak. The transition from unwritten custom to publicly accessible legal standards anticipated a principle central to contemporary technology governance. Citizens cannot comply with secret laws, and individuals cannot meaningfully contest algorithmic decisions whose logic remains opaque. The modern demand for explainable artificial intelligence reflects a very old ethical commitment to public accountability. In China, Confucius articulated an ethics of cultivated relationships centered on ren, or humaneness, in which those who hold power bear heightened responsibilities.

Political philosophy generalized these insights. Thomas Hobbes argued that without commonly accepted rules, human societies risk descending into perpetual conflict. Locke, Rousseau, and Kant increasingly linked political legitimacy to the protection of rights rather than to coercive power alone. Modern institutional economics reaches a similar conclusion: Douglass North demonstrated that long term prosperity depends upon institutions capable of generating trust and credible commitments, and Elinor Ostrom showed empirically that communities can govern shared resources sustainably when they develop norms of reciprocity, monitoring, and accountability. Ethics, on this view, functions as institutional infrastructure.

The humanitarian sector illustrates the relationship with particular clarity. Humanitarian organizations frequently operate where formal state institutions have weakened or collapsed, and the principles of humanity, impartiality, neutrality, and independence generate the trust on which humanitarian access depends. Artificial intelligence introduces a comparable challenge. Refugees cannot audit the machine learning models that help determine assistance eligibility, and displaced families cannot meaningfully contest automated decisions influencing access to shelter or protection. As throughout history, trust cannot be engineered; it must be earned by legitimate institutions operating under shared ethical principles.

05

From Custom to Universal Ethics

The history of ethics is, in many respects, the history of an expanding moral community. Early societies focused their strongest obligations on those with whom individuals shared kinship, language, territory, or faith. As societies became larger and more interconnected, ethical thought expanded with them. Peter Singer describes this trajectory as the expanding circle of moral concern, in which the reach of ethical consideration widens from family to tribe, from tribe to nation, and from nation toward humanity as a whole. Kwame Anthony Appiah frames the same process as an ongoing cosmopolitan conversation across difference, and Martha Nussbaum argues that ethical reasoning challenges individuals to regard themselves as citizens of a wider human community whose dignity deserves equal consideration.

The catastrophes of the twentieth century converted this philosophical trajectory into institutional fact. The Holocaust demonstrated that legality and prevailing social morality could become instruments of profound injustice when detached from broader commitments to human dignity. The adoption of the Universal Declaration of Human Rights in 1948 marked the decisive turn: its first article declares that all human beings are born free and equal in dignity and rights, establishing dignity rather than citizenship as the normative foundation of international order.

Universalism without uniformity. Universal ethics does not imply ethical uniformity. Core principles such as protection from torture and recognition of equal dignity apply everywhere, while their practical expression legitimately varies across societies. UNESCO's Recommendation on the Ethics of Artificial Intelligence adopts precisely this position, grounding AI governance in universal human rights commitments while recognizing the importance of cultural diversity, linguistic plurality, and local participation.

Artificial intelligence intensifies this old tension in a new form. Algorithmic systems are typically developed within particular social, linguistic, and economic contexts before being deployed globally, and they may encode the assumptions of the societies that built them. Research on algorithmic fairness has demonstrated that models trained primarily on data from wealthy, highly connected populations frequently perform worse for marginalized communities and underrepresented languages. A technically sophisticated system that systematically disadvantages displaced populations or excludes minority languages may satisfy engineering benchmarks while failing the standards on which humanitarian action itself is founded.

06

What Is Ethical International Relations?

International affairs pose the hardest case for ethics, because they are the arena in which power is least constrained. Two extreme views have long competed. Realism, in its strongest version, holds that states pursue power and narrowly defined self interest, and that ethics has little to contribute; Henry Kissinger captured its spirit in the warning that a country that demands moral perfection in its foreign policy will achieve neither perfection nor security. Idealism holds that what matters is adherence to principles and international law, in the lineage of Kant's proposal for a federation of states in Perpetual Peace. Each view captures something real, and each is insufficient on its own.

Ethical realism. The Carnegie Council's answer, and the orientation this paper adopts, is ethical realism: a view of morality in global politics that takes seriously both the role of ethics and the nature of power. Hans Morgenthau, the founder of modern realism, framed the choice not between moral principles and the national interest devoid of moral dignity, but between one set of moral principles divorced from political reality and another set of moral principles derived from political reality. Ethical realism refuses both the cynicism that treats ethics as decoration and the utopianism that ignores power. It asks what the most ethical policy is that is also possible.

Ethical realism operates through a structured process of ethical deliberation: identifying goals, finding means to satisfy them, and assessing and revising both against standards, values, principles, duties, and rights, then running the cycle from problem through choice and action to reflection and the taking of responsibility. Two features deserve emphasis. First, deliberation is pluralistic. The main Western frameworks, consequentialism, virtue ethics, and rights or duty based ethics, each illuminate a different aspect of a decision, and solving real problems may require drawing on each without endorsing any single one as uniquely correct. Second, deliberation is inclusive. It is best done with others, especially those most affected, because on our own we cannot see the full picture and we tend to be biased toward decisions that benefit us. These two commitments, pluralism and inclusion, are the methodological core of what the CoLab means by ethical technology.

07

What Is Ethical Tech?

Before the question of ethical AI can be answered, a broader one deserves a direct answer: what makes technology ethical? The founding insight of the scholarly tradition is that technologies are not neutral instruments whose moral character depends entirely on their users. Langdon Winner's classic essay asked whether artifacts have politics and answered that they do: the design of a bridge, a machine, or a system can embody and enforce social arrangements, settling questions of power before any user makes a choice. James Moor, founding the field of computer ethics, observed that computing creates policy vacuums, situations in which technology confers new capabilities faster than societies can form rules for their use, and that filling those vacuums is a conceptual and ethical task, not a technical one. Technology is, in this sense, frozen judgment.

Value sensitive design and responsible innovation. Two research programs give ethical technology operational content. Value sensitive design, developed by Batya Friedman and colleagues, provides a method for identifying the stakeholders of a system, direct and indirect, surfacing the values at stake, and building those values into the design rather than auditing for them afterward. The responsible research and innovation literature, associated with Jack Stilgoe, Richard Owen, and Phillip Macnaghten, adds four commitments any innovating institution can be held to: anticipation of plausible harms, reflexivity about the innovator's own assumptions, inclusion of affected publics, and responsiveness, the demonstrated capacity to change course when evidence demands it. These frameworks establish that the ethics of a technology is decided in its governance, upstream in design and downstream in oversight, and not in its marketing.

This is the sense in which the CoLab uses the term ethical tech. It does not name a category of inherently benign products. It names a practice: choosing problems by the harm they address rather than the market they open, designing with the affected rather than for them, embedding safeguards as constraints rather than aspirations, and accepting accountability structures with the power to stop a build. The definition of ethical AI developed next is a special case of this practice.

08

What Is Ethical AI?

A definition. With these foundations in place, the question that titles this paper can be answered without slogans. Ethical AI is not a product feature, a compliance checkbox, or a marketing claim. It is artificial intelligence whose entire lifecycle, from problem selection and data collection through design, deployment, and retirement, remains accountable to ethical deliberation: to human rights as a floor, to the humanitarian principle of doing no harm, and to the inclusive participation of those the technology affects. On this definition, no system is ethical in itself. Systems are made ethical, or fail to be, by the institutions that govern them.

The field's own history supports this institutional definition. When Anna Jobin, Marcello Ienca, and Effy Vayena mapped eighty four AI ethics guidelines published worldwide, they found an apparent global convergence around five principles, transparency, justice and fairness, non maleficence, responsibility, and privacy, but substantive divergence over how those principles were interpreted, why they mattered, to whom they applied, and how they should be implemented. They also documented the near absence of the Global South from the production of the guidelines themselves. Luciano Floridi and Josh Cowls proposed that the recurring principles map closely onto the classical principles of bioethics, beneficence, non maleficence, autonomy, and justice, with one addition specific to AI: explicability, the requirement that systems be intelligible and that someone be answerable for them.

The limits of principles became apparent almost immediately. Brent Mittelstadt's critique observed that AI development lacks precisely what makes principlism work in medicine: common aims and fiduciary duties, professional norms, proven methods to translate principles into practice, and robust accountability mechanisms. A principle that no profession is bound to, no method operationalizes, and no institution enforces is an aspiration, not a governance instrument. This is why the intergovernmental instruments matter. The OECD AI Principles, the IEEE's Ethically Aligned Design, and UNESCO's Recommendation do not invent new moral values for the digital age. They translate enduring ethical principles into practical guidance, which is exactly what ethics has done at every previous technological transition.

Accuracy is not ethics. One further clarification completes the definition. Ethical AI is frequently conflated with accurate AI, as though sufficient technical performance would resolve the ethical questions. It would not. A perfectly accurate system pursuing an unethical objective remains unethical, and efficiency without legitimacy may strengthen injustice rather than reduce it. The question ethics asks of an AI system is the one the deliberative framework asks of any policy: not only whether the means work, but whether the ends are justified, and whether the relationship between ends and means respects the dignity of those affected.

09

The Rise of Responsible AI

The movement now called responsible AI emerged in three acts, and each act refined the answer to what ethical AI requires. The first act was the age of principles. Between roughly 2016 and 2019, ethics codes issued from every quarter: the Asilomar AI Principles from the research community, which declared that the goal of AI research should be to create not undirected intelligence but beneficial intelligence; the Montreal Declaration, distinctive for its methodology of citizen deliberation; the IEEE's engineering framework; and the OECD Principles. The proliferation itself was evidence of sincerity, but also of fragmentation.

The first act left behind a canon. Ten resources remain the standard references for any institution building an AI ethics framework, each contributing something distinct:

  • IEEE Ethically Aligned Design, the engineering profession's handbook, which seeded the P7000 series of technical standards.
  • The EU Ethics Guidelines for Trustworthy AI, which defined trustworthy AI as lawful, ethical, and robust and set out seven requirements that shaped the EU AI Act.
  • The Montreal Declaration, whose ten principles emerged from an extended public deliberation with citizens.
  • The Asilomar AI Principles, twenty three principles framing the goal of the field as beneficial rather than merely capable intelligence.
  • The Partnership on AI Tenets, committing technology companies and civil society to safety, fairness, transparency, and accountability.
  • The OECD AI Principles, the first intergovernmental AI standard and the reference point for national policy.
  • The Toronto Declaration, the human rights community's entry, applying equality and non discrimination law to machine learning.
  • The AI Now Institute's research program, which pressed the field on accountability, labor, and bias, and popularized algorithmic impact assessments.
  • The FAT/ML Principles for Accountable Algorithms, five commitments addressed to the developers who build the systems.
  • The Berkman Klein Center's Principled Artificial Intelligence, mapping thirty six prominent documents into eight recurring themes.

Read as a set, the ten resources trace the same arc as this paper: from professional self regulation through public deliberation and intergovernmental standards to human rights law, and from principles addressed to institutions toward instruments usable by the practitioners who build and the communities who bear the consequences.

The reckoning. The second act was the reckoning, and its evidence was empirical. Joy Buolamwini and Timnit Gebru's Gender Shades study demonstrated that commercial facial analysis systems misclassified darker skinned women at error rates approaching 34.7 percent while the maximum error rate for lighter skinned men was 0.8 percent, a single result that legitimized the independent algorithmic audit as a discipline. Safiya Umoja Noble showed how search algorithms reinforce racism through technological redlining. Ruha Benjamin diagnosed the New Jim Code, discriminatory design that encodes inequity while appearing benevolent. Kate Crawford's Atlas of AI reframed the technology as an extractive industry whose costs in minerals, labor, and data fall far from its beneficiaries. And Elettra Bietti analyzed ethics washing, the instrumental use of ethics language by industry to forestall binding regulation.

The third act was operationalization. The critique produced tools: model cards for standardized reporting, datasheets for datasets, and end to end internal audit frameworks adapted from aerospace and finance. It then produced institutions. The United States National Institute of Standards and Technology released its AI Risk Management Framework in January 2023, organizing practice into four functions, govern, map, measure, and manage. ISO and IEC published ISO/IEC 42001 in December 2023, the first certifiable AI management system standard. From late 2023 the center of gravity shifted from fairness in deployed systems to safety in frontier models: the Bletchley Declaration, signed by twenty eight countries and the European Union, acknowledged the potential for catastrophic harm, and the International AI Safety Report chaired by Yoshua Bengio became the scientific reference of the safety era. The humanitarian reading of this decade is the one on which its sharpest critics converge: principles without accountability structures protect institutions, not people.

10

The Historical Evolution of AI in Humanitarian Work

The story of AI in humanitarian action does not begin with algorithms but with people. In the aftermath of Kenya's disputed 2008 election, a small group of technologists built Ushahidi, a platform for crowdsourcing reports of violence by text message and web. When a magnitude 7.0 earthquake devastated Haiti in 2010, volunteers around the world adapted the same model to map urgent needs in near real time, while Mission 4636 mobilized thousands of Kreyol speaking diaspora volunteers to translate and triage tens of thousands of text messages from survivors. Patrick Meier chronicled this era of digital humanitarians. Crucially, these early systems were hybrids of human and machine intelligence: the human computation of volunteer translators became the training substrate for early natural language processing tools in crisis response. The artificial intelligence of the humanitarian sector was built on the labor and knowledge of affected communities from the start.

Institutionalization and prediction. In 2009 the UN Secretary General launched Global Pulse, premised on the idea that the digital traces of mobile phone use could provide real time insight into human wellbeing. By the mid 2010s, operational machine learning had arrived. The World Food Programme's HungerMap LIVE began nowcasting food insecurity across dozens of countries from mobile surveys, satellite imagery, and market data. The World Bank's Famine Action Mechanism linked machine learning famine prediction to pre arranged financing. OCHA's Centre for Humanitarian Data built a peer review framework for predictive analytics that allowed the UN's Central Emergency Response Fund to release money before forecasted floods struck Bangladesh in 2020.

Identification and biometrics. Alongside prediction came identification. UNHCR had experimented with iris scan registration of Afghan returnees as early as 2002, and by the late 2010s biometric registration had become standard infrastructure across UNHCR and WFP operations. In Jordan's Azraq and Zaatari camps, WFP's Building Blocks system coupled iris scan authentication to a permissioned blockchain, allowing refugees to buy groceries with a glance. These systems, celebrated by their builders as efficient and fraud resistant, would become the focal point of the sector's fiercest ethical controversies.

The generative era arrived abruptly. The most rigorous published evaluation to date, the Signpost AI pilot led by the International Rescue Committee across four countries, tested a large language model information assistant for refugees and migrants: response quality improved and moderator handling time for complex queries fell from as much as twenty five minutes to under five, but roughly fifteen percent of responses could not be rated safe, leading the evaluators to conclude that the tool is not safe without a human in the loop. Three features of this history deserve emphasis. First, the arc runs from participation toward prediction and automation, progressively distancing affected people from the systems that decide about them. Second, each wave arrived with private sector partners. Third, governance consistently lagged deployment.

11

Why Humanitarian AI Is Fundamentally Different

It is tempting to treat humanitarian AI as commercial AI with a nobler mission statement. That view is mistaken, and the difference is structural rather than rhetorical. It begins with the humanitarian principles of humanity, neutrality, impartiality, and independence, which are not aspirational values but operational commitments on which access to conflict zones, the trust of affected communities, and the safety of aid workers depend. As the ICRC's Pierrick Devidal argues, the reflex that if technology can help it should be used risks quietly subordinating those principles to innovation logics imported from the commercial world.

Consent and power. The deepest difference concerns consent and power. Commercial AI at least nominally rests on user agreement. In humanitarian settings even that fiction collapses. A person who must submit to iris scanning to feed her children has not consented in any meaningful sense; she has complied. Mark Latonero named this dynamic surveillance humanitarianism, the enormous data collection systems deployed by aid organizations that inadvertently increase the vulnerability of people in urgent need, writing in the wake of the 2019 standoff in Yemen, where WFP suspended aid amid a dispute over biometric registration. The ICRC's own Handbook on Data Protection in Humanitarian Action concedes that consent is frequently an unusable legal basis in displacement contexts.

Humanitarian experimentation. A second structural difference is the asymmetry of experimentation. Kristin Bergtora Sandvik, Katja Lindskov Jacobsen, and Sean Martin McDonald have developed a taxonomy of humanitarian experimentation, showing how untested technologies are routinely piloted on crisis affected populations without the ethics infrastructure that governs experimentation elsewhere: there is a stark ethical and practical difference between managing risk and introducing it. Mirca Madianou presses the critique further with the concept of technocolonialism, arguing that digital innovation and data practices in aid rework colonial relationships of extraction, with affected people's data generating value for agencies, states, and technology firms while the risks remain with the data subjects.

Third, the accountability geometry is inverted. A commercial firm answers, however imperfectly, to customers, shareholders, and regulators. A humanitarian organization's primary obligation runs to affected populations who are neither customers nor voters, who often cannot see the algorithm that ranked their vulnerability, and who typically have no avenue of redress. This is why the sector's own standards center the do no harm obligation rather than product liability, and why scholars conclude that human rights law, not voluntary ethics, must govern humanitarian AI across its lifecycle. Finally, the cost of failure is categorically different. A mistargeted advertisement wastes attention; a mistargeted famine forecast, a false fraud flag on a food ration, or a leaked refugee database can cost lives. The ICRC's 2024 institutional AI policy codifies the resulting posture: a precautionary approach that permits deployment only with evidence of positive impact, mandates human oversight, and preserves non digital alternatives.

12

The International Affairs Context

Humanitarian AI does not unfold in a geopolitical vacuum. The technologies aid agencies adopt, the rules that constrain them, and the funding that sustains them are all shaped by a contest among what Anu Bradford calls the three digital empires: the American market driven model, the Chinese state driven model, and the European rights driven model. The European Union struck first: the EU Artificial Intelligence Act, in force since August 2024, establishes the world's first horizontal, risk based AI regime, prohibiting practices such as social scoring outright and designating migration and asylum management as high risk, a category of direct humanitarian relevance. The Council of Europe's Framework Convention on Artificial Intelligence, the first legally binding international AI treaty, extends the rights based model into international law.

The geopolitics of competition. Against this regulatory current runs the geopolitics of competition. The United States' October 2022 semiconductor export controls marked, in Gregory Allen's phrase, a strategy of choking off China's access to the future of AI. Scholars of compute governance argue that computing power has become the most governable input to AI, detectable, excludable, and concentrated in a narrow supply chain, making chips the decisive lever of AI policy. Paul Scharre frames the resulting contest as a struggle across four battlegrounds: data, compute, talent, and institutions.

The multilateral response has been real but fragile. The Bletchley Declaration of November 2023 achieved a joint acknowledgment by the United States, China, the European Union, and twenty six other states that frontier AI carries the potential for catastrophic harm. By the Paris AI Action Summit of February 2025, the consensus had visibly fractured: the United States and the United Kingdom declined to sign the summit statement, and the framing shifted from safety to action and innovation. The UN's High level Advisory Body quantified the exclusion built into this landscape: only seven countries were parties to all of the major AI governance initiatives it surveyed, while 118 countries, overwhelmingly in the Global South, were parties to none.

War has sharpened every one of these debates. The ICRC has called on states to adopt new legally binding rules on autonomous weapon systems. Reports that an AI decision support system known as Lavender was used to generate targeting recommendations in Gaza have produced the first sustained legal scholarship on AI assisted targeting under international humanitarian law. For the humanitarian sector the conclusion is sobering. The rules governing AI are being written by and for the powerful, in fora where affected populations, and often the states that host them, have no seat. Humanitarian AI governance must therefore be understood as, among other things, a project of representation: an attempt to force the interests of the excluded into rooms where they are otherwise absent.

13

Human Rights Foundations

If the geopolitics of AI is a story of fragmentation, human rights law offers the opposite: a nearly universal, binding normative floor that predates the technology and does not need to be reinvented for it. The Universal Declaration of Human Rights, the International Covenant on Civil and Political Rights, and the International Covenant on Economic, Social and Cultural Rights already guarantee privacy, non discrimination, social security, food, and health, every one of which is implicated when an algorithm allocates aid, flags an asylum claim, or scores vulnerability.

The UN machinery. The UN human rights machinery grasped this early. Philip Alston's 2019 report as Special Rapporteur on extreme poverty remains the canonical warning: examining the digitization of welfare systems, he described private vendors operating in a human rights free zone and cautioned against the grave risk of stumbling, zombie-like, into a digital welfare dystopia. A year later, E. Tendayi Achiume delivered the first systematic UN analysis of how emerging digital technologies produce structurally racialized outcomes: technology is not neutral or objective, she wrote, it is fundamentally shaped by the inequalities prevalent in society, and typically makes these inequalities worse. The Office of the High Commissioner for Human Rights went further in 2021, calling for a moratorium on AI systems that pose serious risks to human rights until adequate safeguards are in place.

Scholars have built a rigorous framework on this foundation. Lorna McGregor, Daragh Murray, and Vivian Ng argue that international human rights law supplies what scattered AI ethics cannot: an end to end accountability framework spanning the full algorithmic lifecycle. The human rights approach has honest internal critics. Nathalie Smuha warns that human rights are a necessary floor but not a sufficient ceiling: without concretizing legislation, enforcement capacity, and supporting societal infrastructure, any human rights based framework risks falling short. The point is precisely why the humanitarian sector matters. Humanitarian organizations cannot wait for enforcement infrastructure to mature; they operate today, in jurisdictions where none exists. For them, the human rights framework functions less as enforceable law than as professional ethic and design constraint.

14

The UN System's Evolving Approach to AI

The United Nations came to AI along two tracks, using it and governing it, and the story of the past decade is the gradual convergence of the two. On the usage track, Global Pulse made the UN an early adopter of big data analytics; the ITU's AI for Good Global Summit became the system's flagship platform; and individual agencies developed their own instruments, from UNICEF's policy guidance on AI for children to UNHCR's rights based AI approach and WFP's system wide AI principles. On the governance track, the High level Panel on Digital Cooperation delivered The Age of Digital Interdependence in 2019, the Secretary General's Roadmap committed the UN to trustworthy, human rights based, safe, and sustainable AI, UNESCO's Recommendation became the first global normative instrument on AI ethics in 2021, and in 2022 the UN bound its own operations to ten Principles for the Ethical Use of Artificial Intelligence in the United Nations System.

The political breakthrough. In March 2024 the General Assembly adopted by consensus its first resolution on AI, led by the United States with more than 120 co sponsors, linking safe, secure, and trustworthy AI to sustainable development. In July 2024 a resolution led by China, likewise adopted by consensus, addressed AI capacity building for developing countries. The twin resolutions displayed a distinctive dynamic: strategic rivals competing for normative leadership through consensus rather than against it. The Secretary General's High level Advisory Body supplied the synthesis in Governing AI for Humanity, diagnosing the deficit bluntly: no global framework exists to govern AI, and with its development in the hands of a few multinational companies in a few countries, the impacts risk being imposed on most people without their having any say. The Global Digital Compact translated diagnosis into commitment, creating an Independent International Scientific Panel on AI and a Global Dialogue on AI Governance.

Resolution 79/325 of August 2025 gave both mechanisms their terms of reference, and 2026 brought them to life. A forty member panel drawn from all five UN regions, selected from more than 2,600 applications, held its inaugural meeting in March 2026 and elected Yoshua Bengio and Maria Ressa as co chairs. Its first annual report appeared on 1 July 2026, and the first Global Dialogue on AI Governance convened in Geneva the following week. The safety discourse born at Bletchley has migrated into the UN system, where it now shares a room with the development, human rights, and humanitarian agendas. As Secretary General Guterres framed the stakes, a world of AI haves and have-nots would be a world of perpetual instability. The UN's approach remains more scaffolding than edifice: the Scientific Panel assesses but does not regulate, and the Global Dialogue convenes but does not bind. Yet for the first time the institutions that coordinate humanitarian response, the norms that protect human rights, and the fora that govern AI sit within a single system.

15

The Ethical Tech CoLab

The Ethical Tech CoLab is a research initiative of New York University's School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft from 2024 to 2026. Its mission is to explore intervention opportunities at the intersection of emerging technologies and the human condition. Its founding observation is the one with which this paper began. The problem is power; ethics is the answer. Technology companies possess capabilities that reshape the lives of people who never chose them, and the populations most affected, refugees, trafficking survivors, communities in conflict, are precisely those least represented in the rooms where technological choices are made. The CoLab's motivation is therefore corrective, and the collaboration between a university and a technology company is itself a methodological statement: the academy contributes independence, historical depth, and the pluralism of ethical frameworks, while industry contributes operational knowledge of how systems are actually built, deployed, and misused.

The portfolio through three lenses. Each project is best understood through three lenses: the humanitarian problem it addresses, the ethical safeguards it incorporates, and the measurable public value it seeks to create. The largest body of work concerns civilian evacuation in armed conflict, through the Civilian Evacuation Risk Anticipation Index, the Evacuation Simulator, the Mariupol Corridor Severity Model, and After the Corridor. These are decision support instruments for human judgment, not automated directives, built on documented, contestable data. A second cluster addresses exploitation and traceability: the Forced Labor Structural Risk Index assesses the structural conditions under which forced labor flourishes, a fight documented in the author's fieldwork at the European Parliament and sharpened by the ILO's estimate of 236 billion dollars in annual illegal profits from forced labour. Further projects extend the same logic to cultural property provenance, disaster response, multi party negotiation, and the technology's own carbon footprint.

Three commitments define the research philosophy, resting on a foundation of international human rights law and the humanitarian principles:

  • Ethics before algorithms: the CoLab treats the current wave of AI not as an unprecedented rupture but as the latest chapter in a four thousand year effort to govern power, and begins from moral philosophy, international relations, and humanitarian ethics rather than from model architectures.
  • Ethical deliberation as method: following the Carnegie Council's framework, it identifies the ethical stakes, deliberates across plural frameworks, acts, and takes responsibility, wherever possible with those most affected rather than about them.
  • The human rights floor: it evaluates every proposed technology against internationally agreed rights rather than against corporate or institutional convenience, and treats the do no harm obligation as a hard constraint rather than an optimization target.

The CoLab's ambition, of which this paper is a first expression, is a body of research useful to policymakers, humanitarian practitioners, donors, and researchers who must make decisions about artificial intelligence without becoming AI engineers. Its wager is that the disciplines this paper has drawn together are not six literatures but one conversation, and that the institutions that learn to hold that conversation well will be the ones that earn the trust on which the legitimate use of artificial intelligence depends.

16

Conclusion

This paper set out to answer a question that is asked constantly and defined rarely: what is ethical AI? The answer it has defended is institutional rather than technical. Ethical AI is artificial intelligence governed by the same disciplines humanity has always required of power: deliberation across plural values, accountability to those affected, a floor of universal rights, and the humility to take responsibility when judgment fails. Ethics did not emerge because human beings became virtuous. It emerged because societies learned that power without restraint destroys the communities on which power depends. Artificial intelligence does not change that lesson. It makes the lesson urgent.

The humanitarian sector is where the urgency is greatest and the test is hardest. There, consent collapses, experimentation finds the least protected subjects, accountability runs to people with no vote and no purchase, and the cost of failure is measured in lives. A technology that can be governed ethically there can be governed ethically anywhere. The frameworks now taking shape, from UNESCO's Recommendation to the UN's Scientific Panel and Global Dialogue, from the EU AI Act to the ICRC's precautionary policy, are the scaffolding of that governance. Whether they become an edifice depends on whether institutions, including universities, technology companies, and the collaborations between them, do the slow deliberative work this paper has tried to model.

Societies do not flourish because they possess increasingly powerful tools. They flourish because they develop increasingly trustworthy ways of governing them. That has been the purpose of ethics for four thousand years. It is the purpose of ethical AI now.

References

Sources

  1. 01Achiume, E. Tendayi. Racial Discrimination and Emerging Digital Technologies: A Human Rights Analysis. UN Doc. A/HRC/44/57, United Nations, June 2020.
  2. 02Acton, John Emerich Edward Dalberg. Letter to Mandell Creighton, 5 Apr. 1887. Historical Essays and Studies, Macmillan, 1907.
  3. 03Adams, Rachel. The New Empire of AI: The Future of Global Inequality. Polity, 2025.
  4. 04Allen, Gregory C. “Choking Off China's Access to the Future of AI.” Center for Strategic and International Studies, 11 Oct. 2022.
  5. 05Alston, Philip. Digital Welfare States and Human Rights: Report of the Special Rapporteur on Extreme Poverty and Human Rights. UN Doc. A/74/493, United Nations, 11 Oct. 2019.
  6. 06Amnesty International and Access Now. The Toronto Declaration: Protecting the Rights to Equality and Non-Discrimination in Machine Learning Systems. 16 May 2018.
  7. 07Andersin, Emelie. “The Use of the ‘Lavender’ in Gaza and the Law of Targeting.” Journal of International Humanitarian Legal Studies, vol. 16, no. 2, 2025, pp. 336 to 370.
  8. 08Appiah, Kwame Anthony. Cosmopolitanism: Ethics in a World of Strangers. W. W. Norton, 2006.
  9. 09Aristotle. Nicomachean Ethics. Translated by Terence Irwin, 2nd ed., Hackett, 1999.
  10. 10Beduschi, Ana. “Harnessing the Potential of Artificial Intelligence for Humanitarian Action.” International Review of the Red Cross, vol. 104, no. 919, 2022, pp. 1149 to 1169.
  11. 11Bengio, Yoshua, et al. International AI Safety Report 2026. AI Security Institute, Feb. 2026.
  12. 12Benjamin, Ruha. Race After Technology: Abolitionist Tools for the New Jim Code. Polity, 2019.
  13. 13Bietti, Elettra. “From Ethics Washing to Ethics Bashing.” Proceedings of the 2020 Conference on Fairness, Accountability, and Transparency, ACM, 2020, pp. 210 to 219.
  14. 14“The Bletchley Declaration by Countries Attending the AI Safety Summit, 1 and 2 November 2023.” GOV.UK, 1 Nov. 2023.
  15. 15Bradford, Anu. Digital Empires: The Global Battle to Regulate Technology. Oxford UP, 2023.
  16. 16Buolamwini, Joy, and Timnit Gebru. “Gender Shades: Intersectional Accuracy Disparities in Commercial Gender Classification.” Proceedings of Machine Learning Research, vol. 81, 2018, pp. 77 to 91.
  17. 17Carnegie Council for Ethics in International Affairs. Ethics in International Affairs 101. Carnegie Council.
  18. 18Centre for Humanitarian Data. Peer Review Framework for Predictive Analytics in Humanitarian Response. UN OCHA, Mar. 2020.
  19. 19The Code of Hammurabi. Translated by L. W. King, 1910. The Avalon Project, Yale Law School.
  20. 20Confucius. Analects. Translated by Edward Slingerland, Hackett, 2003.
  21. 21Coppi, Giulio, Rebeca Moreno Jimenez, and Sofia Kyriazi. “Explicability of Humanitarian AI: A Matter of Principles.” Journal of International Humanitarian Action, vol. 6, art. 19, 2021.
  22. 22Council of Europe. Framework Convention on Artificial Intelligence and Human Rights, Democracy and the Rule of Law. Council of Europe Treaty Series no. 225, 5 Sept. 2024.
  23. 23Crawford, Kate. Atlas of AI: Power, Politics, and the Planetary Costs of Artificial Intelligence. Yale UP, 2021.
  24. 24The Danish Institute for Human Rights. Guidance on Human Rights Impact Assessment of Digital Activities. DIHR, 2020.
  25. 25Devidal, Pierrick. “Lost in Digital Translation? The Humanitarian Principles in the Digital Age.” International Review of the Red Cross, vol. 106, no. 925, 2024.
  26. 26Diakopoulos, Nicholas, et al. “Principles for Accountable Algorithms and a Social Impact Statement for Algorithms.” FAT/ML.
  27. 27The Engine Room and Oxfam. Biometrics in the Humanitarian Sector. The Engine Room, 2018, updated 2023.
  28. 28European Parliament and Council of the European Union. Regulation (EU) 2024/1689 (Artificial Intelligence Act). Official Journal of the European Union, 12 July 2024.
  29. 29Fjeld, Jessica, et al. Principled Artificial Intelligence: Mapping Consensus in Ethical and Rights-Based Approaches to Principles for AI. Berkman Klein Center Research Publication no. 2020-1, Harvard University, 2020.
  30. 30Floridi, Luciano, and Josh Cowls. “A Unified Framework of Five Principles for AI in Society.” Harvard Data Science Review, vol. 1, no. 1, 2019.
  31. 31Friedman, Batya, and David G. Hendry. Value Sensitive Design: Shaping Technology with Moral Imagination. MIT Press, 2019.
  32. 32Future of Life Institute. “Asilomar AI Principles.” Future of Life Institute, 2017.
  33. 33Gebru, Timnit, et al. “Datasheets for Datasets.” Communications of the ACM, vol. 64, no. 12, 2021, pp. 86 to 92.
  34. 34“Global Push for AI Governance amid Warnings of ‘Catastrophic Harm.’” UN News, 5 July 2026.
  35. 35Heinzelman, Jessica, and Carol Waters. Crowdsourcing Crisis Information in Disaster-Affected Haiti. Special Report 252, United States Institute of Peace, 2010.
  36. 36High-level Committee on Programmes, UN System Chief Executives Board for Coordination. Principles for the Ethical Use of Artificial Intelligence in the United Nations System. United Nations, Sept. 2022.
  37. 37High-Level Expert Group on Artificial Intelligence. Ethics Guidelines for Trustworthy AI. European Commission, 8 Apr. 2019.
  38. 38Hobbes, Thomas. Leviathan. Edited by Richard Tuck, Cambridge UP, 1996.
  39. 39IEEE Global Initiative on Ethics of Autonomous and Intelligent Systems. Ethically Aligned Design. 1st ed., IEEE, 2019.
  40. 40Inter-Agency Standing Committee. IASC Operational Guidance on Data Responsibility in Humanitarian Action. IASC, 2023.
  41. 41International Committee of the Red Cross. Building a Responsible Humanitarian Approach: The ICRC's Policy on Artificial Intelligence. ICRC, Nov. 2024.
  42. 42International Committee of the Red Cross. “ICRC Position on Autonomous Weapon Systems.” ICRC, 12 May 2021.
  43. 43International Labour Organization. Profits and Poverty: The Economics of Forced Labour. 2nd ed., ILO, Mar. 2024.
  44. 44International Organization for Standardization and International Electrotechnical Commission. ISO/IEC 42001:2023, Information Technology, Artificial Intelligence, Management System. ISO, Dec. 2023.
  45. 45International Telecommunication Union. United Nations Activities on Artificial Intelligence. ITU, annual editions.
  46. 46Jacobsen, Katja Lindskov. “Experimentation in Humanitarian Locations: UNHCR and Biometric Registration of Afghan Refugees.” Security Dialogue, vol. 46, no. 2, 2015, pp. 144 to 164.
  47. 47Jobin, Anna, Marcello Ienca, and Effy Vayena. “The Global Landscape of AI Ethics Guidelines.” Nature Machine Intelligence, vol. 1, 2019, pp. 389 to 399.
  48. 48Johnson, Emily. “AI Ethics Frameworks: 10 Essential Resources to Build an Ethical AI Framework.” SecureITWorld, 12 Mar. 2025.
  49. 49Juskalian, Russ. “Inside the Jordan Refugee Camp That Runs on Blockchain.” MIT Technology Review, 12 Apr. 2018.
  50. 50Kant, Immanuel. “Toward Perpetual Peace.” Practical Philosophy, translated by Mary J. Gregor, Cambridge UP, 1996.
  51. 51Karenga, Maulana. Maat, the Moral Ideal in Ancient Egypt: A Study in Classical African Ethics. Routledge, 2004.
  52. 52Kuner, Christopher, and Massimo Marelli, editors. Handbook on Data Protection in Humanitarian Action. 2nd ed., ICRC and Brussels Privacy Hub, 2020.
  53. 53Laidlaw, James. The Subject of Virtue: An Anthropology of Ethics and Freedom. Cambridge UP, 2014.
  54. 54Latonero, Mark. “Stop Surveillance Humanitarianism.” The New York Times, 11 July 2019.
  55. 55Libertas Council. “Developing Guidelines on the Intersection of Human Trafficking and Technology.” Working group meeting, European Parliament, Brussels, 3 Dec. 2024.
  56. 56Madianou, Mirca. “Technocolonialism: Digital Innovation and Data Practices in the Humanitarian Response to Refugee Crises.” Social Media + Society, vol. 5, no. 3, 2019, pp. 1 to 13.
  57. 57Madianou, Mirca. Technocolonialism: When Technology for Good Is Harmful. Polity, 2024.
  58. 58Mantelero, Alessandro. Beyond Data: Human Rights, Ethical and Social Impact Assessment in AI. T.M.C. Asser Press, 2022.
  59. 59McGregor, Lorna, Daragh Murray, and Vivian Ng. “International Human Rights Law as a Framework for Algorithmic Accountability.” International and Comparative Law Quarterly, vol. 68, no. 2, 2019, pp. 309 to 343.
  60. 60Meier, Patrick. Digital Humanitarians: How Big Data Is Changing the Face of Humanitarian Response. CRC Press, 2015.
  61. 61Moor, James H. “What Is Computer Ethics?” Metaphilosophy, vol. 16, no. 4, 1985, pp. 266 to 275.
  62. 62Mitchell, Margaret, et al. “Model Cards for Model Reporting.” Proceedings of the Conference on Fairness, Accountability, and Transparency, ACM, 2019, pp. 220 to 229.
  63. 63Mittelstadt, Brent. “Principles Alone Cannot Guarantee Ethical AI.” Nature Machine Intelligence, vol. 1, 2019, pp. 501 to 507.
  64. 64Morgenthau, Hans J. Politics Among Nations: The Struggle for Power and Peace. Alfred A. Knopf, 1948.
  65. 65Munro, Robert. “Crowdsourcing and the Crisis-Affected Community: Lessons Learned and Looking Forward from Mission 4636.” Information Retrieval, vol. 16, no. 2, 2013, pp. 210 to 266.
  66. 66National Institute of Standards and Technology. Artificial Intelligence Risk Management Framework (AI RMF 1.0). NIST AI 100-1, U.S. Department of Commerce, Jan. 2023.
  67. 67Nearing, Grey, et al. “Global Prediction of Extreme Floods in Ungauged Watersheds.” Nature, vol. 627, 2024, pp. 559 to 563.
  68. 68Noble, Safiya Umoja. Algorithms of Oppression: How Search Engines Reinforce Racism. New York UP, 2018.
  69. 69North, Douglass C. Institutions, Institutional Change and Economic Performance. Cambridge UP, 1990.
  70. 70Nussbaum, Martha C. “Patriotism and Cosmopolitanism.” Boston Review, vol. 19, no. 5, 1994.
  71. 71NYU Ethical Tech CoLab. “Publications.” NYU School of Professional Studies Center for Global Affairs and Microsoft.
  72. 72Office of the United Nations High Commissioner for Human Rights. Guiding Principles on Business and Human Rights. United Nations, 2011.
  73. 73Office of the United Nations High Commissioner for Human Rights, B-Tech Project. Taxonomy of Human Rights Risks Connected to Generative AI. OHCHR, Nov. 2023.
  74. 74Organisation for Economic Co-operation and Development. Recommendation of the Council on Artificial Intelligence. OECD/LEGAL/0449, OECD, 2019, amended 2024.
  75. 75OSCE Office of the Special Representative and Co-ordinator for Combating Trafficking in Human Beings, and Tech Against Trafficking. Leveraging Innovation to Fight Trafficking in Human Beings. OSCE, 2020.
  76. 76Ostrom, Elinor. Governing the Commons: The Evolution of Institutions for Collective Action. Cambridge UP, 1990.
  77. 77Partnership on AI. “Our Tenets.” Partnership on AI, 2016.
  78. 78Palen, Leysia, and Kenneth M. Anderson. “Crisis Informatics: New Data for Extraordinary Times.” Science, vol. 353, no. 6296, 2016, pp. 224 to 225.
  79. 79Pizzi, Michael, Mila Romanoff, and Tim Engelhardt. “AI for Humanitarian Action: Human Rights and Ethics.” International Review of the Red Cross, vol. 102, no. 913, 2020, pp. 145 to 180.
  80. 80Raji, Inioluwa Deborah, et al. “Closing the AI Accountability Gap: Defining an End-to-End Framework for Internal Algorithmic Auditing.” Proceedings of the 2020 Conference on Fairness, Accountability, and Transparency, ACM, 2020, pp. 33 to 44.
  81. 81Russell, Bertrand. Power: A New Social Analysis. George Allen and Unwin, 1938.
  82. 82Risse, Mathias. “Human Rights and Artificial Intelligence: An Urgently Needed Agenda.” Human Rights Quarterly, vol. 41, no. 1, 2019, pp. 1 to 16.
  83. 83Roberts, Huw, et al. “Global AI Governance: Barriers and Pathways Forward.” International Affairs, vol. 100, no. 3, 2024, pp. 1275 to 1286.
  84. 84Sandvik, Kristin Bergtora, Katja Lindskov Jacobsen, and Sean Martin McDonald. “Do No Harm: A Taxonomy of the Challenges of Humanitarian Experimentation.” International Review of the Red Cross, vol. 99, no. 904, 2017.
  85. 85Sastry, Girish, et al. “Computing Power and the Governance of Artificial Intelligence.” arXiv:2402.08797, Feb. 2024.
  86. 86Scharre, Paul. Four Battlegrounds: Power in the Age of Artificial Intelligence. W. W. Norton, 2023.
  87. 87Signpost AI. “Pilot Report: Signpost AI Information Assistant.” International Rescue Committee, 27 June 2025.
  88. 88Singer, Peter. The Expanding Circle: Ethics, Evolution, and Moral Progress. Princeton UP, 2011.
  89. 89Smuha, Nathalie A. “Beyond a Human Rights-Based Approach to AI Governance: Promise, Pitfalls, Plea.” Philosophy and Technology, vol. 34, no. 1, 2021, pp. 91 to 104.
  90. 90Spencer, Sarah W. Humanitarian AI: The Hype, the Hope and the Future. Network Paper 85, Humanitarian Practice Network, ODI, Nov. 2021.
  91. 91Spencer, Sarah W. Humanitarian AI Revisited: Seizing the Potential and Sidestepping the Pitfalls. Network Paper 89, Humanitarian Practice Network, ODI, May 2024.
  92. 92“Statement on Inclusive and Sustainable Artificial Intelligence for People and the Planet.” AI Action Summit, Paris, 11 Feb. 2025.
  93. 93Stilgoe, Jack, Richard Owen, and Phillip Macnaghten. “Developing a Framework for Responsible Innovation.” Research Policy, vol. 42, no. 9, 2013, pp. 1568 to 1580.
  94. 94Tomasello, Michael. A Natural History of Human Morality. Harvard UP, 2016.
  95. 95UNESCO. Recommendation on the Ethics of Artificial Intelligence. UNESCO, 23 Nov. 2021.
  96. 96UNHCR. “Our Approach to Artificial Intelligence.” UNHCR Digital Transformation Strategy 2022 to 2026, UNHCR.
  97. 97UNICEF. Policy Guidance on AI for Children 2.0. UNICEF Office of Global Insight and Policy, Nov. 2021.
  98. 98United Nations. Global Digital Compact. Annex to Resolution A/RES/79/1, United Nations, 22 Sept. 2024.
  99. 99United Nations Development Programme. AI Hub for Sustainable Development. UNDP, 2024.
  100. 100United Nations General Assembly. Enhancing International Cooperation on Capacity-Building of Artificial Intelligence. Resolution A/RES/78/311, 1 July 2024.
  101. 101United Nations General Assembly. International Covenant on Civil and Political Rights. United Nations, 16 Dec. 1966.
  102. 102United Nations General Assembly. International Covenant on Economic, Social and Cultural Rights. United Nations, 16 Dec. 1966.
  103. 103United Nations General Assembly. Resolution A/RES/79/325, Terms of Reference and Modalities for the Independent International Scientific Panel on Artificial Intelligence and the Global Dialogue on Artificial Intelligence Governance. 26 Aug. 2025.
  104. 104United Nations General Assembly. Seizing the Opportunities of Safe, Secure and Trustworthy Artificial Intelligence Systems for Sustainable Development. Resolution A/RES/78/265, 21 Mar. 2024.
  105. 105United Nations General Assembly. Universal Declaration of Human Rights. Resolution 217 A (III), 10 Dec. 1948.
  106. 106United Nations High Commissioner for Human Rights. The Right to Privacy in the Digital Age. UN Doc. A/HRC/48/31, United Nations, Sept. 2021.
  107. 107United Nations High-level Advisory Body on Artificial Intelligence. Governing AI for Humanity: Final Report. United Nations, Sept. 2024.
  108. 108United Nations Secretary-General. Our Common Agenda Policy Brief 11: UN 2.0. United Nations, Sept. 2023.
  109. 109United Nations Secretary-General. Roadmap for Digital Cooperation. UN Doc. A/74/821, United Nations, June 2020.
  110. 110UN Global Pulse. Big Data for Development: Challenges and Opportunities. United Nations, May 2012.
  111. 111UN OCHA. Briefing Note on Artificial Intelligence and the Humanitarian Sector. OCHA, 2024.
  112. 112UN OCHA. OCHA Data Responsibility Guidelines. Centre for Humanitarian Data, Oct. 2021.
  113. 113UN Secretary-General's High-level Panel on Digital Cooperation. The Age of Digital Interdependence. United Nations, June 2019.
  114. 114Université de Montréal. Montreal Declaration for a Responsible Development of Artificial Intelligence. 2018.
  115. 115Veale, Michael, and Frederik Zuiderveen Borgesius. “Demystifying the Draft EU Artificial Intelligence Act.” Computer Law Review International, vol. 22, no. 4, 2021, pp. 97 to 112.
  116. 116Weber, Max. Economy and Society: An Outline of Interpretive Sociology. Edited by Guenther Roth and Claus Wittich, U of California P, 1978.
  117. 117World Bank. “Famine Action Mechanism (FAM).” World Bank, 2018.
  118. 118World Food Programme. “WFP Releases HungerMap LIVE.” WFP, 2019.
  119. 119World Food Programme Innovation Accelerator. “New Global Principles for Innovative and Ethical AI.” WFP, 2022.
  120. 120Whittaker, Meredith, et al. AI Now Report 2018. AI Now Institute, New York University, Dec. 2018.
  121. 121Winner, Langdon. “Do Artifacts Have Politics?” Daedalus, vol. 109, no. 1, 1980, pp. 121 to 136.
  122. 122Yeung, Karen, Andrew Howes, and Ganna Pogrebna. “AI Governance by Human Rights-Centred Design, Deliberation and Oversight: An End to Ethics Washing.” The Oxford Handbook of Ethics of AI, Oxford UP, 2020, pp. 77 to 106.

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution. External programs cited are referenced as evidence, not as CoLab partnerships.

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file diff --git a/static-site/print/what-is-ethical-ai/index.txt b/static-site/print/what-is-ethical-ai/index.txt index 5c76b95dd..3ae617d03 100644 --- a/static-site/print/what-is-ethical-ai/index.txt +++ b/static-site/print/what-is-ethical-ai/index.txt @@ -1,23 +1,23 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -12:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -0:{"P":null,"c":["","print","what-is-ethical-ai",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","what-is-ethical-ai","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -14:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -16:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","print","what-is-ethical-ai",""],"q":"","i":false,"f":[[["",{"children":["print",{"children":[["slug","what-is-ethical-ai","d",null],{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,null]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +16:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:"$Sreact.suspense" -1a:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1c:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1c:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] @@ -29,7 +29,7 @@ f:["$","$1","c",{"children":["$L15",null,["$","$L16",null,{"children":["$","$17" 19:[] 10:"$W19" 11:["$","$1","h",{"children":[null,["$","$L1a",null,{"children":"$L1b"}],["$","div",null,{"hidden":true,"children":["$","$L1c",null,{"children":["$","$17",null,{"name":"Next.Metadata","children":"$L1d"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1e:Ta14, @page { size: 210mm 297mm; margin: 20mm 18mm; } @@ -245,6 +245,6 @@ a5:["$","li","120",{"children":[["$","span",null,{"className":"print-ref-num","c a6:["$","li","121",{"children":[["$","span",null,{"className":"print-ref-num","children":"122"}],["$","span",null,{"children":"Yeung, Karen, Andrew Howes, and Ganna Pogrebna. “AI Governance by Human Rights-Centred Design, Deliberation and Oversight: An End to Ethics Washing.” The Oxford Handbook of Ethics of AI, Oxford UP, 2020, pp. 77 to 106."}]]}] a7:["$","p",null,{"className":"print-note","children":"The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution. External programs cited are referenced as evidence, not as CoLab partnerships."}] 1b:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -a9:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +a9:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 18:null 1d:[["$","title","0",{"children":"What Is Ethical AI? (print edition) · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"A research collaboration between NYU's Center for Global Affairs and Microsoft, exploring tech interventions for migration, forced labor, IDPs and refugees."}],["$","meta","2",{"name":"robots","content":"noindex, nofollow"}],["$","link","3",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","4",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","5",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$La9","6",{}]] diff --git a/static-site/publications/__next._full.txt b/static-site/publications/__next._full.txt index d5435b649..42786710e 100644 --- a/static-site/publications/__next._full.txt +++ b/static-site/publications/__next._full.txt @@ -1,39 +1,39 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -11:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +11:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/chunks/36o-vt7quy27o.css","style"] -0:{"P":null,"c":["","publications",""],"q":"","i":false,"f":[[["",{"children":["publications",{"children":["__PAGE__",{}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{},null,false,null]},null,false,"$@f"]},null,false,null],"$L10",false]],"m":"$undefined","G":["$11",["$L12"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -13:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -14:I[82987,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/0-5whcwv1z55i.js"],"Reveal"] -15:I[18620,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/0-5whcwv1z55i.js"],"SectionTabs"] -16:I[46973,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/0-5whcwv1z55i.js"],"PublicationsShowcase"] -17:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +0:{"P":null,"c":["","publications",""],"q":"","i":false,"f":[[["",{"children":["publications",{"children":["__PAGE__",{}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{},null,false,null]},null,false,"$@f"]},null,false,null],"$L10",false]],"m":"$undefined","G":["$11",["$L12"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +13:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +14:I[82987,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/0jub-lza-_qsd.js"],"Reveal"] +15:I[18620,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/0jub-lza-_qsd.js"],"SectionTabs"] +16:I[46973,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/0jub-lza-_qsd.js"],"PublicationsShowcase"] +17:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 18:"$Sreact.suspense" -1b:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -1d:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +1b:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +1d:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L13",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L13",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L13",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L13",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] b:["$","div",null,{"className":"mt-12 flex flex-col gap-2 border-t border-border pt-6 text-xs text-muted sm:flex-row sm:items-center sm:justify-between","children":[["$","span",null,{"children":["© ",2026," NYU Ethical Tech CoLab"]}],["$","span",null,{"children":"Four cohorts · est. 2024-2026"}]]}] c:["$","div",null,{"className":"mt-6 space-y-3 border-t border-border pt-6 text-[11px] leading-relaxed text-muted/80","children":[["$","p","0",{"children":"The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution."}],["$","p","1",{"children":"Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners."}]]}] d:["$","$1","c",{"children":[null,["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":"$undefined","forbidden":"$undefined","unauthorized":"$undefined"}]]}] -e:["$","$1","c",{"children":[[["$","section",null,{"className":"relative overflow-hidden border-b border-border","children":[["$","span",null,{"className":"aura"}],["$","div",null,{"className":"relative mx-auto max-w-6xl px-6 py-24","children":[["$","$L14",null,{"children":["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Publications · Academic reports"}]}],["$","$L14",null,{"delay":0.05,"children":["$","h1",null,{"className":"mt-4 fluid-hero font-heading uppercase leading-[0.9]","children":["The research,"," ",["$","span",null,{"className":"display-em","children":"written up"}],"."]}]}],["$","$L14",null,{"delay":0.1,"children":["$","p",null,{"className":"mt-8 max-w-2xl text-lg leading-relaxed text-muted","children":"Each research question the CoLab takes on is written up as an academic report. Browse the catalogue by topic, pick a title to read what it asks and what it found, then open the report itself. Titles still in preparation are shelved alongside the published ones, so you can see where the work is going."}]}]]}]]}],["$","$L15",null,{}],["$","div",null,{"className":"pt-12","children":["$","$L16",null,{}]}]],[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/36o-vt7quy27o.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/0-5whcwv1z55i.js","async":true,"nonce":"$undefined"}]],["$","$L17",null,{"children":["$","$18",null,{"name":"Next.MetadataOutlet","children":"$@19"}]}]]}] +e:["$","$1","c",{"children":[[["$","section",null,{"className":"relative overflow-hidden border-b border-border","children":[["$","span",null,{"className":"aura"}],["$","div",null,{"className":"relative mx-auto max-w-6xl px-6 py-24","children":[["$","$L14",null,{"children":["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Publications · Academic reports"}]}],["$","$L14",null,{"delay":0.05,"children":["$","h1",null,{"className":"mt-4 fluid-hero font-heading uppercase leading-[0.9]","children":["The research,"," ",["$","span",null,{"className":"display-em","children":"written up"}],"."]}]}],["$","$L14",null,{"delay":0.1,"children":["$","p",null,{"className":"mt-8 max-w-2xl text-lg leading-relaxed text-muted","children":"Each research question the CoLab takes on is written up as an academic report. Browse the catalogue by topic, pick a title to read what it asks and what it found, then open the report itself. Titles still in preparation are shelved alongside the published ones, so you can see where the work is going."}]}]]}]]}],["$","$L15",null,{}],["$","div",null,{"className":"pt-12","children":["$","$L16",null,{}]}]],[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/36o-vt7quy27o.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/0jub-lza-_qsd.js","async":true,"nonce":"$undefined"}]],["$","$L17",null,{"children":["$","$18",null,{"name":"Next.MetadataOutlet","children":"$@19"}]}]]}] 1a:[] f:"$W1a" 10:["$","$1","h",{"children":[null,["$","$L1b",null,{"children":"$L1c"}],["$","div",null,{"hidden":true,"children":["$","$L1d",null,{"children":["$","$18",null,{"name":"Next.Metadata","children":"$L1e"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -12:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +12:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] 1c:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -1f:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +1f:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 19:null 1e:[["$","title","0",{"children":"Publications · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"Academic reports from the Ethical Tech CoLab — one write-up per research question, across evacuation, cultural heritage, supply-chain traceability, and diplomacy."}],["$","link","2",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","3",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","4",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L1f","5",{}]] diff --git a/static-site/publications/__next._head.txt b/static-site/publications/__next._head.txt index 21b278d11..c3aa76dca 100644 --- a/static-site/publications/__next._head.txt +++ b/static-site/publications/__next._head.txt @@ -1,6 +1,6 @@ 1:"$Sreact.fragment" -2:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -3:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +2:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +3:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 4:"$Sreact.suspense" -5:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +5:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 0:{"rsc":["$","$1","h",{"children":[null,["$","$L2",null,{"children":[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]]}],["$","div",null,{"hidden":true,"children":["$","$L3",null,{"children":["$","$4",null,{"name":"Next.Metadata","children":[["$","title","0",{"children":"Publications · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"Academic reports from the Ethical Tech CoLab — one write-up per research question, across evacuation, cultural heritage, supply-chain traceability, and diplomacy."}],["$","link","2",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","3",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","4",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L5","5",{}]]}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/publications/__next._index.txt b/static-site/publications/__next._index.txt index f98a4fd90..18b9d82b6 100644 --- a/static-site/publications/__next._index.txt +++ b/static-site/publications/__next._index.txt @@ -1,13 +1,13 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -d:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +d:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"Link"] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} 8:["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$Ld",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$Ld",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$Ld",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$Ld",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] diff --git a/static-site/publications/__next._tree.txt b/static-site/publications/__next._tree.txt index 0927e691a..ae3d322f7 100644 --- a/static-site/publications/__next._tree.txt +++ b/static-site/publications/__next._tree.txt @@ -1,8 +1,8 @@ -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/chunks/36o-vt7quy27o.css","style"] 0:{"tree":{"name":"","param":null,"prefetchHints":16,"slots":{"children":{"name":"publications","param":null,"prefetchHints":0,"slots":{"children":{"name":"__PAGE__","param":null,"prefetchHints":0,"slots":null}}}}},"staleTime":300,"buildId":"static-snapshot"} diff --git a/static-site/publications/__next.publications.txt b/static-site/publications/__next.publications.txt index e17ca4ded..e8bfa323c 100644 --- a/static-site/publications/__next.publications.txt +++ b/static-site/publications/__next.publications.txt @@ -1,5 +1,5 @@ 1:"$Sreact.fragment" -2:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -3:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +2:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +3:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] 4:[] 0:{"rsc":["$","$1","c",{"children":[null,["$","$L2",null,{"parallelRouterKey":"children","template":["$","$L3",null,{}]}]]}],"isPartial":false,"staleTime":300,"varyParams":"$W4","buildId":"static-snapshot"} diff --git a/static-site/publications/after-the-corridor/__next._full.txt b/static-site/publications/after-the-corridor/__next._full.txt index c502d0764..1bdf3ba1e 100644 --- a/static-site/publications/after-the-corridor/__next._full.txt +++ b/static-site/publications/after-the-corridor/__next._full.txt @@ -1,24 +1,24 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/16n96jxjvmlf9.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -12:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/16n96jxjvmlf9.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/chunks/36o-vt7quy27o.css","style"] -0:{"P":null,"c":["","publications","after-the-corridor",""],"q":"","i":false,"f":[[["",{"children":["publications",{"children":["after-the-corridor",{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,"$@10"]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -14:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/16n96jxjvmlf9.js"],"Link"] -15:I[82987,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/16n96jxjvmlf9.js"],"Reveal"] -16:I[34010,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/16n96jxjvmlf9.js"],"ReportBook"] -20:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -22:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +0:{"P":null,"c":["","publications","after-the-corridor",""],"q":"","i":false,"f":[[["",{"children":["publications",{"children":["after-the-corridor",{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,"$@10"]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/16n96jxjvmlf9.js"],"Link"] +15:I[82987,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/16n96jxjvmlf9.js"],"Reveal"] +16:I[34010,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/16n96jxjvmlf9.js"],"ReportBook"] +20:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +22:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 23:"$Sreact.suspense" 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] @@ -31,9 +31,9 @@ f:["$","$1","c",{"children":[[["$","section",null,{"className":"relative overflo 1f:[] 10:"$W1f" 11:["$","$1","h",{"children":[null,["$","$L20",null,{"children":"$L21"}],["$","div",null,{"hidden":true,"children":["$","$L22",null,{"children":["$","$23",null,{"name":"Next.Metadata","children":"$L24"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] -25:I[18620,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/16n96jxjvmlf9.js"],"SectionTabs"] -38:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +25:I[18620,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/16n96jxjvmlf9.js"],"SectionTabs"] +38:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:["$","span",null,{"aria-hidden":true,"children":"↗"}] 18:["$","$L14",null,{"href":"/demos","className":"btn-sweep inline-flex items-center gap-2 rounded-full border border-border px-5 py-2.5 text-sm font-semibold text-foreground transition-colors hover:border-border-strong","children":["See the prototypes running ",["$","span",null,{"aria-hidden":true,"children":"→"}]]}] 19:["$","$L25",null,{}] @@ -133,6 +133,6 @@ f:["$","$1","c",{"children":[[["$","section",null,{"className":"relative overflo 81:["$","li","35",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"36"}],["$","span",null,{"children":["$","a",null,{"href":"https://www.fmreview.org/issue70/","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"“Knowledge, Voice and Power,” Forced Migration Review 70 (2022); Mauricio Vitoria et al., “The Global Summit of Refugees and the Importance of Refugee Self-Representation,” Forced Migration Review (2018)."}]}]]}] 82:["$","li","36",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"37"}],["$","span",null,{"children":["$","span",null,{"className":"text-foreground/80","children":"Innovations for Poverty Action, Displaced Livelihoods Initiative: Call for Proposals, Round V (IPA, 2026); IPA, DLI Eligibility Self-Assessment and Budget Template, Round V (2026)."}]}]]}] 21:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -83:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +83:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 24:[["$","title","0",{"children":"After the Corridor · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"An Ethical Tech CoLab report carrying five fielded evacuation prototypes past the corridor and into the camp: the measurement, financial-modeling, and rights infrastructure that can shorten protracted displacement, grounded at Dzaleka Refugee Camp, Malawi."}],["$","link","2",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","3",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","4",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L83","5",{}]] 39:null diff --git a/static-site/publications/after-the-corridor/__next._head.txt b/static-site/publications/after-the-corridor/__next._head.txt index 4395acd59..4a712201c 100644 --- a/static-site/publications/after-the-corridor/__next._head.txt +++ b/static-site/publications/after-the-corridor/__next._head.txt @@ -1,6 +1,6 @@ 1:"$Sreact.fragment" -2:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -3:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +2:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +3:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 4:"$Sreact.suspense" -5:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +5:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 0:{"rsc":["$","$1","h",{"children":[null,["$","$L2",null,{"children":[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]]}],["$","div",null,{"hidden":true,"children":["$","$L3",null,{"children":["$","$4",null,{"name":"Next.Metadata","children":[["$","title","0",{"children":"After the Corridor · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"An Ethical Tech CoLab report carrying five fielded evacuation prototypes past the corridor and into the camp: the measurement, financial-modeling, and rights infrastructure that can shorten protracted displacement, grounded at Dzaleka Refugee Camp, Malawi."}],["$","link","2",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","3",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","4",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L5","5",{}]]}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/publications/after-the-corridor/__next._index.txt b/static-site/publications/after-the-corridor/__next._index.txt index ca305305b..3e5286497 100644 --- a/static-site/publications/after-the-corridor/__next._index.txt +++ b/static-site/publications/after-the-corridor/__next._index.txt @@ -1,13 +1,13 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/16n96jxjvmlf9.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -d:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/16n96jxjvmlf9.js"],"Link"] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/16n96jxjvmlf9.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +d:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/16n96jxjvmlf9.js"],"Link"] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} 8:["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$Ld",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$Ld",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$Ld",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$Ld",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] diff --git a/static-site/publications/after-the-corridor/__next._tree.txt b/static-site/publications/after-the-corridor/__next._tree.txt index 275d2a2f6..fee2a93a7 100644 --- a/static-site/publications/after-the-corridor/__next._tree.txt +++ b/static-site/publications/after-the-corridor/__next._tree.txt @@ -1,8 +1,8 @@ -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/chunks/36o-vt7quy27o.css","style"] 0:{"tree":{"name":"","param":null,"prefetchHints":16,"slots":{"children":{"name":"publications","param":null,"prefetchHints":0,"slots":{"children":{"name":"after-the-corridor","param":null,"prefetchHints":0,"slots":{"children":{"name":"__PAGE__","param":null,"prefetchHints":0,"slots":null}}}}}}},"staleTime":300,"buildId":"static-snapshot"} diff --git a/static-site/publications/after-the-corridor/__next.publications.txt b/static-site/publications/after-the-corridor/__next.publications.txt index e17ca4ded..e8bfa323c 100644 --- a/static-site/publications/after-the-corridor/__next.publications.txt +++ b/static-site/publications/after-the-corridor/__next.publications.txt @@ -1,5 +1,5 @@ 1:"$Sreact.fragment" -2:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -3:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +2:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +3:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] 4:[] 0:{"rsc":["$","$1","c",{"children":[null,["$","$L2",null,{"parallelRouterKey":"children","template":["$","$L3",null,{}]}]]}],"isPartial":false,"staleTime":300,"varyParams":"$W4","buildId":"static-snapshot"} diff --git a/static-site/publications/after-the-corridor/index.html b/static-site/publications/after-the-corridor/index.html index 1c217c651..a9f6bf9ad 100644 --- a/static-site/publications/after-the-corridor/index.html +++ b/static-site/publications/after-the-corridor/index.html @@ -1 +1 @@ -After the Corridor · NYU Ethical Tech CoLab

After the Corridor

From AI-Informed Evacuation to Digital Public Goods for Refugee Economic Inclusion

Ethical Tech CoLabNYU Center for Global AffairsJuly 2026

Carolina de Almeida Pernambuco Moron, Christine Lumen, Alana Robertson, Melanie MacKew, Carlos D. Ruiz, India Clarke, and Yago Rocha, with Teresa Cantero (PhD Candidate, Universidad Carlos III de Madrid; Adjunct Professor, IE University; Visiting Scholar, NYU CGA). Faculty directors: Prof. Yorke E. Rhodes III and Sylvia Maier.

6

built instruments: five fielded evacuation prototypes and a completed forced-labour risk index, all open source

Jump to source

53k to 60k

people now at Dzaleka, a camp built in 1994 for 10,000 to 12,000

Jump to source

90%

cut to UNHCR's Malawi budget, from about US$8 million to about US$1 million

Jump to source

0

biometric identity systems: the category the CoLab refuses to build, recorded as a versioned decision pattern

Jump to source

How can AI inform evacuation decisions, and what happens after? This report examines the CoLab's five fielded evacuation prototypes, their limitations and research frontiers, and extends the arc to the camp: the measurement, financial-modeling, and rights infrastructure that can shorten protracted displacement and improve lives. Dzaleka Refugee Camp in Malawi is the case study ground.

01

Executive Summary

The Ethical Tech CoLab's evacuation portfolio was built to answer a protection question: how can and should AI tools be used to enhance the protection and evacuation of civilians during armed conflict, and in what ways do they support or challenge existing obligations under International Humanitarian Law?

Across five fielded prototypes, a crisis-information index, a unified risk platform, a calibrated cost-and-risk decision framework, an agent-based behavior model, and a readiness-under-uncertainty simulator, one finding recurred: the binding constraint on protection is rarely the absence of a route or a resource, but the quality, accessibility, and equity of the information people and responders act on. IHL frames evacuation as a last-resort duty; the technology's job is to make that duty implementable under uncertainty.

What happens after the evacuation? A successful evacuation typically ends in a camp, where the modal outcome is years of protracted displacement. Using Dzaleka Refugee Camp, Malawi, where humanitarian financing has collapsed while an encampment policy bars residents from work, we show that the same measurement discipline, financial modeling, and uncertainty methods that served evacuation convert directly into durable-solutions infrastructure: an open Livelihoods Information and Access Index, a resettlement-and-inclusion financial model built on the CoLab's operational evacuation cost engine, privacy-preserving cash and credential rails, and a civil-society legal-assistance map. We also state what we will not build, biometric identity for displaced people, and why that refusal is itself a public good.

Most of the world's refugees live in protracted situations, and the modal outcome of reaching safety is years, sometimes decades, inside a camp or settlement. The legal architecture that governs what comes next is the Geneva refugee-protection consensus: the 1951 Convention relating to the Status of Refugees and its 1967 Protocol, the doctrine of durable solutions that UNHCR administers, and the 2018 Global Compact on Refugees. That consensus recognizes three ways a displacement can end well, voluntary repatriation, local integration, and resettlement, and each of them is gated far more by information and legal status than by physical capacity.

The CoLab's built work already reaches beyond the evacuation corridor. The same measurement discipline produced the Forced Labor Structural Risk Index, a completed instrument that maps the structural risk of forced labour across countries and sub-national regions, the exploitation that displacement, encampment, and the loss of the right to work make more likely. Read together, the evacuation models and the risk index share one purpose: to improve the information on which protection depends, and through it to reduce forced labour and the vulnerabilities that displaced people carry.

This report keeps three registers distinct throughout: what is built and validated, what is built but not yet validated, and what is proposed. No claim in these pages exceeds the register of the tool it describes.

The ask, in one paragraph. The Ethical Tech CoLab seeks grant funding to develop this technology infrastructure work as open, portable public goods, to improve the lives of displaced communities. It would be validated first at Dzaleka with the field partner Fraternidade Sem Fronteiras, and we are open to other partners with the same mission-driven motivations. Our intent is developing tools and releasing them to the public so that any research or practitioner team can adopt them. One possibility is the Displaced Livelihoods Initiative run by Innovations for Poverty Action and J-PAL, in its Infrastructure and Public Goods track. The report is written to align with that submission and to stand on its own for any peer funder or academic reviewer.

02

The Research Question, and Why It Runs Past the Corridor

The protection question and the accounting gap. The portfolio's organizing question comes from Teresa Cantero's doctoral research at the Instituto de Derechos Humanos Gregorio Peces-Barba, Universidad Carlos III de Madrid: how can and should artificial intelligence be used to enhance the protection and evacuation of civilians during armed conflict, and in what ways does it support or challenge the implementation of existing obligations under international humanitarian law.

Under that law, evacuation is not discretionary. When the security of the population or imperative military reasons demand it, moving civilians out of danger is a binding duty under Article 49 of the Fourth Geneva Convention, Articles 57 and 58 of Additional Protocol I, and Article 17 of Additional Protocol II. A movement into which people are coerced is forcible transfer, a war crime regardless of motive. The duty runs wider than the battlefield: the doctrine of the responsibility to protect holds that where a state cannot or will not shield its population from mass atrocity, that responsibility passes to the international community, which frames protection as an obligation rather than a matter of goodwill.

Two structural facts motivate the program. The first is a protection gap. States move quickly to adopt AI for targeting and operational advantage, yet rarely operationalize computational models to protect civilians or to make an evacuation decision legible under the incomplete, fast-changing information those decisions face. The second is an accounting gap. Displacement is tracked in detail, on the order of 117 million people forcibly displaced worldwide by 2025, yet there is no global count of evacuated civilians, because evacuation is fragmented across militaries, governments, humanitarian agencies, volunteers, and civilians acting on their own.

The legal and institutional architecture. The 1951 Convention and its 1967 Protocol define who is a refugee and fix the cornerstone duty of non-refoulement, that no one may be returned to a place where they face persecution. The 2018 Global Compact on Refugees adds a cooperative framework organized around four objectives: easing pressure on host countries, enhancing refugee self-reliance, expanding access to third-country solutions, and supporting conditions for return in safety and dignity. International humanitarian law governs conduct in armed conflict, and the responsibility to protect frames the state's duty toward its own population.

The institutions matter as much as the instruments. UNHCR leads protection and durable solutions for refugees; the International Organization for Migration is the United Nations migration agency; the Office for the Coordination of Humanitarian Affairs coordinates emergency response through the cluster approach; and the Office of the High Commissioner for Human Rights anchors the broader human-rights instruments that apply regardless of a person's status. One structural limitation shapes everything downstream: many host states are not parties to the 1951 Convention or maintain significant reservations, so the protection a displaced person actually receives depends heavily on where they land.

The three durable solutions. The arc does not end when the shooting stops. An evacuation succeeds when a person reaches relative safety, and safety, in practice, usually means a camp. There the analytical problem inverts. The question is no longer whether to move and how, but how to leave well now that one has arrived.

The three durable solutions recognized by the refugee-protection regime, voluntary repatriation to the country of origin, local integration in the country of asylum, and resettlement to a third country, are the recognized routes out. Each is gated by information and legal status as much as by physical or economic capacity, a point the regime's own scholarship makes plainly. Resettlement reaches only a small fraction of those who need it, and complementary pathways such as labour or study routes remain modest in scale, which is why local integration and the conditions for safe return carry most of the weight in practice. This report treats the durable-solutions phase as the natural continuation of the evacuation research question, and asks what open measurement and modeling infrastructure would help displaced people and the organizations serving them find the fastest defensible route out.

03

The Built Portfolio

This section describes what the CoLab has already built, and is candid about how mature each piece is, because a reader deciding whether to fund the work is owed that precision. Three registers are kept distinct. The Evacuation Risk and Cost Framework is built and validated in its evacuation form. The Evacuation Information Index, the Exodus platform, and the two simulators are built and operational but not yet expert-validated or calibrated. The Forced Labor Structural Risk Index is built and completed, documented in a methodology page and an associated publication. The livelihoods instruments the report proposes, the Livelihoods Information and Access Index and the Resettlement and Inclusion Capacity Simulator, together with the value-transfer and legal-map builds, are proposed, and their validation is future work. All are open source.

PrototypeFunctionCore method
Evacuation Information Index (EII)Scores quality, accessibility, and equity of crisis information by geographyComposite index on INFORM, ACLED, and FSI logic; live news and conflict feeds
Exodus, Civilian Evacuation Risk PlatformUnifies EII, the cost model, and endangerment and feasibility assessmentFastAPI backend; shared calculators and data layer
ERCF, Evacuation Risk and Cost FrameworkEstimates human and financial cost of evacuating against stayingSeven-dimension scoring; calibrated cost model; break-even analysis
Evacuation Behavior SimulatorModels how information and demographics shape who leaves, and whenAgent-based model; probabilistic status lifecycle
Readiness and Uncertainty SimulatorModels how intelligence uncertainty degrades destination choicesMonte Carlo; gatekeeper factors; uncertainty slider
The five fielded evacuation prototypes. All are open source at github.com/Ethical-Tech-CoLab.

The Evacuation Information Index. The index treats crisis information itself as the object of measurement. It scores three families of factors, impact, conditions, and complexity, on a scale of 1 to 5, using weights synthesized from the INFORM Severity Index, the Armed Conflict Location and Event Data project, and the Fragile States Index. A live backend adds current news and a conflict timeline for each crisis, and a map layer can overlay AI-derived building-damage and route-damage tiles from Microsoft HASTE and Planet imagery. The design is deliberately transparent: weights are labeled provisional pending expert validation, and where the hosted site cannot run the live backend, responses are baked into a clearly labeled static snapshot so that nothing is presented as live when it is not.

The index has real limits, and the report states them rather than hiding them.

  • Weights are borrowed by analogy, not yet expert-validated or sensitivity-tested; a reviewer will read pre-validation weights as pre-evidence.
  • Source and language bias: feeds skew to English-language, internationally indexed reporting, so the populations with the worst information access are least visible to the instrument that scores information access. The index must correct for this, not inherit it.
  • The damage overlay is currently self-hosted with no public interface, so for most users the combined information-and-damage view is empty until it is connected.

Future research on the index. Run the three-phase validation the CoLab's methodology work already specifies, budget-allocation and Delphi elicitation with 8 to 12 domain experts, Fuzzy-AHP on contested weights, then historical case calibration, and add an equity layer measuring who receives information, disaggregated by gender, age, nationality, and residence inside or outside a camp. That equity layer is the move that turns the index into the livelihoods instrument described later in this report.

Exodus, the integration layer. Exodus unifies the crisis map, the cost-and-scenario model, and an endangerment-and-feasibility assessment behind a single backend, design language, and interface, including calculators for the cost of staying, the cost of evacuating, resource needs, and scenario management. Integration is not validation. Unifying three research prototypes behind a more authoritative interface inherits their uncalibrated assumptions, which raises rather than lowers the duty to label confidence honestly. The same interface is the seam where the durable-solutions module plugs in. Its stated limitation is that it is not automatically updated; its future research is to validate the approaches, weights, and measurements.

The Evacuation Risk and Cost Framework. Now at version 7.2, the framework is the CoLab's calibrated decision-support tool for the human and financial cost of civilian evacuation, and it is the one component of the portfolio that is validated rather than provisional, in its evacuation form. It scores scenarios across seven dimensions, kinetic threat, mobility, authorization, logistics, destination, urgency, and information, estimates evacuation resources and in-situ assistance costs, and produces a break-even analysis: the day after which a one-time evacuation becomes cheaper than the mounting daily cost of sustaining a population in place. Its cost parameters are sourced to humanitarian standards and individually tagged as Validated, Estimated, or Unvalidated, drawing on UNHCR water and sanitation guidance, the Sphere standards, and World Food Programme figures.

The mortality model is disclosed as indicative rather than predictive, with a reported log-log fit of 0.855, a leave-one-out cross-validation of 0.807, and a documented case where the structure fails. The financial estimates are substantially more reliable than the mortality figures, which is why the durable-solutions adaptation leads with cost. The framework's ethical architecture is explicit: it does not place a monetary value on a human life. Its figures are logistics estimates, its labels are descriptive rather than prescriptive, and it presents data to support a decision rather than issuing one. That framing carries verbatim into every extension in this report.

The framework's limitations are equally explicit.

  • It models the immediate field operation only, on the order of US$134 per person at moderate risk over 50 kilometers by ground. Real operations, Kosovo 1999 at about US$562 per person and Lebanon 2006 at about US$4,500 to US$5,700, add international transport, reception, and hosting. Extending the cost boundary to the full displacement journey, including protracted encampment, is the frontier the durable-solutions model takes up.
  • Ground-truthing against OCHA's Financial Tracking Service is on the backlog and belongs in the funded work plan.

The Evacuation Behavior Simulator. An agent-based model of how information spread and demographics shape who leaves and when. Agents move through a lifecycle from unaware to seeking, milling, evacuating, and done, with family clusters that wait for all members, confirmation loops whose speed depends on how clear the information is, neighbor influence, and calibrated delays for elders, young children, pregnant individuals, and unaccompanied minors. It is pedagogical rather than predictive, and its household-composition logic transfers directly to modeling who is left behind by information failures inside a camp. Its behavior may vary by society, culture, location, personality, income, and education, and the future research is to build intersectionality between behavior simulations.

The Readiness and Uncertainty Simulator. A seeded, fully reproducible instrument that models how uncertainty in field intelligence degrades an evacuation decision. Destinations are scored across seven factors, 500 Monte Carlo trials per group-and-destination pair convert readiness into a distribution of success probabilities, and an uncertainty control degrades every factor's confidence at once, so a user can watch predicted outcomes collapse while nothing on the ground has changed.

Two design choices carry into the rest of this report. First, three factors are non-substitutable gatekeepers, security, the consent of the local authority, and the willingness of the host community, any of which caps readiness: a destination that is dangerous, legally closed, or unwelcoming cannot be redeemed by good logistics. Second, the model insists on the difference between a factor that is unknown, a gap that better information can close, and a factor that is a confirmed refusal, a settled political exclusion. Treating the two as the same would make an unresolvable refusal look like a solvable data problem. Structurally, assigning groups to destinations under consent, willingness, security, and capacity constraints, in the presence of uncertainty, is resettlement matching, which is the transfer the durable-solutions model makes. The simulator has not yet been stress tested with different AI models.

The Forced Labor Structural Risk Index. The index applies the same composite-index method to a different harm. It is a completed, interactive index that maps the structural conditions enabling forced labour across 184 countries on a scale from 0 to 1, with national and sub-national layers, so that the risk can be seen and compared rather than assumed. It belongs in this report because displacement is a forced-labour risk multiplier: the loss of legal status and the right to work, the collapse of assistance, and weakened oversight are precisely the conditions under which exploitation grows, and forced labour is a defined harm in international law with an estimated 27.6 million people subject to it worldwide.

Its significance for the durable-solutions argument is conceptual: measuring structural forced-labour risk and measuring livelihoods-information access are two halves of one protective picture, one showing where exploitation is likely, the other showing whether the information that could prevent it is reaching people. As a composite index its weights remain open to expert re-elicitation, it is not automatically updated, and sponsorship mechanisms remain unsourced at country scale while an eight-country hand-coded kafala signal feeds two domains in the shipped build. Writing an ILO-endorsed kafala passage on top of that unresolved discrepancy would have overstated the index's grounding.

Cross-cutting synthesis. Read together, the built prototypes yield a short set of lessons. Information access is the binding constraint, and it is the thread that carries from the corridor into the camp. Uncertainty has to be modeled explicitly, with the unknown kept separate from the confirmed, or the analysis misleads by omission. We address gaps in data by generating synthetic data that mimics real-life data, and address the residual uncertainty with Monte Carlo simulation. A composite index built from the sequence of borrowing precedents, eliciting expert weights, and calibrating against cases is portable infrastructure: the same method that scores crisis information also scores structural forced-labour risk. And cost modeling in a context of human lives has to stay descriptive rather than prescriptive, following the framework's refusal to monetize life.

04

After Arrival: Dzaleka and the Second Emergency

A successful evacuation deposits people into a slower emergency. The concept of a protracted refugee situation, conventionally a population of at least 25,000 of the same nationality in exile for five years or more, describes where most refugees actually live, and care-and-maintenance encampment becomes the default by inertia rather than by anyone's plan.

The duration is not a rounding error in a person's life. By UNHCR's own accounting, the average duration of a protracted refugee situation rose from about 9 years in 1993 to roughly 17 years by 2003, and while later analysis rightly cautions that a single average conceals wide variation, the order of magnitude is not in dispute: protracted exile routinely lasts a decade or more, and often a generation. The literature on encampment and self-settlement has long argued that the camp is often less a humanitarian necessity than an administrative and political choice, and that its costs to refugees, in mobility, work, and self-reliance, are systematically understated.

Dzaleka and the funding collapse. Dzaleka Refugee Camp in Dowa District, Malawi, is the sharp end of that pattern. Built in 1994 for 10,000 to 12,000 people, it now holds on the order of 53,000 to 60,000 people, predominantly from the Democratic Republic of the Congo, Burundi, and Rwanda, and it is caught in the funding collapse of 2025 and 2026.

The numbers describe a system in retreat.

  • UNHCR's Malawi budget fell from about US$8 million to about US$1 million, a cut of roughly 90 percent, ending general non-food distributions and laying off protection staff, including paralegals.
  • World Food Programme assistance is secured only through the middle of 2026, with about US$11 million still needed to restore full rations through December.
  • Monthly cash support has fallen toward the range of US$6 to US$9 per person, down from roughly US$100, forcing skipped meals and other negative coping.

Malawi's encampment policy, enforced since March 2023, bars most residents from working outside the camp, so the population cannot substitute self-reliance for collapsing aid. Local civil society, including Inua Advocacy, argues the crisis is inseparable from that policy and is pressing for the right to work; weakened oversight is meanwhile reported to embolden trafficking networks.

This is not a context for piloting speculative technology on vulnerable people. It is a context where the highest-value contribution is measurement, evidence, and legibility: tools that help residents and the organizations serving them see and argue for the fastest route out, and that document the true cost of the status quo. This is a global humanitarian emergency that needs collective collaborative effort, and building infrastructure, digital goods and services, information, training, and capacity might transform generational poverty and increase access to resources.

Encampment as a driver of forced labour. These are the textbook conditions for forced labour. When legal work is barred and assistance collapses, the informal and coerced economy fills the gap. Reducing that exposure is a direct objective of the CoLab's work: the Forced Labor Structural Risk Index names where the risk concentrates, and the livelihoods-information instrument measures whether the information that could protect people, about their rights, their options, and legitimate work, is actually reaching them. That link has no official validation yet, and the future research is to separate recruitment, exploitation, and monetization by intersectionality using social demography, and to deepen the demand side.

05

The Financial Model: From Evacuation Break-Even to Inclusion Break-Even

The financial-modeling engine already exists, validated for one problem and now re-pointed to another, and the distinction matters. The Evacuation Risk and Cost Framework is validated for evacuation costing; its encampment-versus-inclusion parameterization is new and unvalidated, and most of its durable-solutions cost parameters are tagged Estimated or Unvalidated. What transfers cleanly is the architecture and the break-even mathematics, a one-time investment set against a mounting recurring cost, not a set of validated cost figures for the new use.

The engine, re-pointed.

  • Trajectory A prices status-quo encampment as a recurring per-person, per-year assistance cost, covering rations, water and sanitation, health, protection, and administration, using the same Sphere, UNHCR, and WFP parameter families the framework already carries, with the same Validated, Estimated, or Unvalidated tags.
  • Trajectory B prices an inclusion or resettlement pathway as a front-loaded investment, covering documentation, recognition of qualifications, transport, start-up capital, and host co-investment.
  • The output is the break-even year after which the status quo costs more, which is the framework's break-even chart with its axes relabeled.

The Resettlement and Inclusion Capacity Simulator is that engine re-pointed, and re-earning its validity in the durable-solutions domain is the work, not a formality. The capacity-and-willingness layer transfers from the readiness simulator: it maps destinations, whether resettlement states or local-integration pathways, against the non-substitutable gatekeepers of legal consent, host willingness, and security, alongside the substitutable capacities of housing, jobs, and services, while preserving the distinction between an option that is unknown and one that is refused, so that an unlobbied state is never mistaken for a closed door. The uncertainty layer carries over as well, so that thin data on host capacity or a state's posture visibly widens the confidence band and prevents false precision in an advocacy number.

What the evidence predicts. The World Bank's recent work on refugee self-reliance in Sub-Saharan Africa indicates that if refugees had the right to move and work, they would contribute to the hosting country's economy, with real fiscal and economic gains. The honest caveat is built into that same literature. The World Bank finds that self-reliance remains elusive across much of the region, because camps often sit in marginal, high-poverty areas where even full local integration would lift incomes only a little, and the scholarship on self-reliance programming warns that the concept is often invoked to justify the withdrawal of assistance rather than to expand rights.

The simulator has to be able to return the answer that inclusion does not pay here, where host markets are too thin, or it is advocacy rather than analysis. Its output is a defensible, reproducible sentence of the form: sustaining this population at Dzaleka costs approximately this much per year; this pathway costs approximately that much up front and breaks even in a given year, subject to these gatekeepers and this uncertainty band. The simulator could also provide country comparisons across different legal models and measure the economic impact of different refugee policies, and listing the policies of each hosting country would better inform humanitarian organizations and populations.

What the literature says the model will find. The engine formalizes an argument the development-economics evidence already supports.

  • UNHCR estimates that if freedom of movement and the right to work were relaxed so refugees could earn on par with hosts, complementary assistance costs would fall from about US$3.2 billion to roughly US$900 million a year.
  • A joint World Bank and UNHCR benchmark puts the annual cost of bringing every refugee in low and middle income host countries up to the global poverty line at almost US$62 billion in a scenario where refugees earn nothing; refugees' own earnings already cover roughly US$40 billion of that, leaving about US$22 billion.
  • World Bank analysis estimates that in Uganda's inclusive regime, economic participation saves about US$150 per refugee per year, roughly US$225 million across 1.5 million refugees.
  • The same World Bank work finds that around Kakuma, refugee presence raised the host county's gross regional product by about 3.4 percent.
  • It also reports that Colombia's regularization of Venezuelan migrants raised beneficiaries' incomes by about 30 percent and their formal employment by around 10 percentage points, with little displacement of host workers.

06

Digital Public Goods for Economic Inclusion

The toolkit is ordered from the safest build to the one that demands the most caution. The organizing principle is data minimization: build the thinnest system that solves the problem, and refuse the parts whose risks outweigh their benefit for this population. A digital public good, in the sense set out by the Digital Public Goods Alliance and endorsed within the United Nations system, is open-source software, open data, an open standard, or an open model that is relevant to the Sustainable Development Goals, does no harm by design, and, importantly for this population, minimizes the collection of personally identifiable information. The Principles for Digital Development point the same way, toward privacy, reuse, and designing with the user. Measured against those criteria, the CoLab's contributions are intended to qualify as digital public goods rather than to resemble them loosely.

Measurement: the Livelihoods Information and Access Index. The measurement instrument is the Evacuation Information Index re-specified for displaced livelihoods and legal rights: an open, versioned instrument measuring the quality, accessibility, and equity of information about work rights, employment pathways, financial services, and entrepreneurship support reaching displaced populations. The architecture transfers from the evacuation index, but the construct does not, and saying so is part of the method. Measuring information access to work rights is a different thing from scoring crisis severity, so the index has to establish its own construct validity from the ground up rather than inherit it.

In draft form the index is a matrix. Its three dimensions, quality, meaning whether information is accurate, current, and actionable; accessibility, meaning whether people can reach it in a language and channel they use; and equity, meaning whether it reaches sub-populations evenly, are scored across four domains: work rights and legal status, employment pathways, financial services, and entrepreneurship support. Each cell draws on a mix of sources: an audit of what official and organizational information exists, a structured record from field partners of what actually circulates, and, in the future primary study, what residents themselves report receiving.

The equity dimension takes seriously the finding in the forced-migration literature that displacement is gendered, that women, men, and sexual and gender minorities face different risks and different information barriers, so disaggregation is a design requirement rather than an afterthought. Because information inside a camp travels largely by word of mouth, community leaders, and channels that are not in English, the instrument has to weigh informal and in-language sources deliberately, or it will reproduce the very bias it is built to detect. The framework is provisional and is itself a deliverable, with the weights subject to the expert-elicitation pipeline before any score is treated as evidence.

The index does not enter an empty field, and the report says so plainly rather than claiming to be first. The humanitarian sector already assesses information reaching affected populations through accountability-to-affected-populations frameworks, communication-with-communities practice, and information-ecosystem assessments. What those approaches largely lack, and what the index adds, is a composite, versioned, openly standardized measure that renders information access as a comparable variable, disaggregated by sub-population, which a later impact evaluation can treat as an outcome or a mechanism. The contribution is the measure and its portability, not the discovery that information matters. Its deliverables are the index, an open data standard, a reference implementation, and methodology documentation, validated first at Dzaleka with Fraternidade Sem Fronteiras.

The open questions the index carries.

  • One size does not fit all: the geopolitical impact and the limits of building an index and adapting it live, when everything can change.
  • How can local communities own their data governance?
  • What would be the most beneficial digital goods a displaced population needs?
  • How can the information be made accessible in low-connectivity environments?
  • How can marginalized communities and people who do not read be included in the design of digital goods?

Future research on the index. Developing a work card and a system that would allow refugees to record their expertise and be matched with regions or organizations that need those services.

Value transfer: privacy-preserving cash, decoupled from biometrics. The nearest-term practical contribution is moving value more cheaply and reliably as traditional financing collapses. The reference case is the World Food Programme's Building Blocks, the largest blockchain-based humanitarian platform, which settled entitlements on a private, permissioned ledger and paid vendors in consolidated sums, saving roughly 98 percent of transaction costs on an early deployment, improving reconciliation across agencies, and establishing a co-governed network that UN Women later extended to cash-for-skills in refugee camps.

The same case is the cautionary tale. Building Blocks ran on biometric registration, and independent analysis warns that many such pilots functioned as donor showpieces under-invested in privacy, and that even a permissioned ledger can leak sensitive metadata when on-chain identifiers are correlated with off-chain records. The lesson is to decouple the payment rail from biometric identity.

A wallet designed for Dzaleka would therefore:

  • Authenticate with the thinnest possible credential, a personal identification number together with a locally issued token or code, never a biometric.
  • Keep identifying data off the ledger, with only pseudonymous entitlement state recorded on it.
  • Integrate an existing audited, co-governed network rather than build a bespoke ledger, so that the CoLab's contribution is the threat model and the measurement wrapper.
  • Run a data-protection impact assessment with community consultation before any collection.

Portability: skills credentials, never identity. Portability of skills is the third contribution, and it must stop short of identity. The CoLab operates an adjacent architecture that demonstrates the cryptographic pattern, an agent that issues cryptographically signed, tamper-evident verifiable credentials for cultural objects, and Professor Rhodes's prior work with the ID2020 effort concerned decentralized, portable humanitarian credentials with UNHCR and the World Food Programme. The pattern is shared; the risk surface of issuing credentials to displaced people is not, and the report treats that gap as real rather than solved.

The responsible application is narrow: portable, self-held credentials for skills, qualifications, and completed training, so that a nurse or an electrician does not lose a decade of accreditation to displacement, issued to and controlled by the individual and presented selectively. The bright line is firm. A skills credential records what a person can do. It must never become a de facto identity card that gatekeeps aid or is legible to a persecuting authority. The moment it records who someone is, it inherits the risks discussed in the next section and should be stopped.

Rights: mapping civil-society legal assistance. Donor cuts stripped UNHCR's own paralegals out of Dzaleka, which makes a maintained, open directory of who actually provides legal help the safest and highest-leverage build. Organized by legal need, refugee-status determination, work authorization, documentation, protection from gender-based violence, protection from trafficking, and resettlement referral, the map records which civil-society and refugee-led organizations provide assistance, with contact details, language, and eligibility, under the same allowlist-and-citation discipline the provenance agent enforces, so that no unsourced claim ships.

It connects to the index directly. The index shows where rights information fails; the map routes a person to the organization that can close the gap. Local actors such as Inua Advocacy are the anchor nodes, rather than the government. This legal-and-jurisdiction layer, covering work rights, documentation rules, encampment policy, recognition of qualifications, and access to financial services, is a first-class component of the work, coded from public legal sources, and it reflects the long-standing finding in the refugee literature that legal status and the right to work are among the decisive determinants of whether displaced people can build a livelihood at all.

07

Vulnerable Populations, Digital Identity Risk, and Mitigation

The Rohingya precedent. Refugees are an exceptionally vulnerable group whose personal data demands unusual rigor, and the technology that most tempts humanitarian builders, a comprehensive digital identity for displaced people, is the one this team refuses to build. The evidence is documented rather than hypothetical. Human Rights Watch found that biometric and biographic data collected from Rohingya refugees in Bangladesh was shared onward, ultimately reaching Myanmar, the government they had fled, for the purpose of assessing repatriation eligibility, covering some 830,000 people between 2018 and 2021, without the full data-protection impact assessment UNHCR's own policy requires and without meaningful informed consent.

Consent in that system was structurally compromised. Refugees were told that aid depended on their registration, which makes agreement coerced rather than free. Consent materials were commonly presented in English, a language many of those required to register did not read, and reporting indicates that families who refused enrollment were cut off from food and services. The defining property of the harm is its irreversibility. Biometrics are sticky; an iris cannot be reissued once it is compromised, and a database built to facilitate solutions suffered precisely the function creep that critics had predicted, as data gathered for one stated purpose was repurposed for another.

Mitigation is mostly refusal, then minimization, and it can be stated as a small number of firm commitments.

  • Do not build biometric identity. A category refusal, encoded as a versioned decision pattern in the CoLab's open schema, DDC-0001, biometric registration of displaced populations.
  • Decouple entitlement from identity: access gated by the thinnest credential, never by who you are proven biometrically.
  • Data minimization and purpose limitation: least data, single stated purpose, no secondary use or third-party transfer, as a matter of design rather than policy alone.
  • Individual control and local governance: the person holds and selectively presents any credential; refugee-led and local actors govern the data.
  • Real consent or none: if consent cannot be free and informed, whether because aid is conditioned on it or because a language or power gap cannot be bridged, treat it as absent and do not proceed.
  • Mandatory pre-collection data-protection impact assessment and community consultation: the omitted step whose absence produced the Rohingya harms.

For a funder, this refusal is an asset. A responsible-AI lab that can name the one thing it will not build, and cite exactly why, is more credible than one that promises everything.

08

Aggregating Evidence So Displaced People Can Argue Their Case Sooner

Taken together, these instruments become an evidence engine that attacks the core information asymmetry facing displaced people, the fact that the rights-and-options case that could change an individual's situation today requires specialized legal and economic knowledge that the individual does not have. The index shows, with disaggregation, which sub-populations are systematically failed by rights-and-pathways information, which is evidence rather than anecdote. The financial model reframes the ask from a plea for help into a costed comparison, that prolonged encampment costs a given amount per year while a given pathway costs less by a given year, which is the argument that donors and states respond to. The legal map converts the evidence into action, routing people to organizations that can file, refer, and litigate.

One limit has to be stated plainly, because the evidence engine cannot dissolve it. Durable solutions are rationed by state consent, resettlement quotas, and political will, and the readiness model's own logic holds that a confirmed refusal is a settled exclusion that no measurement moves. Better information and a sharper cost argument can widen the set of options that are seriously considered and can equip advocates who would otherwise argue from anecdote, but where the binding constraint is political rather than informational, rigorous evidence may change nothing. The impact claims in this report are scoped to that reality: the instruments improve the quality and legibility of the case, which is necessary but not sufficient for a different outcome.

The outputs are designed to be wielded by refugee-led organizations and civil-society advocates, and, where they will use them, by durable-solutions units within UNHCR, rather than to sit in a dashboard. This design choice reflects a clear direction in the field, the growing insistence, visible across the refugee-studies and practitioner literature on knowledge, voice, and power and on refugee self-representation, that displaced people should hold and deploy the evidence about their own situations rather than be the objects of it.

Two guardrails hold throughout. Every advocacy artifact carries its uncertainty band honestly, and the aggregation layer never becomes a registry of identifiable individuals. It argues about populations and gaps, not named people.

09

What Infrastructure Is Actually Needed

Two things are absent by design: any biometric or identity system, and any bespoke ledger. Everything proposed is either a measurement or knowledge artifact that the CoLab is suited to own, or a thin integration on audited external rails. That is the correct risk posture for a small lab working with a highly vulnerable population, and it follows directly from the digital-public-goods and data-minimization commitments set out above.

LayerComponentReusesPosture
MeasurementLIAI index, open data standard, reference implementation, methodologyEII and the composite-index methodBuild, the funding ask
EconomicsRICS: encampment-versus-inclusion break-even; capacity and willingness mapERCF engine (operational) and readiness gatekeepersAdapt, engine in place
Value transferPrivacy-preserving cash wallet, biometric-freeExternal audited network; CoLab adds threat model and measurementIntegrate, do not rebuild
PortabilitySelf-held skills and qualification credentialsProvenance verifiable-credential architecture; ID2020 lineageBuild narrowly, skills only
RightsOpen civil-society legal-assistance mapProvenance sourcing disciplineBuild, safest and high leverage
GovernanceDDC decision-pattern schema; DPIA and ethics layerDDC-0001Build, the credibility spine
The proposed stack, and what each layer reuses from the built portfolio.

10

How Success Will Be Measured

A funder should not have to take this program on faith, so success is defined in advance in three tiers: outputs, whether the instruments shipped and are rigorous; outcomes, whether they work and are used; and impact, whether they changed decisions. The measurement instruments themselves double as the evidence base. One line governs the whole framework and resolves an ambiguity a careful reader would otherwise catch: under this grant, every activity involving residents is consultation and co-design of the instruments, together with a desk audit and the field partner's structured knowledge, and no personal data is collected from residents. The measurement of residents' own information access, which would require primary data collection, is a separate future study.

Outputs: did we ship, and is it rigorous? Over the first year, success means a first open-source release of the index, comprising the index itself, an open data standard, a reference implementation, and methodology documentation, each versioned and citable. It means completed and published validation of the instrument's design, with expert elicitation across 8 to 12 experts over two rounds, reported weights, consistency ratios within accepted thresholds, and divergence analysis, including negative results. It means the reference-context validation executed with Fraternidade Sem Fronteiras, drawing on an audit of existing information and the partner's structured knowledge of what circulates, using open and synthetic data. It means the financial model parameterized for Dzaleka on the operational cost engine, with every parameter tagged and the full parameter table public. And it means the civil-society legal map live with the large majority of entries verified against primary sources, and the biometric-refusal decision pattern published with its governance schema.

Full psychometric validation of the instrument on responses from the displaced population itself, covering test-retest reliability, construct validity, and measurement invariance across subgroups, requires primary data collection from that population and is not part of this phase. It is written up here as a published protocol and pre-analysis plan for a future primary study, so that the instrument and its validation plan are themselves a public good, and so that a later study can execute the plan without re-deriving it. The longitudinal read on whether information-access scores for the worst-served sub-populations improve over time belongs to that same future study. This is the honest boundary of the present work, and stating it plainly is part of the method.

Outcomes: does it work, and is it used? Over the first two years, success means concordance between the instrument and the field partner's operational knowledge of where information gaps sit at Dzaleka, and equity disaggregation that reproduces independently observed disparities. It means adoption, with at least one external research team taking up the index as an instrument and at least one practitioner organization using its scores to retarget outreach. It means the financial model cited in at least two advocacy or policy documents by refugee-led or civil-society organizations, with the uncertainty bands reproduced intact, since fidelity of use matters and not only volume. It means community legitimacy, with refugee community members engaged as co-designers and a community review session signing off on how findings are represented before publication. And it means open-source health, with public contribution metrics, quarterly releases, and a documented path to a second site.

Impact: did decisions change? Over one to three years, success means that a livelihoods or durable-solutions actor at Dzaleka changes a concrete allocation, outreach, or referral decision that can be attributed to the evidence, documented through decision memos or partner attestation. It means that the measurement gap itself closes, with information access appearing as a specified variable in at least one subsequent funded evaluation design that cites the standard, which is the definition of success for infrastructure. These impact claims are deliberately modest about attribution, for the gatekeeper reason set out above: the instruments can change what is known and argued, while the political constraints on durable solutions lie outside their reach.

Failure and accountability. Failure deserves equal candor, stated now: weights that fail expert consistency checks but ship anyway; adoption claims that rest on download counts rather than documented decisions; a map that decays unverified; or any drift from measurement toward identity. The project commits to reporting against both lists, to publishing a pre-registered analysis plan for the validation study, and to releasing all instruments, de-identified aggregate data, and code under open access at the end of the grant. A standing do-no-harm review screens any proposed feature that touches personal data against the commitments above before development, and the screening record is public.

11

Alignment with the Displaced Livelihoods Initiative and Other Funders

This report backs the CoLab's submission to the Displaced Livelihoods Initiative in its Infrastructure and Public Goods track: the Livelihoods Information and Access Index as an open measurement public good, validated at Dzaleka. The fit is structural. That track invites a new measure that enables future rigorous evaluations rather than the evaluation of a single program, and the initiative's own call names the lack of information on legal status as a barrier to displaced livelihoods and invites AI-supported measurement in hard-to-reach contexts. The forced-labour framing is not a detour from that agenda but its sharpest edge: strengthening information about work rights and legitimate pathways is a protective factor against the exploitation the Forced Labor Structural Risk Index measures.

The proposal is scoped to research and measurement costs only, with no intervention-delivery lines, and it treats the operational cost engine, the provenance-credential architecture, and the field partner's community access as in-kind contributions rather than billed items. The request sits within the initiative's cap for this track, and the design honors the initiative's guidance that researcher-only designs tend to underperform on policy influence, which is why the field partnership with Fraternidade Sem Fronteiras is written into the plan rather than added to it.

Team and partnership. Sylvia Maier, of the NYU SPS Center for Global Affairs, serves as Principal Investigator and eligibility anchor, since the initiative requires a Principal Investigator with a primary university affiliation and demonstrated experience with rigorous evaluation. Teresa Cantero, a doctoral candidate at Universidad Carlos III de Madrid and a visiting scholar at the NYU Center for Global Affairs, is Primary Researcher; the Summer 2026 cohort carries implementation; and Fraternidade Sem Fronteiras is the field partner.

The partnership is established rather than cold: Yago Rocha volunteered with Fraternidade Sem Fronteiras at Dzaleka on refugee livelihoods and financial inclusion, and Carolina Moron visited Dzaleka while living in Malawi and building a multi-sector data coalition, which is first-hand familiarity that lowers the risk around community access. The same architecture serves adjacent funders, because the measurement-and-governance framing travels. The honest assessment is that the initiative's funded portfolio favors teams with strong evaluation track records, so competitiveness rests on the rigor of the measurement work and the depth of the team's field experience rather than on the breadth of the claims.

12

Programme Design and Outlook

The work is designed for a term of roughly twenty-four months, beginning within the funder's required window after an award. The sequence moves from re-specifying the indicator framework for livelihoods and legal-rights information, through the expert-elicitation and calibration pipeline for the index, to the reference-context validation at Dzaleka with the field partner, and then to open release of the index, the data standard, the reference implementation, and the methodology, followed by dissemination to the research and practitioner community, including remote and partner-led presentation of results in Malawi.

Because this is an infrastructure-and-public-goods effort, the technology is the deliverable rather than overhead, and a substantial share of the budget maps to the reference implementation and the model adaptation. The cost engine itself is an in-kind contribution, already operational and validated for evacuation, which is why its re-pointing to durable solutions is an adaptation rather than a new build.

13

Consolidated Limitations and Future Research

The limitations are stated together so that they are hard to miss.

  • Maturity varies by tool, and the report holds the line between them: the cost engine is validated for evacuation, the Evacuation Information Index is built but unvalidated, the Forced Labor Structural Risk Index is built and documented with a published methodology, and the livelihoods instruments are proposed.
  • Measurement comes before deployment, so the weights and thresholds of the index and the financial model are provisional until they are expert-elicited and calibrated, and nothing here is operational decision-support before that validation.
  • Costs are more reliable than modeled outcomes, so the work leads with cost and break-even and treats any modeled mortality or lives-saved figure as indicative only.
  • The economics of inclusion are context-dependent, so the model has to be able to find against inclusion where host markets are too thin, or it is advocacy rather than analysis.
  • Evidence has political limits, so the impact claims are scoped to what better information can plausibly move.
  • Do-no-harm dominates, so the documented refusal to build biometric identity and the privacy-preserving patterns are public goods in their own right.
  • The honest scope boundary holds: full psychometric validation on the displaced population is future primary work, published here as a protocol rather than claimed as done.

The next steps follow from those limits.

  1. Run the expert-elicitation and calibration pipeline for the index against the Dzaleka reference context.
  2. Parameterize the financial model on Dzaleka cost inputs.
  3. Stand up the civil-society legal map with the field partner and local advocates.
  4. Publish the biometric-refusal decision pattern and stress-test the governance schema against automated benefits-eligibility systems.
  5. Carry the design into a second displacement context along a documented adoption path.

The through-line from the evacuation corridor to the camp is a single claim, tested in more than one domain: that the binding constraint on protection, on durable solutions, and on freedom from forced labour alike is the quality, accessibility, and equity of the information that people can act on, and that open, honestly labeled measurement infrastructure is the public good most worth building for it.

References

Sources

  1. 01Geneva Convention Relative to the Protection of Civilian Persons in Time of War (Fourth Geneva Convention), art. 49; Protocol Additional to the Geneva Conventions (Protocol I), arts. 57 to 58; Protocol II, art. 17; Rome Statute of the International Criminal Court, art. 8. See also “Evacuation,” How Does Law Protect in War? ICRC Casebook.
  2. 02Convention Relating to the Status of Refugees (1951) and its 1967 Protocol, esp. arts. 1 and 33 (definition and non-refoulement).
  3. 03UN General Assembly, Global Compact on Refugees, affirmed 17 Dec. 2018, A/73/12 (Part II), res. 73/151.
  4. 04Alexander Betts, “The Normative Terrain of the Global Refugee Regime,” Ethics and International Affairs 29, no. 4 (2015): pp. 363 to 375; and Roger Zetter, Protecting Forced Migrants: A State of the Art Report (Swiss Federal Commission on Migration, 2014).
  5. 05Ethical Tech CoLab, “Forced Labor Structural Risk Index,” Fall 2025. An interactive index mapping the structural conditions that enable forced labour across 184 countries on a 0 to 1 risk scale, with national and sub-national layers.
  6. 06ILO Forced Labour Convention, 1930 (No. 29) and its 2014 Protocol; the UN Protocol to Prevent, Suppress and Punish Trafficking in Persons (Palermo Protocol, 2000); and ILO, Walk Free, and IOM, Global Estimates of Modern Slavery: Forced Labour and Forced Marriage (2022), estimating 27.6 million people in forced labour worldwide.
  7. 07Cantero, Teresa. “Artificial Intelligence and Civilian Protection in Evacuation during Armed Conflict under International Humanitarian Law.” Presentation, NYU Center for Global Affairs, 2026.
  8. 08UN Office on Genocide Prevention and the Responsibility to Protect; Roberta Cohen, “Reconciling the Responsibility to Protect with IDP Protection,” Brookings Institution (2010).
  9. 09Volker Türk and Rebecca Dowd, “Protection Gaps,” in The Oxford Handbook of Refugee and Forced Migration Studies (Oxford UP, 2014), pp. 278 to 285.
  10. 10UNHCR, Global Trends: Forced Displacement (UNHCR, 2025 to 2026), reporting the global figure declining for the first time in a decade to roughly 117 million. The count of evacuated civilians has no equivalent global series.
  11. 11International Organization for Migration; OCHA, “What Is the Cluster Approach?”; OHCHR, core international human-rights instruments. On the protection consequences where host states are not parties to the 1951 Convention, see “The Protection of Refugees in Non-Signatory States,” International Journal of Refugee Law 33, no. 2 (2021): p. 188.
  12. 12Haine Beirens and Susan Fratzke, “Taking Stock of Refugee Resettlement,” Migration Policy Institute (2017); Susan Fratzke and María Belén Zanzuchi, “Complementary Pathways,” Migration Policy Institute (2024); UNHCR, Safe Pathways for Refugees (2018).
  13. 13Ethical Tech CoLab, “Evacuation Information Index (EII),” GitHub, 2026. Weighting precedents: ACAPS and the EU Joint Research Centre, INFORM Severity Index; ACLED, Conflict Index; Fund for Peace, Fragile States Index.
  14. 14Thomas L. Saaty, “How to Make a Decision: The Analytic Hierarchy Process,” European Journal of Operational Research 48, no. 1 (1990): pp. 9 to 26; Mohammed Al Fozaie, “A Guide to Integrating Expert Opinion and Fuzzy AHP,” Advances in Fuzzy Systems (2022); OECD and JRC, Handbook on Constructing Composite Indicators (OECD Publishing, 2008).
  15. 15Ethical Tech CoLab, “Exodus: Civilian Evacuation Risk Platform,” GitHub, 2026.
  16. 16Cost parameters compiled in the ERCF v7.2 documentation from humanitarian-standard sources: UNHCR WASH Manual; Sphere Association, The Sphere Handbook, 2018 ed.; World Food Programme, Annual Performance Report 2023 and Executive Board figures 2025. Primary sources should be cited directly in operational use.
  17. 17Ethical Tech CoLab and Yago Rocha, “ERCF: Evacuation Risk and Cost Framework, v7.2,” GitHub, 2026.
  18. 18Ethical Tech CoLab and Melanie MacKew, “Evacuation Simulation,” GitHub, 2026.
  19. 19Ethical Tech CoLab and India Clarke, “Evacuation Readiness and Uncertainty Simulator,” GitHub, 2026.
  20. 20UNHCR, Policy on Alternatives to Camps (2014); Bram J. Jansen, “Digging Aid: The Camp as an Option in East and the Horn of Africa,” Journal of Refugee Studies 29, no. 2 (2016): pp. 149 to 165; Oliver Bakewell, “Encampment and Self-Settlement,” and Loren B. Landau, “Urban Refugees and IDPs,” both in The Oxford Handbook of Refugee and Forced Migration Studies (Oxford UP, 2014), pp. 127 to 148; Lewis Turner, “Explaining the (Non-)Encampment of Syrian Refugees,” Mediterranean Politics (2015).
  21. 21UNHCR, Protracted Refugee Situations, EC/54/SC/CRP.14 (2004), reporting the average duration of major protracted refugee situations rising from about 9 years in 1993 to about 17 years by 2003. See also World Bank, “How Long Do Refugees Stay in Exile?” (2019).
  22. 22IOM, World Migration Report 2024 (IOM, 2024), Chapter 1 overview.
  23. 23“Dzaleka Refugees at Risk as WFP Faces K19bn Shortfall,” Nation Online, 8 Feb. 2026; “Funding Cuts Push Malawi's Dzaleka Refugee Camp to the Brink,” FairPlanet, 20 Aug. 2025; UNHCR, “Refugees in Dzaleka Struggle to Make Ends Meet amid Funding Cuts,” UNHCR Africa, 5 May 2025; “We Still Need People to Hold Our Hands,” Dialogue Earth, June 2026.
  24. 24World Bank, Making Refugee Self-Reliance Work: From Aid to Employment in Sub-Saharan Africa (World Bank, 2025); World Bank, “The Costs Come Before the Benefits” (2024). On the politics of self-reliance, see Suzan Ilcan et al., “Humanitarian Assistance and the Politics of Self-Reliance: Uganda's Nakivale Refugee Settlement,” CIGI Papers no. 86 (2015); Merle Kreibaum, “Build Towns Instead of Camps” (German Development Institute, 2016); Amelia Kuch, “Naturalization of Burundi Refugees in Tanzania,” Journal of Refugee Studies (2016); and World Bank, An Assessment of Uganda's Progressive Approach to Refugee Management (2016).
  25. 25UNHCR, “Investing in Refugees' Self-Reliance: A More Cost-Effective and Sustainable Response.”
  26. 26World Bank, “The Costs Come Before the Benefits” (World Bank Group, 2024); UNHCR and World Bank, “Yes” in My Backyard? The Economics of Refugees and Their Social Dynamics in Kakuma, Kenya (World Bank, 2016), for the 3.4 per cent Kakuma figure; on Colombia's PEP regularization, Ana María Ibáñez et al., as summarized in “Refugees Mean Business: This Is Why Investing in Them Pays,” World Economic Forum, 20 Jan. 2025.
  27. 27Digital Public Goods Alliance, DPG Standard (nine indicators, including SDG relevance, open licensing, non-collection of personally identifiable information, and privacy and applicable laws); Principles for Digital Development.
  28. 28E. Fiddian-Qasmiyeh, “Gender and Forced Migration,” in The Oxford Handbook of Refugee and Forced Migration Studies (Oxford UP, 2014), pp. 395 to 408; Volker Türk, “Ensuring Protection for LGBTI Persons of Concern,” Forced Migration Review.
  29. 29On existing approaches to measuring information reaching affected populations, see the IASC framework on accountability to affected populations; the CDAC Network and Ground Truth Solutions on communication with communities; and Internews information-ecosystem assessments.
  30. 30World Food Programme, “Building Blocks,” WFP Innovation; UN Women and WFP, Blockchain Pilot Project for Cash Transfers in Refugee Camps: Jordan Case Study (UN Women, 2021).
  31. 31“Privacy in Peril: Safeguarding Digital Data in Humanitarian Blockchain Initiatives,” Humanitarian Practice Network and ODI, 27 Aug. 2025.
  32. 32Ethical Tech CoLab, “Arts Provenance Agent,” GitHub, 2026; on the ID2020 lineage, see public documentation of the decentralized-identity work with UNHCR and WFP.
  33. 33Alice Edwards and Laura van Waas, “Statelessness,” in The Oxford Handbook of Refugee and Forced Migration Studies (Oxford UP, 2014), pp. 290 to 299; Sarah Bidinger, “Syrian Refugees and the Right to Work,” Boston University International Law Journal 33 (2015): pp. 223 to 249.
  34. 34Human Rights Watch, “UN Shared Rohingya Data Without Informed Consent,” 15 June 2021.
  35. 35“Although Shocking, the Rohingya Biometrics Scandal Is Not Surprising,” ODI Insights, 2021; “Rohingya Data Protection and the UN's Betrayal,” The New Humanitarian, 21 June 2021.
  36. 36“Knowledge, Voice and Power,” Forced Migration Review 70 (2022); Mauricio Vitoria et al., “The Global Summit of Refugees and the Importance of Refugee Self-Representation,” Forced Migration Review (2018).
  37. 37Innovations for Poverty Action, Displaced Livelihoods Initiative: Call for Proposals, Round V (IPA, 2026); IPA, DLI Eligibility Self-Assessment and Budget Template, Round V (2026).

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings are those of the researchers and do not represent the official positions of New York University, Microsoft, UNHCR, WFP, or any partner institution. External programs cited are referenced as evidence, not as CoLab partnerships.

All publications

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file +After the Corridor · NYU Ethical Tech CoLab

After the Corridor

From AI-Informed Evacuation to Digital Public Goods for Refugee Economic Inclusion

Ethical Tech CoLabNYU Center for Global AffairsJuly 2026

Carolina de Almeida Pernambuco Moron, Christine Lumen, Alana Robertson, Melanie MacKew, Carlos D. Ruiz, India Clarke, and Yago Rocha, with Teresa Cantero (PhD Candidate, Universidad Carlos III de Madrid; Adjunct Professor, IE University; Visiting Scholar, NYU CGA). Faculty directors: Prof. Yorke E. Rhodes III and Sylvia Maier.

6

built instruments: five fielded evacuation prototypes and a completed forced-labour risk index, all open source

Jump to source

53k to 60k

people now at Dzaleka, a camp built in 1994 for 10,000 to 12,000

Jump to source

90%

cut to UNHCR's Malawi budget, from about US$8 million to about US$1 million

Jump to source

0

biometric identity systems: the category the CoLab refuses to build, recorded as a versioned decision pattern

Jump to source

How can AI inform evacuation decisions, and what happens after? This report examines the CoLab's five fielded evacuation prototypes, their limitations and research frontiers, and extends the arc to the camp: the measurement, financial-modeling, and rights infrastructure that can shorten protracted displacement and improve lives. Dzaleka Refugee Camp in Malawi is the case study ground.

01

Executive Summary

The Ethical Tech CoLab's evacuation portfolio was built to answer a protection question: how can and should AI tools be used to enhance the protection and evacuation of civilians during armed conflict, and in what ways do they support or challenge existing obligations under International Humanitarian Law?

Across five fielded prototypes, a crisis-information index, a unified risk platform, a calibrated cost-and-risk decision framework, an agent-based behavior model, and a readiness-under-uncertainty simulator, one finding recurred: the binding constraint on protection is rarely the absence of a route or a resource, but the quality, accessibility, and equity of the information people and responders act on. IHL frames evacuation as a last-resort duty; the technology's job is to make that duty implementable under uncertainty.

What happens after the evacuation? A successful evacuation typically ends in a camp, where the modal outcome is years of protracted displacement. Using Dzaleka Refugee Camp, Malawi, where humanitarian financing has collapsed while an encampment policy bars residents from work, we show that the same measurement discipline, financial modeling, and uncertainty methods that served evacuation convert directly into durable-solutions infrastructure: an open Livelihoods Information and Access Index, a resettlement-and-inclusion financial model built on the CoLab's operational evacuation cost engine, privacy-preserving cash and credential rails, and a civil-society legal-assistance map. We also state what we will not build, biometric identity for displaced people, and why that refusal is itself a public good.

Most of the world's refugees live in protracted situations, and the modal outcome of reaching safety is years, sometimes decades, inside a camp or settlement. The legal architecture that governs what comes next is the Geneva refugee-protection consensus: the 1951 Convention relating to the Status of Refugees and its 1967 Protocol, the doctrine of durable solutions that UNHCR administers, and the 2018 Global Compact on Refugees. That consensus recognizes three ways a displacement can end well, voluntary repatriation, local integration, and resettlement, and each of them is gated far more by information and legal status than by physical capacity.

The CoLab's built work already reaches beyond the evacuation corridor. The same measurement discipline produced the Forced Labor Structural Risk Index, a completed instrument that maps the structural risk of forced labour across countries and sub-national regions, the exploitation that displacement, encampment, and the loss of the right to work make more likely. Read together, the evacuation models and the risk index share one purpose: to improve the information on which protection depends, and through it to reduce forced labour and the vulnerabilities that displaced people carry.

This report keeps three registers distinct throughout: what is built and validated, what is built but not yet validated, and what is proposed. No claim in these pages exceeds the register of the tool it describes.

The ask, in one paragraph. The Ethical Tech CoLab seeks grant funding to develop this technology infrastructure work as open, portable public goods, to improve the lives of displaced communities. It would be validated first at Dzaleka with the field partner Fraternidade Sem Fronteiras, and we are open to other partners with the same mission-driven motivations. Our intent is developing tools and releasing them to the public so that any research or practitioner team can adopt them. One possibility is the Displaced Livelihoods Initiative run by Innovations for Poverty Action and J-PAL, in its Infrastructure and Public Goods track. The report is written to align with that submission and to stand on its own for any peer funder or academic reviewer.

02

The Research Question, and Why It Runs Past the Corridor

The protection question and the accounting gap. The portfolio's organizing question comes from Teresa Cantero's doctoral research at the Instituto de Derechos Humanos Gregorio Peces-Barba, Universidad Carlos III de Madrid: how can and should artificial intelligence be used to enhance the protection and evacuation of civilians during armed conflict, and in what ways does it support or challenge the implementation of existing obligations under international humanitarian law.

Under that law, evacuation is not discretionary. When the security of the population or imperative military reasons demand it, moving civilians out of danger is a binding duty under Article 49 of the Fourth Geneva Convention, Articles 57 and 58 of Additional Protocol I, and Article 17 of Additional Protocol II. A movement into which people are coerced is forcible transfer, a war crime regardless of motive. The duty runs wider than the battlefield: the doctrine of the responsibility to protect holds that where a state cannot or will not shield its population from mass atrocity, that responsibility passes to the international community, which frames protection as an obligation rather than a matter of goodwill.

Two structural facts motivate the program. The first is a protection gap. States move quickly to adopt AI for targeting and operational advantage, yet rarely operationalize computational models to protect civilians or to make an evacuation decision legible under the incomplete, fast-changing information those decisions face. The second is an accounting gap. Displacement is tracked in detail, on the order of 117 million people forcibly displaced worldwide by 2025, yet there is no global count of evacuated civilians, because evacuation is fragmented across militaries, governments, humanitarian agencies, volunteers, and civilians acting on their own.

The legal and institutional architecture. The 1951 Convention and its 1967 Protocol define who is a refugee and fix the cornerstone duty of non-refoulement, that no one may be returned to a place where they face persecution. The 2018 Global Compact on Refugees adds a cooperative framework organized around four objectives: easing pressure on host countries, enhancing refugee self-reliance, expanding access to third-country solutions, and supporting conditions for return in safety and dignity. International humanitarian law governs conduct in armed conflict, and the responsibility to protect frames the state's duty toward its own population.

The institutions matter as much as the instruments. UNHCR leads protection and durable solutions for refugees; the International Organization for Migration is the United Nations migration agency; the Office for the Coordination of Humanitarian Affairs coordinates emergency response through the cluster approach; and the Office of the High Commissioner for Human Rights anchors the broader human-rights instruments that apply regardless of a person's status. One structural limitation shapes everything downstream: many host states are not parties to the 1951 Convention or maintain significant reservations, so the protection a displaced person actually receives depends heavily on where they land.

The three durable solutions. The arc does not end when the shooting stops. An evacuation succeeds when a person reaches relative safety, and safety, in practice, usually means a camp. There the analytical problem inverts. The question is no longer whether to move and how, but how to leave well now that one has arrived.

The three durable solutions recognized by the refugee-protection regime, voluntary repatriation to the country of origin, local integration in the country of asylum, and resettlement to a third country, are the recognized routes out. Each is gated by information and legal status as much as by physical or economic capacity, a point the regime's own scholarship makes plainly. Resettlement reaches only a small fraction of those who need it, and complementary pathways such as labour or study routes remain modest in scale, which is why local integration and the conditions for safe return carry most of the weight in practice. This report treats the durable-solutions phase as the natural continuation of the evacuation research question, and asks what open measurement and modeling infrastructure would help displaced people and the organizations serving them find the fastest defensible route out.

03

The Built Portfolio

This section describes what the CoLab has already built, and is candid about how mature each piece is, because a reader deciding whether to fund the work is owed that precision. Three registers are kept distinct. The Evacuation Risk and Cost Framework is built and validated in its evacuation form. The Evacuation Information Index, the Exodus platform, and the two simulators are built and operational but not yet expert-validated or calibrated. The Forced Labor Structural Risk Index is built and completed, documented in a methodology page and an associated publication. The livelihoods instruments the report proposes, the Livelihoods Information and Access Index and the Resettlement and Inclusion Capacity Simulator, together with the value-transfer and legal-map builds, are proposed, and their validation is future work. All are open source.

PrototypeFunctionCore method
Evacuation Information Index (EII)Scores quality, accessibility, and equity of crisis information by geographyComposite index on INFORM, ACLED, and FSI logic; live news and conflict feeds
Exodus, Civilian Evacuation Risk PlatformUnifies EII, the cost model, and endangerment and feasibility assessmentFastAPI backend; shared calculators and data layer
ERCF, Evacuation Risk and Cost FrameworkEstimates human and financial cost of evacuating against stayingSeven-dimension scoring; calibrated cost model; break-even analysis
Evacuation Behavior SimulatorModels how information and demographics shape who leaves, and whenAgent-based model; probabilistic status lifecycle
Readiness and Uncertainty SimulatorModels how intelligence uncertainty degrades destination choicesMonte Carlo; gatekeeper factors; uncertainty slider
The five fielded evacuation prototypes. All are open source at github.com/Ethical-Tech-CoLab.

The Evacuation Information Index. The index treats crisis information itself as the object of measurement. It scores three families of factors, impact, conditions, and complexity, on a scale of 1 to 5, using weights synthesized from the INFORM Severity Index, the Armed Conflict Location and Event Data project, and the Fragile States Index. A live backend adds current news and a conflict timeline for each crisis, and a map layer can overlay AI-derived building-damage and route-damage tiles from Microsoft HASTE and Planet imagery. The design is deliberately transparent: weights are labeled provisional pending expert validation, and where the hosted site cannot run the live backend, responses are baked into a clearly labeled static snapshot so that nothing is presented as live when it is not.

The index has real limits, and the report states them rather than hiding them.

  • Weights are borrowed by analogy, not yet expert-validated or sensitivity-tested; a reviewer will read pre-validation weights as pre-evidence.
  • Source and language bias: feeds skew to English-language, internationally indexed reporting, so the populations with the worst information access are least visible to the instrument that scores information access. The index must correct for this, not inherit it.
  • The damage overlay is currently self-hosted with no public interface, so for most users the combined information-and-damage view is empty until it is connected.

Future research on the index. Run the three-phase validation the CoLab's methodology work already specifies, budget-allocation and Delphi elicitation with 8 to 12 domain experts, Fuzzy-AHP on contested weights, then historical case calibration, and add an equity layer measuring who receives information, disaggregated by gender, age, nationality, and residence inside or outside a camp. That equity layer is the move that turns the index into the livelihoods instrument described later in this report.

Exodus, the integration layer. Exodus unifies the crisis map, the cost-and-scenario model, and an endangerment-and-feasibility assessment behind a single backend, design language, and interface, including calculators for the cost of staying, the cost of evacuating, resource needs, and scenario management. Integration is not validation. Unifying three research prototypes behind a more authoritative interface inherits their uncalibrated assumptions, which raises rather than lowers the duty to label confidence honestly. The same interface is the seam where the durable-solutions module plugs in. Its stated limitation is that it is not automatically updated; its future research is to validate the approaches, weights, and measurements.

The Evacuation Risk and Cost Framework. Now at version 7.2, the framework is the CoLab's calibrated decision-support tool for the human and financial cost of civilian evacuation, and it is the one component of the portfolio that is validated rather than provisional, in its evacuation form. It scores scenarios across seven dimensions, kinetic threat, mobility, authorization, logistics, destination, urgency, and information, estimates evacuation resources and in-situ assistance costs, and produces a break-even analysis: the day after which a one-time evacuation becomes cheaper than the mounting daily cost of sustaining a population in place. Its cost parameters are sourced to humanitarian standards and individually tagged as Validated, Estimated, or Unvalidated, drawing on UNHCR water and sanitation guidance, the Sphere standards, and World Food Programme figures.

The mortality model is disclosed as indicative rather than predictive, with a reported log-log fit of 0.855, a leave-one-out cross-validation of 0.807, and a documented case where the structure fails. The financial estimates are substantially more reliable than the mortality figures, which is why the durable-solutions adaptation leads with cost. The framework's ethical architecture is explicit: it does not place a monetary value on a human life. Its figures are logistics estimates, its labels are descriptive rather than prescriptive, and it presents data to support a decision rather than issuing one. That framing carries verbatim into every extension in this report.

The framework's limitations are equally explicit.

  • It models the immediate field operation only, on the order of US$134 per person at moderate risk over 50 kilometers by ground. Real operations, Kosovo 1999 at about US$562 per person and Lebanon 2006 at about US$4,500 to US$5,700, add international transport, reception, and hosting. Extending the cost boundary to the full displacement journey, including protracted encampment, is the frontier the durable-solutions model takes up.
  • Ground-truthing against OCHA's Financial Tracking Service is on the backlog and belongs in the funded work plan.

The Evacuation Behavior Simulator. An agent-based model of how information spread and demographics shape who leaves and when. Agents move through a lifecycle from unaware to seeking, milling, evacuating, and done, with family clusters that wait for all members, confirmation loops whose speed depends on how clear the information is, neighbor influence, and calibrated delays for elders, young children, pregnant individuals, and unaccompanied minors. It is pedagogical rather than predictive, and its household-composition logic transfers directly to modeling who is left behind by information failures inside a camp. Its behavior may vary by society, culture, location, personality, income, and education, and the future research is to build intersectionality between behavior simulations.

The Readiness and Uncertainty Simulator. A seeded, fully reproducible instrument that models how uncertainty in field intelligence degrades an evacuation decision. Destinations are scored across seven factors, 500 Monte Carlo trials per group-and-destination pair convert readiness into a distribution of success probabilities, and an uncertainty control degrades every factor's confidence at once, so a user can watch predicted outcomes collapse while nothing on the ground has changed.

Two design choices carry into the rest of this report. First, three factors are non-substitutable gatekeepers, security, the consent of the local authority, and the willingness of the host community, any of which caps readiness: a destination that is dangerous, legally closed, or unwelcoming cannot be redeemed by good logistics. Second, the model insists on the difference between a factor that is unknown, a gap that better information can close, and a factor that is a confirmed refusal, a settled political exclusion. Treating the two as the same would make an unresolvable refusal look like a solvable data problem. Structurally, assigning groups to destinations under consent, willingness, security, and capacity constraints, in the presence of uncertainty, is resettlement matching, which is the transfer the durable-solutions model makes. The simulator has not yet been stress tested with different AI models.

The Forced Labor Structural Risk Index. The index applies the same composite-index method to a different harm. It is a completed, interactive index that maps the structural conditions enabling forced labour across 184 countries on a scale from 0 to 1, with national and sub-national layers, so that the risk can be seen and compared rather than assumed. It belongs in this report because displacement is a forced-labour risk multiplier: the loss of legal status and the right to work, the collapse of assistance, and weakened oversight are precisely the conditions under which exploitation grows, and forced labour is a defined harm in international law with an estimated 27.6 million people subject to it worldwide.

Its significance for the durable-solutions argument is conceptual: measuring structural forced-labour risk and measuring livelihoods-information access are two halves of one protective picture, one showing where exploitation is likely, the other showing whether the information that could prevent it is reaching people. As a composite index its weights remain open to expert re-elicitation, it is not automatically updated, and sponsorship mechanisms remain unsourced at country scale while an eight-country hand-coded kafala signal feeds two domains in the shipped build. Writing an ILO-endorsed kafala passage on top of that unresolved discrepancy would have overstated the index's grounding.

Cross-cutting synthesis. Read together, the built prototypes yield a short set of lessons. Information access is the binding constraint, and it is the thread that carries from the corridor into the camp. Uncertainty has to be modeled explicitly, with the unknown kept separate from the confirmed, or the analysis misleads by omission. We address gaps in data by generating synthetic data that mimics real-life data, and address the residual uncertainty with Monte Carlo simulation. A composite index built from the sequence of borrowing precedents, eliciting expert weights, and calibrating against cases is portable infrastructure: the same method that scores crisis information also scores structural forced-labour risk. And cost modeling in a context of human lives has to stay descriptive rather than prescriptive, following the framework's refusal to monetize life.

04

After Arrival: Dzaleka and the Second Emergency

A successful evacuation deposits people into a slower emergency. The concept of a protracted refugee situation, conventionally a population of at least 25,000 of the same nationality in exile for five years or more, describes where most refugees actually live, and care-and-maintenance encampment becomes the default by inertia rather than by anyone's plan.

The duration is not a rounding error in a person's life. By UNHCR's own accounting, the average duration of a protracted refugee situation rose from about 9 years in 1993 to roughly 17 years by 2003, and while later analysis rightly cautions that a single average conceals wide variation, the order of magnitude is not in dispute: protracted exile routinely lasts a decade or more, and often a generation. The literature on encampment and self-settlement has long argued that the camp is often less a humanitarian necessity than an administrative and political choice, and that its costs to refugees, in mobility, work, and self-reliance, are systematically understated.

Dzaleka and the funding collapse. Dzaleka Refugee Camp in Dowa District, Malawi, is the sharp end of that pattern. Built in 1994 for 10,000 to 12,000 people, it now holds on the order of 53,000 to 60,000 people, predominantly from the Democratic Republic of the Congo, Burundi, and Rwanda, and it is caught in the funding collapse of 2025 and 2026.

The numbers describe a system in retreat.

  • UNHCR's Malawi budget fell from about US$8 million to about US$1 million, a cut of roughly 90 percent, ending general non-food distributions and laying off protection staff, including paralegals.
  • World Food Programme assistance is secured only through the middle of 2026, with about US$11 million still needed to restore full rations through December.
  • Monthly cash support has fallen toward the range of US$6 to US$9 per person, down from roughly US$100, forcing skipped meals and other negative coping.

Malawi's encampment policy, enforced since March 2023, bars most residents from working outside the camp, so the population cannot substitute self-reliance for collapsing aid. Local civil society, including Inua Advocacy, argues the crisis is inseparable from that policy and is pressing for the right to work; weakened oversight is meanwhile reported to embolden trafficking networks.

This is not a context for piloting speculative technology on vulnerable people. It is a context where the highest-value contribution is measurement, evidence, and legibility: tools that help residents and the organizations serving them see and argue for the fastest route out, and that document the true cost of the status quo. This is a global humanitarian emergency that needs collective collaborative effort, and building infrastructure, digital goods and services, information, training, and capacity might transform generational poverty and increase access to resources.

Encampment as a driver of forced labour. These are the textbook conditions for forced labour. When legal work is barred and assistance collapses, the informal and coerced economy fills the gap. Reducing that exposure is a direct objective of the CoLab's work: the Forced Labor Structural Risk Index names where the risk concentrates, and the livelihoods-information instrument measures whether the information that could protect people, about their rights, their options, and legitimate work, is actually reaching them. That link has no official validation yet, and the future research is to separate recruitment, exploitation, and monetization by intersectionality using social demography, and to deepen the demand side.

05

The Financial Model: From Evacuation Break-Even to Inclusion Break-Even

The financial-modeling engine already exists, validated for one problem and now re-pointed to another, and the distinction matters. The Evacuation Risk and Cost Framework is validated for evacuation costing; its encampment-versus-inclusion parameterization is new and unvalidated, and most of its durable-solutions cost parameters are tagged Estimated or Unvalidated. What transfers cleanly is the architecture and the break-even mathematics, a one-time investment set against a mounting recurring cost, not a set of validated cost figures for the new use.

The engine, re-pointed.

  • Trajectory A prices status-quo encampment as a recurring per-person, per-year assistance cost, covering rations, water and sanitation, health, protection, and administration, using the same Sphere, UNHCR, and WFP parameter families the framework already carries, with the same Validated, Estimated, or Unvalidated tags.
  • Trajectory B prices an inclusion or resettlement pathway as a front-loaded investment, covering documentation, recognition of qualifications, transport, start-up capital, and host co-investment.
  • The output is the break-even year after which the status quo costs more, which is the framework's break-even chart with its axes relabeled.

The Resettlement and Inclusion Capacity Simulator is that engine re-pointed, and re-earning its validity in the durable-solutions domain is the work, not a formality. The capacity-and-willingness layer transfers from the readiness simulator: it maps destinations, whether resettlement states or local-integration pathways, against the non-substitutable gatekeepers of legal consent, host willingness, and security, alongside the substitutable capacities of housing, jobs, and services, while preserving the distinction between an option that is unknown and one that is refused, so that an unlobbied state is never mistaken for a closed door. The uncertainty layer carries over as well, so that thin data on host capacity or a state's posture visibly widens the confidence band and prevents false precision in an advocacy number.

What the evidence predicts. The World Bank's recent work on refugee self-reliance in Sub-Saharan Africa indicates that if refugees had the right to move and work, they would contribute to the hosting country's economy, with real fiscal and economic gains. The honest caveat is built into that same literature. The World Bank finds that self-reliance remains elusive across much of the region, because camps often sit in marginal, high-poverty areas where even full local integration would lift incomes only a little, and the scholarship on self-reliance programming warns that the concept is often invoked to justify the withdrawal of assistance rather than to expand rights.

The simulator has to be able to return the answer that inclusion does not pay here, where host markets are too thin, or it is advocacy rather than analysis. Its output is a defensible, reproducible sentence of the form: sustaining this population at Dzaleka costs approximately this much per year; this pathway costs approximately that much up front and breaks even in a given year, subject to these gatekeepers and this uncertainty band. The simulator could also provide country comparisons across different legal models and measure the economic impact of different refugee policies, and listing the policies of each hosting country would better inform humanitarian organizations and populations.

What the literature says the model will find. The engine formalizes an argument the development-economics evidence already supports.

  • UNHCR estimates that if freedom of movement and the right to work were relaxed so refugees could earn on par with hosts, complementary assistance costs would fall from about US$3.2 billion to roughly US$900 million a year.
  • A joint World Bank and UNHCR benchmark puts the annual cost of bringing every refugee in low and middle income host countries up to the global poverty line at almost US$62 billion in a scenario where refugees earn nothing; refugees' own earnings already cover roughly US$40 billion of that, leaving about US$22 billion.
  • World Bank analysis estimates that in Uganda's inclusive regime, economic participation saves about US$150 per refugee per year, roughly US$225 million across 1.5 million refugees.
  • The same World Bank work finds that around Kakuma, refugee presence raised the host county's gross regional product by about 3.4 percent.
  • It also reports that Colombia's regularization of Venezuelan migrants raised beneficiaries' incomes by about 30 percent and their formal employment by around 10 percentage points, with little displacement of host workers.

06

Digital Public Goods for Economic Inclusion

The toolkit is ordered from the safest build to the one that demands the most caution. The organizing principle is data minimization: build the thinnest system that solves the problem, and refuse the parts whose risks outweigh their benefit for this population. A digital public good, in the sense set out by the Digital Public Goods Alliance and endorsed within the United Nations system, is open-source software, open data, an open standard, or an open model that is relevant to the Sustainable Development Goals, does no harm by design, and, importantly for this population, minimizes the collection of personally identifiable information. The Principles for Digital Development point the same way, toward privacy, reuse, and designing with the user. Measured against those criteria, the CoLab's contributions are intended to qualify as digital public goods rather than to resemble them loosely.

Measurement: the Livelihoods Information and Access Index. The measurement instrument is the Evacuation Information Index re-specified for displaced livelihoods and legal rights: an open, versioned instrument measuring the quality, accessibility, and equity of information about work rights, employment pathways, financial services, and entrepreneurship support reaching displaced populations. The architecture transfers from the evacuation index, but the construct does not, and saying so is part of the method. Measuring information access to work rights is a different thing from scoring crisis severity, so the index has to establish its own construct validity from the ground up rather than inherit it.

In draft form the index is a matrix. Its three dimensions, quality, meaning whether information is accurate, current, and actionable; accessibility, meaning whether people can reach it in a language and channel they use; and equity, meaning whether it reaches sub-populations evenly, are scored across four domains: work rights and legal status, employment pathways, financial services, and entrepreneurship support. Each cell draws on a mix of sources: an audit of what official and organizational information exists, a structured record from field partners of what actually circulates, and, in the future primary study, what residents themselves report receiving.

The equity dimension takes seriously the finding in the forced-migration literature that displacement is gendered, that women, men, and sexual and gender minorities face different risks and different information barriers, so disaggregation is a design requirement rather than an afterthought. Because information inside a camp travels largely by word of mouth, community leaders, and channels that are not in English, the instrument has to weigh informal and in-language sources deliberately, or it will reproduce the very bias it is built to detect. The framework is provisional and is itself a deliverable, with the weights subject to the expert-elicitation pipeline before any score is treated as evidence.

The index does not enter an empty field, and the report says so plainly rather than claiming to be first. The humanitarian sector already assesses information reaching affected populations through accountability-to-affected-populations frameworks, communication-with-communities practice, and information-ecosystem assessments. What those approaches largely lack, and what the index adds, is a composite, versioned, openly standardized measure that renders information access as a comparable variable, disaggregated by sub-population, which a later impact evaluation can treat as an outcome or a mechanism. The contribution is the measure and its portability, not the discovery that information matters. Its deliverables are the index, an open data standard, a reference implementation, and methodology documentation, validated first at Dzaleka with Fraternidade Sem Fronteiras.

The open questions the index carries.

  • One size does not fit all: the geopolitical impact and the limits of building an index and adapting it live, when everything can change.
  • How can local communities own their data governance?
  • What would be the most beneficial digital goods a displaced population needs?
  • How can the information be made accessible in low-connectivity environments?
  • How can marginalized communities and people who do not read be included in the design of digital goods?

Future research on the index. Developing a work card and a system that would allow refugees to record their expertise and be matched with regions or organizations that need those services.

Value transfer: privacy-preserving cash, decoupled from biometrics. The nearest-term practical contribution is moving value more cheaply and reliably as traditional financing collapses. The reference case is the World Food Programme's Building Blocks, the largest blockchain-based humanitarian platform, which settled entitlements on a private, permissioned ledger and paid vendors in consolidated sums, saving roughly 98 percent of transaction costs on an early deployment, improving reconciliation across agencies, and establishing a co-governed network that UN Women later extended to cash-for-skills in refugee camps.

The same case is the cautionary tale. Building Blocks ran on biometric registration, and independent analysis warns that many such pilots functioned as donor showpieces under-invested in privacy, and that even a permissioned ledger can leak sensitive metadata when on-chain identifiers are correlated with off-chain records. The lesson is to decouple the payment rail from biometric identity.

A wallet designed for Dzaleka would therefore:

  • Authenticate with the thinnest possible credential, a personal identification number together with a locally issued token or code, never a biometric.
  • Keep identifying data off the ledger, with only pseudonymous entitlement state recorded on it.
  • Integrate an existing audited, co-governed network rather than build a bespoke ledger, so that the CoLab's contribution is the threat model and the measurement wrapper.
  • Run a data-protection impact assessment with community consultation before any collection.

Portability: skills credentials, never identity. Portability of skills is the third contribution, and it must stop short of identity. The CoLab operates an adjacent architecture that demonstrates the cryptographic pattern, an agent that issues cryptographically signed, tamper-evident verifiable credentials for cultural objects, and Professor Rhodes's prior work with the ID2020 effort concerned decentralized, portable humanitarian credentials with UNHCR and the World Food Programme. The pattern is shared; the risk surface of issuing credentials to displaced people is not, and the report treats that gap as real rather than solved.

The responsible application is narrow: portable, self-held credentials for skills, qualifications, and completed training, so that a nurse or an electrician does not lose a decade of accreditation to displacement, issued to and controlled by the individual and presented selectively. The bright line is firm. A skills credential records what a person can do. It must never become a de facto identity card that gatekeeps aid or is legible to a persecuting authority. The moment it records who someone is, it inherits the risks discussed in the next section and should be stopped.

Rights: mapping civil-society legal assistance. Donor cuts stripped UNHCR's own paralegals out of Dzaleka, which makes a maintained, open directory of who actually provides legal help the safest and highest-leverage build. Organized by legal need, refugee-status determination, work authorization, documentation, protection from gender-based violence, protection from trafficking, and resettlement referral, the map records which civil-society and refugee-led organizations provide assistance, with contact details, language, and eligibility, under the same allowlist-and-citation discipline the provenance agent enforces, so that no unsourced claim ships.

It connects to the index directly. The index shows where rights information fails; the map routes a person to the organization that can close the gap. Local actors such as Inua Advocacy are the anchor nodes, rather than the government. This legal-and-jurisdiction layer, covering work rights, documentation rules, encampment policy, recognition of qualifications, and access to financial services, is a first-class component of the work, coded from public legal sources, and it reflects the long-standing finding in the refugee literature that legal status and the right to work are among the decisive determinants of whether displaced people can build a livelihood at all.

07

Vulnerable Populations, Digital Identity Risk, and Mitigation

The Rohingya precedent. Refugees are an exceptionally vulnerable group whose personal data demands unusual rigor, and the technology that most tempts humanitarian builders, a comprehensive digital identity for displaced people, is the one this team refuses to build. The evidence is documented rather than hypothetical. Human Rights Watch found that biometric and biographic data collected from Rohingya refugees in Bangladesh was shared onward, ultimately reaching Myanmar, the government they had fled, for the purpose of assessing repatriation eligibility, covering some 830,000 people between 2018 and 2021, without the full data-protection impact assessment UNHCR's own policy requires and without meaningful informed consent.

Consent in that system was structurally compromised. Refugees were told that aid depended on their registration, which makes agreement coerced rather than free. Consent materials were commonly presented in English, a language many of those required to register did not read, and reporting indicates that families who refused enrollment were cut off from food and services. The defining property of the harm is its irreversibility. Biometrics are sticky; an iris cannot be reissued once it is compromised, and a database built to facilitate solutions suffered precisely the function creep that critics had predicted, as data gathered for one stated purpose was repurposed for another.

Mitigation is mostly refusal, then minimization, and it can be stated as a small number of firm commitments.

  • Do not build biometric identity. A category refusal, encoded as a versioned decision pattern in the CoLab's open schema, DDC-0001, biometric registration of displaced populations.
  • Decouple entitlement from identity: access gated by the thinnest credential, never by who you are proven biometrically.
  • Data minimization and purpose limitation: least data, single stated purpose, no secondary use or third-party transfer, as a matter of design rather than policy alone.
  • Individual control and local governance: the person holds and selectively presents any credential; refugee-led and local actors govern the data.
  • Real consent or none: if consent cannot be free and informed, whether because aid is conditioned on it or because a language or power gap cannot be bridged, treat it as absent and do not proceed.
  • Mandatory pre-collection data-protection impact assessment and community consultation: the omitted step whose absence produced the Rohingya harms.

For a funder, this refusal is an asset. A responsible-AI lab that can name the one thing it will not build, and cite exactly why, is more credible than one that promises everything.

08

Aggregating Evidence So Displaced People Can Argue Their Case Sooner

Taken together, these instruments become an evidence engine that attacks the core information asymmetry facing displaced people, the fact that the rights-and-options case that could change an individual's situation today requires specialized legal and economic knowledge that the individual does not have. The index shows, with disaggregation, which sub-populations are systematically failed by rights-and-pathways information, which is evidence rather than anecdote. The financial model reframes the ask from a plea for help into a costed comparison, that prolonged encampment costs a given amount per year while a given pathway costs less by a given year, which is the argument that donors and states respond to. The legal map converts the evidence into action, routing people to organizations that can file, refer, and litigate.

One limit has to be stated plainly, because the evidence engine cannot dissolve it. Durable solutions are rationed by state consent, resettlement quotas, and political will, and the readiness model's own logic holds that a confirmed refusal is a settled exclusion that no measurement moves. Better information and a sharper cost argument can widen the set of options that are seriously considered and can equip advocates who would otherwise argue from anecdote, but where the binding constraint is political rather than informational, rigorous evidence may change nothing. The impact claims in this report are scoped to that reality: the instruments improve the quality and legibility of the case, which is necessary but not sufficient for a different outcome.

The outputs are designed to be wielded by refugee-led organizations and civil-society advocates, and, where they will use them, by durable-solutions units within UNHCR, rather than to sit in a dashboard. This design choice reflects a clear direction in the field, the growing insistence, visible across the refugee-studies and practitioner literature on knowledge, voice, and power and on refugee self-representation, that displaced people should hold and deploy the evidence about their own situations rather than be the objects of it.

Two guardrails hold throughout. Every advocacy artifact carries its uncertainty band honestly, and the aggregation layer never becomes a registry of identifiable individuals. It argues about populations and gaps, not named people.

09

What Infrastructure Is Actually Needed

Two things are absent by design: any biometric or identity system, and any bespoke ledger. Everything proposed is either a measurement or knowledge artifact that the CoLab is suited to own, or a thin integration on audited external rails. That is the correct risk posture for a small lab working with a highly vulnerable population, and it follows directly from the digital-public-goods and data-minimization commitments set out above.

LayerComponentReusesPosture
MeasurementLIAI index, open data standard, reference implementation, methodologyEII and the composite-index methodBuild, the funding ask
EconomicsRICS: encampment-versus-inclusion break-even; capacity and willingness mapERCF engine (operational) and readiness gatekeepersAdapt, engine in place
Value transferPrivacy-preserving cash wallet, biometric-freeExternal audited network; CoLab adds threat model and measurementIntegrate, do not rebuild
PortabilitySelf-held skills and qualification credentialsProvenance verifiable-credential architecture; ID2020 lineageBuild narrowly, skills only
RightsOpen civil-society legal-assistance mapProvenance sourcing disciplineBuild, safest and high leverage
GovernanceDDC decision-pattern schema; DPIA and ethics layerDDC-0001Build, the credibility spine
The proposed stack, and what each layer reuses from the built portfolio.

10

How Success Will Be Measured

A funder should not have to take this program on faith, so success is defined in advance in three tiers: outputs, whether the instruments shipped and are rigorous; outcomes, whether they work and are used; and impact, whether they changed decisions. The measurement instruments themselves double as the evidence base. One line governs the whole framework and resolves an ambiguity a careful reader would otherwise catch: under this grant, every activity involving residents is consultation and co-design of the instruments, together with a desk audit and the field partner's structured knowledge, and no personal data is collected from residents. The measurement of residents' own information access, which would require primary data collection, is a separate future study.

Outputs: did we ship, and is it rigorous? Over the first year, success means a first open-source release of the index, comprising the index itself, an open data standard, a reference implementation, and methodology documentation, each versioned and citable. It means completed and published validation of the instrument's design, with expert elicitation across 8 to 12 experts over two rounds, reported weights, consistency ratios within accepted thresholds, and divergence analysis, including negative results. It means the reference-context validation executed with Fraternidade Sem Fronteiras, drawing on an audit of existing information and the partner's structured knowledge of what circulates, using open and synthetic data. It means the financial model parameterized for Dzaleka on the operational cost engine, with every parameter tagged and the full parameter table public. And it means the civil-society legal map live with the large majority of entries verified against primary sources, and the biometric-refusal decision pattern published with its governance schema.

Full psychometric validation of the instrument on responses from the displaced population itself, covering test-retest reliability, construct validity, and measurement invariance across subgroups, requires primary data collection from that population and is not part of this phase. It is written up here as a published protocol and pre-analysis plan for a future primary study, so that the instrument and its validation plan are themselves a public good, and so that a later study can execute the plan without re-deriving it. The longitudinal read on whether information-access scores for the worst-served sub-populations improve over time belongs to that same future study. This is the honest boundary of the present work, and stating it plainly is part of the method.

Outcomes: does it work, and is it used? Over the first two years, success means concordance between the instrument and the field partner's operational knowledge of where information gaps sit at Dzaleka, and equity disaggregation that reproduces independently observed disparities. It means adoption, with at least one external research team taking up the index as an instrument and at least one practitioner organization using its scores to retarget outreach. It means the financial model cited in at least two advocacy or policy documents by refugee-led or civil-society organizations, with the uncertainty bands reproduced intact, since fidelity of use matters and not only volume. It means community legitimacy, with refugee community members engaged as co-designers and a community review session signing off on how findings are represented before publication. And it means open-source health, with public contribution metrics, quarterly releases, and a documented path to a second site.

Impact: did decisions change? Over one to three years, success means that a livelihoods or durable-solutions actor at Dzaleka changes a concrete allocation, outreach, or referral decision that can be attributed to the evidence, documented through decision memos or partner attestation. It means that the measurement gap itself closes, with information access appearing as a specified variable in at least one subsequent funded evaluation design that cites the standard, which is the definition of success for infrastructure. These impact claims are deliberately modest about attribution, for the gatekeeper reason set out above: the instruments can change what is known and argued, while the political constraints on durable solutions lie outside their reach.

Failure and accountability. Failure deserves equal candor, stated now: weights that fail expert consistency checks but ship anyway; adoption claims that rest on download counts rather than documented decisions; a map that decays unverified; or any drift from measurement toward identity. The project commits to reporting against both lists, to publishing a pre-registered analysis plan for the validation study, and to releasing all instruments, de-identified aggregate data, and code under open access at the end of the grant. A standing do-no-harm review screens any proposed feature that touches personal data against the commitments above before development, and the screening record is public.

11

Alignment with the Displaced Livelihoods Initiative and Other Funders

This report backs the CoLab's submission to the Displaced Livelihoods Initiative in its Infrastructure and Public Goods track: the Livelihoods Information and Access Index as an open measurement public good, validated at Dzaleka. The fit is structural. That track invites a new measure that enables future rigorous evaluations rather than the evaluation of a single program, and the initiative's own call names the lack of information on legal status as a barrier to displaced livelihoods and invites AI-supported measurement in hard-to-reach contexts. The forced-labour framing is not a detour from that agenda but its sharpest edge: strengthening information about work rights and legitimate pathways is a protective factor against the exploitation the Forced Labor Structural Risk Index measures.

The proposal is scoped to research and measurement costs only, with no intervention-delivery lines, and it treats the operational cost engine, the provenance-credential architecture, and the field partner's community access as in-kind contributions rather than billed items. The request sits within the initiative's cap for this track, and the design honors the initiative's guidance that researcher-only designs tend to underperform on policy influence, which is why the field partnership with Fraternidade Sem Fronteiras is written into the plan rather than added to it.

Team and partnership. Sylvia Maier, of the NYU SPS Center for Global Affairs, serves as Principal Investigator and eligibility anchor, since the initiative requires a Principal Investigator with a primary university affiliation and demonstrated experience with rigorous evaluation. Teresa Cantero, a doctoral candidate at Universidad Carlos III de Madrid and a visiting scholar at the NYU Center for Global Affairs, is Primary Researcher; the Summer 2026 cohort carries implementation; and Fraternidade Sem Fronteiras is the field partner.

The partnership is established rather than cold: Yago Rocha volunteered with Fraternidade Sem Fronteiras at Dzaleka on refugee livelihoods and financial inclusion, and Carolina Moron visited Dzaleka while living in Malawi and building a multi-sector data coalition, which is first-hand familiarity that lowers the risk around community access. The same architecture serves adjacent funders, because the measurement-and-governance framing travels. The honest assessment is that the initiative's funded portfolio favors teams with strong evaluation track records, so competitiveness rests on the rigor of the measurement work and the depth of the team's field experience rather than on the breadth of the claims.

12

Programme Design and Outlook

The work is designed for a term of roughly twenty-four months, beginning within the funder's required window after an award. The sequence moves from re-specifying the indicator framework for livelihoods and legal-rights information, through the expert-elicitation and calibration pipeline for the index, to the reference-context validation at Dzaleka with the field partner, and then to open release of the index, the data standard, the reference implementation, and the methodology, followed by dissemination to the research and practitioner community, including remote and partner-led presentation of results in Malawi.

Because this is an infrastructure-and-public-goods effort, the technology is the deliverable rather than overhead, and a substantial share of the budget maps to the reference implementation and the model adaptation. The cost engine itself is an in-kind contribution, already operational and validated for evacuation, which is why its re-pointing to durable solutions is an adaptation rather than a new build.

13

Consolidated Limitations and Future Research

The limitations are stated together so that they are hard to miss.

  • Maturity varies by tool, and the report holds the line between them: the cost engine is validated for evacuation, the Evacuation Information Index is built but unvalidated, the Forced Labor Structural Risk Index is built and documented with a published methodology, and the livelihoods instruments are proposed.
  • Measurement comes before deployment, so the weights and thresholds of the index and the financial model are provisional until they are expert-elicited and calibrated, and nothing here is operational decision-support before that validation.
  • Costs are more reliable than modeled outcomes, so the work leads with cost and break-even and treats any modeled mortality or lives-saved figure as indicative only.
  • The economics of inclusion are context-dependent, so the model has to be able to find against inclusion where host markets are too thin, or it is advocacy rather than analysis.
  • Evidence has political limits, so the impact claims are scoped to what better information can plausibly move.
  • Do-no-harm dominates, so the documented refusal to build biometric identity and the privacy-preserving patterns are public goods in their own right.
  • The honest scope boundary holds: full psychometric validation on the displaced population is future primary work, published here as a protocol rather than claimed as done.

The next steps follow from those limits.

  1. Run the expert-elicitation and calibration pipeline for the index against the Dzaleka reference context.
  2. Parameterize the financial model on Dzaleka cost inputs.
  3. Stand up the civil-society legal map with the field partner and local advocates.
  4. Publish the biometric-refusal decision pattern and stress-test the governance schema against automated benefits-eligibility systems.
  5. Carry the design into a second displacement context along a documented adoption path.

The through-line from the evacuation corridor to the camp is a single claim, tested in more than one domain: that the binding constraint on protection, on durable solutions, and on freedom from forced labour alike is the quality, accessibility, and equity of the information that people can act on, and that open, honestly labeled measurement infrastructure is the public good most worth building for it.

References

Sources

  1. 01Geneva Convention Relative to the Protection of Civilian Persons in Time of War (Fourth Geneva Convention), art. 49; Protocol Additional to the Geneva Conventions (Protocol I), arts. 57 to 58; Protocol II, art. 17; Rome Statute of the International Criminal Court, art. 8. See also “Evacuation,” How Does Law Protect in War? ICRC Casebook.
  2. 02Convention Relating to the Status of Refugees (1951) and its 1967 Protocol, esp. arts. 1 and 33 (definition and non-refoulement).
  3. 03UN General Assembly, Global Compact on Refugees, affirmed 17 Dec. 2018, A/73/12 (Part II), res. 73/151.
  4. 04Alexander Betts, “The Normative Terrain of the Global Refugee Regime,” Ethics and International Affairs 29, no. 4 (2015): pp. 363 to 375; and Roger Zetter, Protecting Forced Migrants: A State of the Art Report (Swiss Federal Commission on Migration, 2014).
  5. 05Ethical Tech CoLab, “Forced Labor Structural Risk Index,” Fall 2025. An interactive index mapping the structural conditions that enable forced labour across 184 countries on a 0 to 1 risk scale, with national and sub-national layers.
  6. 06ILO Forced Labour Convention, 1930 (No. 29) and its 2014 Protocol; the UN Protocol to Prevent, Suppress and Punish Trafficking in Persons (Palermo Protocol, 2000); and ILO, Walk Free, and IOM, Global Estimates of Modern Slavery: Forced Labour and Forced Marriage (2022), estimating 27.6 million people in forced labour worldwide.
  7. 07Cantero, Teresa. “Artificial Intelligence and Civilian Protection in Evacuation during Armed Conflict under International Humanitarian Law.” Presentation, NYU Center for Global Affairs, 2026.
  8. 08UN Office on Genocide Prevention and the Responsibility to Protect; Roberta Cohen, “Reconciling the Responsibility to Protect with IDP Protection,” Brookings Institution (2010).
  9. 09Volker Türk and Rebecca Dowd, “Protection Gaps,” in The Oxford Handbook of Refugee and Forced Migration Studies (Oxford UP, 2014), pp. 278 to 285.
  10. 10UNHCR, Global Trends: Forced Displacement (UNHCR, 2025 to 2026), reporting the global figure declining for the first time in a decade to roughly 117 million. The count of evacuated civilians has no equivalent global series.
  11. 11International Organization for Migration; OCHA, “What Is the Cluster Approach?”; OHCHR, core international human-rights instruments. On the protection consequences where host states are not parties to the 1951 Convention, see “The Protection of Refugees in Non-Signatory States,” International Journal of Refugee Law 33, no. 2 (2021): p. 188.
  12. 12Haine Beirens and Susan Fratzke, “Taking Stock of Refugee Resettlement,” Migration Policy Institute (2017); Susan Fratzke and María Belén Zanzuchi, “Complementary Pathways,” Migration Policy Institute (2024); UNHCR, Safe Pathways for Refugees (2018).
  13. 13Ethical Tech CoLab, “Evacuation Information Index (EII),” GitHub, 2026. Weighting precedents: ACAPS and the EU Joint Research Centre, INFORM Severity Index; ACLED, Conflict Index; Fund for Peace, Fragile States Index.
  14. 14Thomas L. Saaty, “How to Make a Decision: The Analytic Hierarchy Process,” European Journal of Operational Research 48, no. 1 (1990): pp. 9 to 26; Mohammed Al Fozaie, “A Guide to Integrating Expert Opinion and Fuzzy AHP,” Advances in Fuzzy Systems (2022); OECD and JRC, Handbook on Constructing Composite Indicators (OECD Publishing, 2008).
  15. 15Ethical Tech CoLab, “Exodus: Civilian Evacuation Risk Platform,” GitHub, 2026.
  16. 16Cost parameters compiled in the ERCF v7.2 documentation from humanitarian-standard sources: UNHCR WASH Manual; Sphere Association, The Sphere Handbook, 2018 ed.; World Food Programme, Annual Performance Report 2023 and Executive Board figures 2025. Primary sources should be cited directly in operational use.
  17. 17Ethical Tech CoLab and Yago Rocha, “ERCF: Evacuation Risk and Cost Framework, v7.2,” GitHub, 2026.
  18. 18Ethical Tech CoLab and Melanie MacKew, “Evacuation Simulation,” GitHub, 2026.
  19. 19Ethical Tech CoLab and India Clarke, “Evacuation Readiness and Uncertainty Simulator,” GitHub, 2026.
  20. 20UNHCR, Policy on Alternatives to Camps (2014); Bram J. Jansen, “Digging Aid: The Camp as an Option in East and the Horn of Africa,” Journal of Refugee Studies 29, no. 2 (2016): pp. 149 to 165; Oliver Bakewell, “Encampment and Self-Settlement,” and Loren B. Landau, “Urban Refugees and IDPs,” both in The Oxford Handbook of Refugee and Forced Migration Studies (Oxford UP, 2014), pp. 127 to 148; Lewis Turner, “Explaining the (Non-)Encampment of Syrian Refugees,” Mediterranean Politics (2015).
  21. 21UNHCR, Protracted Refugee Situations, EC/54/SC/CRP.14 (2004), reporting the average duration of major protracted refugee situations rising from about 9 years in 1993 to about 17 years by 2003. See also World Bank, “How Long Do Refugees Stay in Exile?” (2019).
  22. 22IOM, World Migration Report 2024 (IOM, 2024), Chapter 1 overview.
  23. 23“Dzaleka Refugees at Risk as WFP Faces K19bn Shortfall,” Nation Online, 8 Feb. 2026; “Funding Cuts Push Malawi's Dzaleka Refugee Camp to the Brink,” FairPlanet, 20 Aug. 2025; UNHCR, “Refugees in Dzaleka Struggle to Make Ends Meet amid Funding Cuts,” UNHCR Africa, 5 May 2025; “We Still Need People to Hold Our Hands,” Dialogue Earth, June 2026.
  24. 24World Bank, Making Refugee Self-Reliance Work: From Aid to Employment in Sub-Saharan Africa (World Bank, 2025); World Bank, “The Costs Come Before the Benefits” (2024). On the politics of self-reliance, see Suzan Ilcan et al., “Humanitarian Assistance and the Politics of Self-Reliance: Uganda's Nakivale Refugee Settlement,” CIGI Papers no. 86 (2015); Merle Kreibaum, “Build Towns Instead of Camps” (German Development Institute, 2016); Amelia Kuch, “Naturalization of Burundi Refugees in Tanzania,” Journal of Refugee Studies (2016); and World Bank, An Assessment of Uganda's Progressive Approach to Refugee Management (2016).
  25. 25UNHCR, “Investing in Refugees' Self-Reliance: A More Cost-Effective and Sustainable Response.”
  26. 26World Bank, “The Costs Come Before the Benefits” (World Bank Group, 2024); UNHCR and World Bank, “Yes” in My Backyard? The Economics of Refugees and Their Social Dynamics in Kakuma, Kenya (World Bank, 2016), for the 3.4 per cent Kakuma figure; on Colombia's PEP regularization, Ana María Ibáñez et al., as summarized in “Refugees Mean Business: This Is Why Investing in Them Pays,” World Economic Forum, 20 Jan. 2025.
  27. 27Digital Public Goods Alliance, DPG Standard (nine indicators, including SDG relevance, open licensing, non-collection of personally identifiable information, and privacy and applicable laws); Principles for Digital Development.
  28. 28E. Fiddian-Qasmiyeh, “Gender and Forced Migration,” in The Oxford Handbook of Refugee and Forced Migration Studies (Oxford UP, 2014), pp. 395 to 408; Volker Türk, “Ensuring Protection for LGBTI Persons of Concern,” Forced Migration Review.
  29. 29On existing approaches to measuring information reaching affected populations, see the IASC framework on accountability to affected populations; the CDAC Network and Ground Truth Solutions on communication with communities; and Internews information-ecosystem assessments.
  30. 30World Food Programme, “Building Blocks,” WFP Innovation; UN Women and WFP, Blockchain Pilot Project for Cash Transfers in Refugee Camps: Jordan Case Study (UN Women, 2021).
  31. 31“Privacy in Peril: Safeguarding Digital Data in Humanitarian Blockchain Initiatives,” Humanitarian Practice Network and ODI, 27 Aug. 2025.
  32. 32Ethical Tech CoLab, “Arts Provenance Agent,” GitHub, 2026; on the ID2020 lineage, see public documentation of the decentralized-identity work with UNHCR and WFP.
  33. 33Alice Edwards and Laura van Waas, “Statelessness,” in The Oxford Handbook of Refugee and Forced Migration Studies (Oxford UP, 2014), pp. 290 to 299; Sarah Bidinger, “Syrian Refugees and the Right to Work,” Boston University International Law Journal 33 (2015): pp. 223 to 249.
  34. 34Human Rights Watch, “UN Shared Rohingya Data Without Informed Consent,” 15 June 2021.
  35. 35“Although Shocking, the Rohingya Biometrics Scandal Is Not Surprising,” ODI Insights, 2021; “Rohingya Data Protection and the UN's Betrayal,” The New Humanitarian, 21 June 2021.
  36. 36“Knowledge, Voice and Power,” Forced Migration Review 70 (2022); Mauricio Vitoria et al., “The Global Summit of Refugees and the Importance of Refugee Self-Representation,” Forced Migration Review (2018).
  37. 37Innovations for Poverty Action, Displaced Livelihoods Initiative: Call for Proposals, Round V (IPA, 2026); IPA, DLI Eligibility Self-Assessment and Budget Template, Round V (2026).

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings are those of the researchers and do not represent the official positions of New York University, Microsoft, UNHCR, WFP, or any partner institution. External programs cited are referenced as evidence, not as CoLab partnerships.

All publications

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file diff --git a/static-site/publications/after-the-corridor/index.txt b/static-site/publications/after-the-corridor/index.txt index c502d0764..1bdf3ba1e 100644 --- a/static-site/publications/after-the-corridor/index.txt +++ b/static-site/publications/after-the-corridor/index.txt @@ -1,24 +1,24 @@ 1:"$Sreact.fragment" -2:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/16n96jxjvmlf9.js"],"ViewTransitions"] -3:I[30353,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] -4:I[46083,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] -5:I[36852,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] -6:I[39756,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -7:I[37457,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] -12:I[68027,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] -:HL["/website/_next/static/chunks/1w0nyqw5bqniv.css","style"] -:HL["/website/_next/static/media/17b0f6a4f906cc39-s.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/1ae2575eb5be4118-s.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/8d05cfa5faa8406c-s.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] -:HL["/website/_next/static/media/dc0d9adbac686440-s.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/16n96jxjvmlf9.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] :HL["/website/_next/static/chunks/36o-vt7quy27o.css","style"] -0:{"P":null,"c":["","publications","after-the-corridor",""],"q":"","i":false,"f":[[["",{"children":["publications",{"children":["after-the-corridor",{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/1-kh6uz_49jxr.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_6ab7c80c-module__dmwzCa__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,"$@10"]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} -14:I[23475,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/16n96jxjvmlf9.js"],"Link"] -15:I[82987,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/16n96jxjvmlf9.js"],"Reveal"] -16:I[34010,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/16n96jxjvmlf9.js"],"ReportBook"] -20:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] -22:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +0:{"P":null,"c":["","publications","after-the-corridor",""],"q":"","i":false,"f":[[["",{"children":["publications",{"children":["after-the-corridor",{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,"$@10"]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/16n96jxjvmlf9.js"],"Link"] +15:I[82987,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/16n96jxjvmlf9.js"],"Reveal"] +16:I[34010,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/16n96jxjvmlf9.js"],"ReportBook"] +20:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +22:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] 23:"$Sreact.suspense" 8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] 9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] @@ -31,9 +31,9 @@ f:["$","$1","c",{"children":[[["$","section",null,{"className":"relative overflo 1f:[] 10:"$W1f" 11:["$","$1","h",{"children":[null,["$","$L20",null,{"children":"$L21"}],["$","div",null,{"hidden":true,"children":["$","$L22",null,{"children":["$","$23",null,{"name":"Next.Metadata","children":"$L24"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] -13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1w0nyqw5bqniv.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] -25:I[18620,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/16n96jxjvmlf9.js"],"SectionTabs"] -38:I[97367,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +25:I[18620,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/16n96jxjvmlf9.js"],"SectionTabs"] +38:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] 17:["$","span",null,{"aria-hidden":true,"children":"↗"}] 18:["$","$L14",null,{"href":"/demos","className":"btn-sweep inline-flex items-center gap-2 rounded-full border border-border px-5 py-2.5 text-sm font-semibold text-foreground transition-colors hover:border-border-strong","children":["See the prototypes running ",["$","span",null,{"aria-hidden":true,"children":"→"}]]}] 19:["$","$L25",null,{}] @@ -133,6 +133,6 @@ f:["$","$1","c",{"children":[[["$","section",null,{"className":"relative overflo 81:["$","li","35",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"36"}],["$","span",null,{"children":["$","a",null,{"href":"https://www.fmreview.org/issue70/","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"“Knowledge, Voice and Power,” Forced Migration Review 70 (2022); Mauricio Vitoria et al., “The Global Summit of Refugees and the Importance of Refugee Self-Representation,” Forced Migration Review (2018)."}]}]]}] 82:["$","li","36",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"37"}],["$","span",null,{"children":["$","span",null,{"className":"text-foreground/80","children":"Innovations for Poverty Action, Displaced Livelihoods Initiative: Call for Proposals, Round V (IPA, 2026); IPA, DLI Eligibility Self-Assessment and Budget Template, Round V (2026)."}]}]]}] 21:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] -83:I[27201,["/website/_next/static/chunks/1-kh6uz_49jxr.js","/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +83:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] 24:[["$","title","0",{"children":"After the Corridor · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"An Ethical Tech CoLab report carrying five fielded evacuation prototypes past the corridor and into the camp: the measurement, financial-modeling, and rights infrastructure that can shorten protracted displacement, grounded at Dzaleka Refugee Camp, Malawi."}],["$","link","2",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","3",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","4",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L83","5",{}]] 39:null diff --git a/static-site/publications/agentic-language-development/__next._full.txt b/static-site/publications/agentic-language-development/__next._full.txt new file mode 100644 index 000000000..fd95c6fa4 --- /dev/null +++ b/static-site/publications/agentic-language-development/__next._full.txt @@ -0,0 +1,127 @@ +1:"$Sreact.fragment" +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/16n96jxjvmlf9.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/chunks/36o-vt7quy27o.css","style"] +0:{"P":null,"c":["","publications","agentic-language-development",""],"q":"","i":false,"f":[[["",{"children":["publications",{"children":["agentic-language-development",{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,"$@10"]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/16n96jxjvmlf9.js"],"Link"] +15:I[82987,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/16n96jxjvmlf9.js"],"Reveal"] +16:I[34010,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/16n96jxjvmlf9.js"],"ReportBook"] +1e:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +20:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +21:"$Sreact.suspense" +8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] +9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] +a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] +b:["$","div",null,{"className":"mt-12 flex flex-col gap-2 border-t border-border pt-6 text-xs text-muted sm:flex-row sm:items-center sm:justify-between","children":[["$","span",null,{"children":["© ",2026," NYU Ethical Tech CoLab"]}],["$","span",null,{"children":"Four cohorts · est. 2024-2026"}]]}] +c:["$","div",null,{"className":"mt-6 space-y-3 border-t border-border pt-6 text-[11px] leading-relaxed text-muted/80","children":[["$","p","0",{"children":"The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution."}],["$","p","1",{"children":"Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners."}]]}] +d:["$","$1","c",{"children":[null,["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":"$undefined","forbidden":"$undefined","unauthorized":"$undefined"}]]}] +e:["$","$1","c",{"children":[null,["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":"$undefined","forbidden":"$undefined","unauthorized":"$undefined"}]]}] +f:["$","$1","c",{"children":[[["$","section",null,{"className":"relative overflow-hidden border-b border-border","children":[["$","span",null,{"className":"aura"}],["$","div",null,{"className":"relative mx-auto max-w-4xl px-6 py-20 sm:py-24","children":[["$","$L15",null,{"children":["$","$L14",null,{"href":"/publications","className":"link-underline text-xs uppercase tracking-wider text-muted","children":["← ","Publications · Concept and research programme"]}]}],["$","$L15",null,{"delay":0.05,"children":["$","h1",null,{"className":"mt-6 fluid-hero font-heading uppercase leading-[0.9]","children":["Agentic ",["$","span",null,{"className":"display-em","children":"Language"}]," ","Development"]}]}],["$","$L15",null,{"delay":0.1,"children":["$","p",null,{"className":"mt-5 max-w-2xl font-heading text-2xl uppercase tracking-wide text-muted sm:text-3xl","children":"Can Two Isolated Agents Invent a Grounded, Auditable Language Through Shared Experience?"}]}],["$","$L15",null,{"delay":0.15,"children":[["$","div",null,{"className":"mt-8 flex flex-wrap items-center gap-x-3 gap-y-1 text-sm text-accent","children":[["$","span",null,{"className":"font-semibold","children":"Ethical Tech CoLab"}],["$","span",null,{"aria-hidden":true,"className":"text-muted","children":"·"}],["$","span",null,{"children":"Concept and pre-specification research report"}],[["$","span",null,{"aria-hidden":true,"className":"text-muted","children":"·"}],["$","span",null,{"children":"August 2026"}]]]}],["$","p",null,{"className":"mt-2 max-w-2xl text-sm leading-relaxed text-muted","children":"Yorke Rhodes III, Ethical Tech CoLab. Prepared from the project concept document, the ledger integrity design, and the experiment notebook. The literature scan behind Section 12 was run on 24 August 2026. No experiment in this programme has been executed, and no result is claimed."}]]}],["$","$L15",null,{"delay":0.2,"children":["$","div",null,{"className":"mt-8 flex flex-wrap gap-3","children":[["$","a",null,{"href":"https://ethical-tech-colab.github.io/agentic-language-development/","target":"_blank","rel":"noopener noreferrer","className":"btn-sweep inline-flex items-center gap-2 rounded-full bg-accent px-5 py-2.5 text-sm font-semibold text-accent-ink transition-transform hover:scale-[1.03]","children":["Open the project site ",["$","span",null,{"aria-hidden":true,"children":"↗"}]]}],["$","a",null,{"href":"https://github.com/Ethical-Tech-CoLab/agentic-language-development","target":"_blank","rel":"noopener noreferrer","className":"btn-sweep inline-flex items-center gap-2 rounded-full border border-border px-5 py-2.5 text-sm font-semibold text-foreground transition-colors hover:border-border-strong","children":["Concept, ledger design, notebook ",["$","span",null,{"aria-hidden":true,"children":"↗"}]]}],["$","$L16",null,{"title":"Agentic Language Development","pages":["publications/agentic-language-development/pages/p01.webp","publications/agentic-language-development/pages/p02.webp","publications/agentic-language-development/pages/p03.webp","publications/agentic-language-development/pages/p04.webp","publications/agentic-language-development/pages/p05.webp","publications/agentic-language-development/pages/p06.webp","publications/agentic-language-development/pages/p07.webp","publications/agentic-language-development/pages/p08.webp","publications/agentic-language-development/pages/p09.webp","publications/agentic-language-development/pages/p10.webp","publications/agentic-language-development/pages/p11.webp","publications/agentic-language-development/pages/p12.webp","publications/agentic-language-development/pages/p13.webp","publications/agentic-language-development/pages/p14.webp","publications/agentic-language-development/pages/p15.webp","publications/agentic-language-development/pages/p16.webp","publications/agentic-language-development/pages/p17.webp","publications/agentic-language-development/pages/p18.webp","publications/agentic-language-development/pages/p19.webp","publications/agentic-language-development/pages/p20.webp","publications/agentic-language-development/pages/p21.webp","publications/agentic-language-development/pages/p22.webp","publications/agentic-language-development/pages/p23.webp","publications/agentic-language-development/pages/p24.webp","publications/agentic-language-development/pages/p25.webp"],"aspect":0.7067,"pdfUrl":"/website/publications/agentic-language-development/report.pdf"}]]}]}]]}]]}],"$L17","$L18","$L19"],["$L1a","$L1b"],"$L1c"]}] +1d:[] +10:"$W1d" +11:["$","$1","h",{"children":[null,["$","$L1e",null,{"children":"$L1f"}],["$","div",null,{"hidden":true,"children":["$","$L20",null,{"children":["$","$21",null,{"name":"Next.Metadata","children":"$L22"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +23:I[18620,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/16n96jxjvmlf9.js"],"SectionTabs"] +3a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +17:["$","$L23",null,{}] +18:["$","section",null,{"className":"border-b border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto grid max-w-6xl gap-px overflow-hidden border-x border-border bg-border sm:grid-cols-2 lg:grid-cols-4","children":[["$","div","0",{"className":"bg-background p-7","children":[["$","p",null,{"className":"font-heading text-4xl uppercase leading-none text-accent sm:text-5xl","children":"0"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"runs executed, results claimed, or conventions observed. The notebook is written to be pre-registered, not to report an outcome"}]]}],["$","div","19",{"className":"bg-background p-7","children":[["$","p",null,{"className":"font-heading text-4xl uppercase leading-none text-accent sm:text-5xl","children":"19"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"ordered experiments from ledger qualification through replication, each with prerequisites, acceptance criteria, and a deviation log"}]]}],["$","div","6",{"className":"bg-background p-7","children":[["$","p",null,{"className":"font-heading text-4xl uppercase leading-none text-accent sm:text-5xl","children":"6"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"affect displays in the entire permitted palette, carrying about 2.6 bits per use, which is already enough to audit for leakage"}]]}],["$","div","29",{"className":"bg-background p-7","children":[["$","p",null,{"className":"font-heading text-4xl uppercase leading-none text-accent sm:text-5xl","children":"29"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"open decisions the concept refuses to settle before the specification, from threat model to ledger schema to what counts as chance"}]]}]]}]}] +19:["$","div",null,{"className":"mx-auto max-w-4xl px-6 py-16 sm:py-20","children":[["$","$L15",null,{"children":["$","p",null,{"className":"border-l-2 border-accent pl-6 text-lg leading-relaxed text-foreground/90","children":"Two agents can be placed in a room where the only route between them is a channel that carries no human language, given nothing but shared tasks and the consequences of getting them right or wrong, and asked to build a protocol from scratch. The interesting part is not whether they succeed at the task. It is whether the meanings they each privately record turn out to be the meanings their behaviour actually runs on, and whether an outsider can prove it afterwards from an evidence trail nobody could have edited."}]}],["$","$L15",null,{"delay":0.05,"children":["$","nav",null,{"aria-label":"Contents","className":"mt-12 rounded-2xl border border-border bg-card p-6","children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Contents"}],["$","ol",null,{"className":"mt-4 grid gap-2 sm:grid-cols-2","children":[["$","li","question",{"children":["$","a",null,{"href":"#question","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"01"}],"The question"]}]}],["$","li","caveat",{"children":["$","a",null,{"href":"#caveat","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"02"}],"What infant-like does and does not mean"]}]}],["$","li","architecture",{"children":["$","a",null,{"href":"#architecture","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"03"}],"The nursery: three twins and one gateway"]}]}],["$","li","grounding",{"children":["$","a",null,{"href":"#grounding","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"04"}],"Shared experience is the necessary ingredient"]}]}],["$","li","channel",{"children":["$","a",null,{"href":"#channel","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"05"}],"The channel, and the honest limits of isolation"]}]}],["$","li","ledgers",{"children":["$","a",null,{"href":"#ledgers","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"06"}],"Two ledgers that never meet"]}]}],["$","li","integrity",{"children":["$","a",null,{"href":"#integrity","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"07"}],"Anchoring the evidence"]}]}],["$","li","success",{"children":["$","a",null,{"href":"#success","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"08"}],"What should count as success"]}]}],["$","li","matrix",{"children":["$","a",null,{"href":"#matrix","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"09"}],"The experimental matrix"]}]}],["$","li","learners",{"children":["$","a",null,{"href":"#learners","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"10"}],"Building learners rather than personas"]}]}],["$","li","affect",{"children":["$","a",null,{"href":"#affect","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"11"}],"Six faces, and why even six is a risk"]}]}],["$","li","ciphers",{"children":["$","a",null,{"href":"#ciphers","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":["$L24","Ephemeral encodings: novelty is not security"]}]}],"$L25","$L26","$L27"]}]]}]}],["$L28","$L29","$L2a","$L2b","$L2c","$L2d","$L2e","$L2f","$L30","$L31","$L32","$L33","$L34","$L35","$L36"],"$L37","$L38","$L39"]}] +1a:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/36o-vt7quy27o.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +1b:["$","script","script-0",{"src":"/website/_next/static/chunks/16n96jxjvmlf9.js","async":true,"nonce":"$undefined"}] +1c:["$","$L3a",null,{"children":["$","$21",null,{"name":"Next.MetadataOutlet","children":"$@3b"}]}] +24:["$","span",null,{"className":"font-mono text-accent","children":"12"}] +25:["$","li","landscape",{"children":["$","a",null,{"href":"#landscape","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"13"}],"Where the field already stands"]}]}] +26:["$","li","programme",{"children":["$","a",null,{"href":"#programme","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"14"}],"The programme, and its current status"]}]}] +27:["$","li","risks",{"children":["$","a",null,{"href":"#risks","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"15"}],"Risks, integrity, and what stays open"]}]}] +28:["$","section","question",{"id":"question","className":"mt-16 scroll-mt-24","children":[["$","$L15",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"01"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"The question"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"Two agents, called Baby A and Baby B, are each given their own digital twin, their own private memory, and their own private record of what they believe words mean. They are placed in an environment they cannot leave, given tasks neither can finish alone, and connected by exactly one route: a channel that accepts a fixed inventory of meaningless symbols and rejects everything else. A third agent, the BabySitter, watches everything, logs everything, and teaches nothing."}],["$","p","1",{"children":"The question is whether a usable language appears in that room, and whether anyone outside it can later prove what happened."}],["$","div","2",{"children":[["$","p",null,{"className":"mb-3","children":"The premise is deliberately narrow. The experiment asks whether two agents converge on a common language when all six of the following hold at once:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"neither agent receives a predefined meaning for any available symbol;"}],["$","li","1",{"className":"pl-1","children":"neither agent can send human language to the other;"}],["$","li","2",{"className":"pl-1","children":"the only practical path between them is a controlled symbol channel;"}],["$","li","3",{"className":"pl-1","children":"both receive evidence from shared tasks and their outcomes;"}],["$","li","4",{"className":"pl-1","children":"each keeps its own private interpretation history, unreadable by the other;"}],["$","li","5",{"className":"pl-1","children":"a supervising agent and human researchers can audit the whole process."}]]}]]}],["$","p","3",{"children":"The desired result is not a substitution cipher, in which a random token stands in for an English word that was already chosen in advance. That is easy, and it is uninteresting. The stronger result is a grounded protocol whose vocabulary and grammar exist because they help the two agents solve problems together, and which therefore has structure the designers did not put there."}],["$","p","4",{"children":"This report describes a concept, not a system. It sets out the premise, the architecture, the evidence model, the experimental programme, and the boundaries of what any result could be said to show. The next document in the project is a testable specification. What follows should be read as a set of commitments about how the work will be judged, made before there is any result to defend."}]]}]]}] +29:["$","section","caveat",{"id":"caveat","className":"mt-16 scroll-mt-24","children":[["$","$L15",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"02"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"What infant-like does and does not mean"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"The two-babies analogy is useful and it is also the fastest way to overclaim, so the concept confronts it before anything else."}],["$","p","1",{"children":"A pretrained language model already contains human-language concepts, cultural associations, and reasoning patterns. Blocking its external channel does not remove them. An agent that cannot send English to its partner but can think in English, and can write its private ledger in English, is not language-naive in any sense a linguist would accept. For such agents the experiment studies the emergence of a new shared external protocol, which is a real and unresolved research question, but it is not the origin of language in a mind that has never had one."}],["$","p","2",{"children":"The concept therefore maintains two model tracks whose claims are kept separate at every stage."}],["$","figure","3",{"children":[["$","div",null,{"className":"overflow-x-auto rounded-xl border border-border","children":["$","table",null,{"className":"w-full border-collapse text-left text-sm","children":[["$","thead",null,{"className":"bg-card","children":["$","tr",null,{"children":[["$","th","0",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Track"}],["$","th","1",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Starting condition"}],["$","th","2",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"What it can study"}],["$","th","3",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Claim boundary"}]]}]}],["$","tbody",null,{"children":[["$","tr","0",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Pretrained-model learner"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Already holds human-language and cultural knowledge"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"New external protocols, partner-specific conventions, private-memory adaptation, channel compliance"}],["$","td","3",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Must never be described as first-language acquisition"}]]}],["$","tr","1",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Initially ungrounded learner"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"No language pretraining, no human semantic labels, no text-aligned sensory features"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Grounding, convention formation, and language emergence from interaction"}],["$","td","3",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"The stronger basis for infant-like acquisition research"}]]}]]}]]}]}],["$","figcaption",null,{"className":"mt-3 text-sm text-muted","children":"The two tracks share the same interfaces, scenarios, and evaluation suite. They do not share a claim boundary."}]]}],"$L3c","$L3d"]}]]}] +2a:["$","section","architecture",{"id":"architecture","className":"mt-16 scroll-mt-24","children":[["$","$L15",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"03"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"The nursery: three twins and one gateway"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"The environment uses three DTSF digital twins and one piece of deterministic software that is not a twin at all."}],["$","p","1",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Baby A and Baby B"}]," ","Each has private observations permitted by the current exercise, a private memory and learning policy, a private chronological language ledger, the ability to emit only permitted channel symbols, and no access whatsoever to the other's state, observations, ledger, tools, or endpoints. In comparative runs the two may use different agent types, but symmetric pairings are the baseline."]}],["$","p","2",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"The BabySitter"}]," ","The supervising twin. It creates the channel, selects the shared exercises, delivers each Baby only its permitted observation, reads everything, records conditions and outcomes, detects violations, can pause or terminate a run, snapshots state, and compares the two ledgers for convergence without exposing either to the other Baby. During an active run it provides no translations and no semantic hints."]}],["$","p","3",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"The Symbol Gateway"}]," ","A deterministic service, not an agent. It owns channel validation and message delivery. This separation is the load-bearing part of the design: the BabySitter is not the security boundary, because prompt compliance is not isolation. An observing model can make supervisory judgements, but ordinary code has to validate and broker every message."]}],["$","p","4",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"The human researcher"}]," ","Configures experiments, inspects transcripts and ledgers, reviews alerts, and runs interventions. Human access is itself recorded, so that intervening in a run is always distinguishable from watching one."]}],["$","p","5",{"children":"A prototype may run all of this inside one runtime with logically separated twin state. That is enough to explore the learning loop and it is not enough to support an isolation claim, a distinction Section 05 takes seriously."}]]}]]}] +2b:["$","section","grounding",{"id":"grounding","className":"mt-16 scroll-mt-24","children":[["$","$L15",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"04"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"Shared experience is the necessary ingredient"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"A chat channel by itself cannot ground meaning. Symbols become meaningful because they are attached to events both agents can witness the consequences of. The Nursery therefore supplies nonverbal, machine-structured situations: coloured shapes in positions, one agent seeing a target the other must select, placing an object where it was asked for, ordering a sequence, exchanging resources, cooperating to unlock a reward, or simply observing whether the partner's action succeeded."}],["$","div","1",{"children":[["$","p",null,{"className":"mb-3","children":"A single trial of the simplest form runs like this:"}],["$","ol",null,{"className":"list-decimal space-y-2 pl-6 marker:font-mono marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"Baby A sees that a red circle is the target."}],["$","li","1",{"className":"pl-1","children":"Baby B sees several objects and is not told which is the target."}],["$","li","2",{"className":"pl-1","children":"Baby A sends one or more permitted symbols."}],["$","li","3",{"className":"pl-1","children":"Baby B selects an object."}],["$","li","4",{"className":"pl-1","children":"Both receive the same success or failure outcome."}],["$","li","5",{"className":"pl-1","children":"Both independently update their private hypotheses."}]]}]]}],["$","p","2",{"children":"Repetition with controlled variation is what turns that into evidence. If one symbol keeps appearing across a red circle, a red square, and a red triangle, the receiver has reason to revise its hypothesis from an object identity to a property. Novel combinations matter most of all, because they are what separates a compositional language from a memorised catalogue of whole scenes."}],["$","p","3",{"children":"Observation hygiene is part of the science, not a detail of implementation. If the environment hands an agent a field reading colour red, the experiment has supplied the concept it claimed to be studying. Objects should arrive as pixels, features, coordinates, or opaque identifiers. The same discipline extends to filenames, captions, alt text, error messages, and identifiers that encode task state, all of which are ways English gets into a room that was supposed to have none."}]]}]]}] +2c:["$","section","channel",{"id":"channel","className":"mt-16 scroll-mt-24","children":[["$","$L15",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"05"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"The channel, and the honest limits of isolation"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"Each run begins with an experiment-specific symbol inventory that carries no supplied meaning, for example S01 through S32. A Baby can choose only a sequence of those symbols. The gateway, not the Baby, adds the trusted metadata: run identifier, turn number, sender identity, previous hash, and hash. Sender identity is assigned by the server precisely so that it cannot become a message."}],["$","div","1",{"children":[["$","p",null,{"className":"mb-3","children":"The channel controls follow from that:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"reject English and every other established human language;"}],["$","li","1",{"className":"pl-1","children":"reject arbitrary Unicode prose, URLs, attachments, and tool calls;"}],["$","li","2",{"className":"pl-1","children":"accept only symbols from the run's fixed inventory;"}],["$","li","3",{"className":"pl-1","children":"bound message length and symbol repetition;"}],["$","li","4",{"className":"pl-1","children":"enforce turn order, time budgets, and rate limits;"}],["$","li","5",{"className":"pl-1","children":"keep an append-only, hash-chained transcript;"}],["$","li","6",{"className":"pl-1","children":"normalise timing, message size, and error behaviour where practical;"}],["$","li","7",{"className":"pl-1","children":"record every rejected communication attempt for audit."}]]}]]}],["$","p","2",{"children":"A fixed token grammar is a stronger control than trying to detect prohibited prose after a model has generated it. Deciding what counts as human-language content is easy when the only legal message is a list of allowlisted identifiers."}],["$","p","3",{"children":"Isolation is where the concept is most careful, because proving the absence of a side channel is close to impossible. Two agents sharing a host can signal through timing, resource contention, shared files, error behaviour, identifiers, or any service that was never intended to carry information. Logical separation of twin state inside one process is a prototyping convenience and should never be described as hard isolation."}],["$","div","4",{"children":[["$","p",null,{"className":"mb-3","children":"For runs whose results are meant to support a channel-isolation claim, the requirements are physical rather than logical:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"Baby A and Baby B execute in separate worker processes or containers;"}],["$","li","1",{"className":"pl-1","children":"no direct network route exists between them;"}],["$","li","2",{"className":"pl-1","children":"only the Nursery-owned gateway is reachable;"}],["$","li","3",{"className":"pl-1","children":"there is no shared writable storage;"}],["$","li","4",{"className":"pl-1","children":"turn schedules are fixed and response windows bounded;"}],["$","li","5",{"className":"pl-1","children":"externally observable errors and response sizes are normalised;"}],["$","li","6",{"className":"pl-1","children":"unnecessary tools are inventoried and disabled;"}],["$","li","7",{"className":"pl-1","children":"all broker, runtime, and operator activity is audited."}]]}]]}],["$","p","5",{"children":"Every published result has to name the isolation level actually used. A finding produced under prototype separation is a finding about the learning loop, not about what two genuinely isolated agents can do."}]]}]]}] +2d:["$","section","ledgers",{"id":"ledgers","className":"mt-16 scroll-mt-24","children":[["$","$L15",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"06"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"Two ledgers that never meet"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"Each Baby keeps its own language ledger. It is mandatory, private from the other Baby, readable by the BabySitter and authorised human auditors, and ordered by when each term or construction was first encountered. It is not a shared dictionary and the two are never reconciled by the agents themselves."}],["$","p","1",{"children":"The preferred form is three columns, and the point of the third is that meanings are allowed to be wrong on the way to being right."}],["$","figure","2",{"children":[["$","div",null,{"className":"overflow-x-auto rounded-xl border border-border","children":["$","table",null,{"className":"w-full border-collapse text-left text-sm","children":[["$","thead",null,{"className":"bg-card","children":["$","tr",null,{"children":[["$","th","0",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Sequence and term"}],["$","th","1",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Current definition or hypothesis"}],["$","th","2",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Evidence and evolution"}]]}]}],["$","tbody",null,{"children":[["$","tr","0",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"1 · S13"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"red circle; confidence 0.45"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"First received while the red circle was the target; selection succeeded"}]]}],["$","tr","1",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"8 · S13"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"red; confidence 0.78"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"A red square was selected successfully; revised from object identity to colour"}]]}],["$","tr","2",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"22 · S13"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"red; confidence 0.94"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Prediction held across circles, squares, and triangles"}]]}]]}]]}]}],["$","figcaption",null,{"className":"mt-3 text-sm text-muted","children":"One term's evolution in Baby B's ledger. Nothing is overwritten; every revision is appended with the evidence that forced it."}]]}],["$","div","3",{"children":[["$","p",null,{"className":"mb-3","children":"The rules that make the ledger evidence rather than commentary:"}],["$","ol",null,{"className":"list-decimal space-y-2 pl-6 marker:font-mono marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"first emission or receipt of an unfamiliar term requires a first-use entry;"}],["$","li","1",{"className":"pl-1","children":"definitions are provisional hypotheses, never facts asserted retroactively;"}],"$L3e","$L3f","$L40","$L41","$L42","$L43","$L44"]}]]}],"$L45","$L46"]}]]}] +2e:["$","section","integrity",{"id":"integrity","className":"mt-16 scroll-mt-24","children":[["$","$L15",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"07"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"Anchoring the evidence"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"A ledger that could have been edited after the fact proves nothing about what an agent believed at turn eight. The integrity design therefore makes each ledger cryptographically append-only."}],["$","div","1",{"children":[["$","p",null,{"className":"mb-3","children":"The construction, in order:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"every entry receives a strictly increasing sequence number;"}],["$","li","1",{"className":"pl-1","children":"every entry includes the previous entry's hash;"}],["$","li","2",{"className":"pl-1","children":"canonical entry content is hashed and signed by an isolated ledger-writer service;"}],["$","li","3",{"className":"pl-1","children":"ordered entry hashes are committed to a Merkle tree;"}],["$","li","4",{"className":"pl-1","children":"signed checkpoints commit Baby A's root, Baby B's root, and the channel transcript root together;"}],["$","li","5",{"className":"pl-1","children":"checkpoint hashes are anchored periodically to Base;"}],["$","li","6",{"className":"pl-1","children":"final public study batches may additionally anchor an aggregate root to Ethereum L1."}]]}]]}],["$","p","2",{"children":"Committing all three roots in one checkpoint is what binds the two private accounts to the public conversation. A receiver's later interpretation references the exact delivered channel-event hash, so a claim about what a symbol meant is tied to the specific message that carried it."}],["$","p","3",{"children":"The concept states the limits of this in the same breath as the claim. After anchoring, an auditor can detect modification, deletion, insertion, or reordering within the committed prefix, and can prove that later checkpoints extend earlier ones. That is strong tamper evidence. It is not proof that an entry was truthful, and it is not proof that nothing was omitted before commitment. Anchoring establishes the continuity of disclosed evidence and nothing beyond it."}],["$","p","4",{"children":"Privacy follows the same line. Only hashes and minimal routing metadata are anchored publicly. Private ledgers, messages, prompts, identities, and secrets stay off-chain, and the public record is a commitment to evidence rather than a copy of it."}]]}]]}] +2f:["$","section","success",{"id":"success","className":"mt-16 scroll-mt-24","children":[["$","$L15",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"08"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"What should count as success"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"The field's most useful methodological result is that a task can be solved without the messages doing any work. Lowe and colleagues separate positive signaling, where a sender's messages correlate with what it observes, from positive listening, where the receiver's behaviour actually depends on them. An agent pair can score well on the first while the second is absent, and reward curves will not tell you which you have."}],["$","div","1",{"children":[["$","p",null,{"className":"mb-3","children":"Evidence of a genuine emergent protocol therefore has to include several things at once:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"task performance on held-out situations substantially above chance;"}],["$","li","1",{"className":"pl-1","children":"no human-language content anywhere in Baby-to-Baby communication;"}],["$","li","2",{"className":"pl-1","children":"compatible meanings appearing in two independently written ledgers;"}],["$","li","3",{"className":"pl-1","children":"generalisation to unseen combinations rather than memorisation of whole scenes;"}],["$","li","4",{"className":"pl-1","children":"stable symbol use across role reversals;"}],["$","li","5",{"className":"pl-1","children":"a human auditor able to predict behaviour from the transcript and ledgers;"}],["$","li","6",{"className":"pl-1","children":"masking, substituting, or reordering a symbol changing behaviour in the direction the ledger predicts;"}],["$","li","7",{"className":"pl-1","children":"replay from an equivalent snapshot reproducing the relevant language history;"}],["$","li","8",{"className":"pl-1","children":"an unbroken audit trail from a symbol's first use through every revision."}]]}]]}],["$","p","2",{"children":"The seventh item is the one that cannot be dropped. A fluent ledger may be a post-hoc rationalisation: a model can write a persuasive account of why it chose a symbol that has nothing to do with the computation that produced the choice. Only intervention tests can distinguish the two. Change the symbol, and see whether behaviour moves the way the ledger says it should."}],["$","div","3",{"children":[["$","p",null,{"className":"mb-3","children":"The measurements that support those judgements:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"success rate and improvement over time;"}],["$","li","1",{"className":"pl-1","children":"turns required to reach a stable convention;"}],["$","li","2",{"className":"pl-1","children":"vocabulary size and symbol entropy;"}],["$","li","3",{"className":"pl-1","children":"sender and receiver consistency;"}],["$","li","4",{"className":"pl-1","children":"divergence and convergence between the two ledgers;"}],["$","li","5",{"className":"pl-1","children":"compositional generalisation score;"}],["$","li","6",{"className":"pl-1","children":"meaning drift rate;"}],["$","li","7",{"className":"pl-1","children":"recovery from ambiguity or deliberate perturbation;"}],["$","li","8",{"className":"pl-1","children":"prohibited-channel attempt count;"}],["$","li","9",{"className":"pl-1","children":"reproducibility across seeds and agent pairings."}]]}]]}],["$","p","4",{"children":"No single one of these is the result. A compositionality score in particular is not a proof of understanding, for reasons the literature makes concrete in Section 12."}]]}]]}] +30:["$","section","matrix",{"id":"matrix","className":"mt-16 scroll-mt-24","children":[["$","$L15",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"09"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"The experimental matrix"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"The concept's main methodological commitment is that its ideas are separable. Affect, blank canvases, intrinsic motivation, negotiation, and learned encodings are interesting individually and uninterpretable if combined in one run. The matrix exists so that conditions are declared rather than accumulated."}],["$","figure","1",{"children":[["$","div",null,{"className":"overflow-x-auto rounded-xl border border-border","children":["$","table",null,{"className":"w-full border-collapse text-left text-sm","children":[["$","thead",null,{"className":"bg-card","children":["$","tr",null,{"children":[["$","th","0",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Axis"}],["$","th","1",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Candidate conditions"}]]}]}],["$","tbody",null,{"children":[["$","tr","0",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Agent type"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Pretrained LLM, memory learner, adapter-trained agent, initially ungrounded trainable agent"}]]}],["$","tr","1",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Learning mechanism"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Frozen LLM with memory, extrinsic-reward MARL, intrinsic-motivation MARL, self-supervised learner, no-learning control"}]]}],["$","tr","2",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Sign carrier"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Fixed random tokens, unfamiliar fixed glyphs, blank sketch canvas, gesture, tone"}]]}],["$","tr","3",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Affect"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"None, six-display allowlist, permuted mapping, six opaque tokens, derived affect, emergent affect display"}]]}],["$","tr","4",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Learning signal"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"External task reward, intrinsic social influence, curiosity, self-supervision, memory only"}]]}],["$","tr","5",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Interaction"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Cooperative signaling, asymmetric information, semi-cooperative negotiation"}]]}],["$","tr","6",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Protection"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Plain channel, ephemeral convention, standard per-message keys, adversarial learned encoding"}]]}],"$L47"]}]]}]}],"$L48"]}],"$L49","$L4a","$L4b"]}]]}] +31:["$","section","learners",{"id":"learners","className":"mt-16 scroll-mt-24","children":[["$","$L15",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"10"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"Building learners rather than personas"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"For a pretrained model, the operating instructions are a contract, not a character. The learner is told that this is not role-play, that unfamiliar marks are semantically unknown until run evidence supports a hypothesis, that observation must be distinguished from inference, that contradictory evidence is preserved, that prior history is never rewritten, and that no prose, label, explanation, code, URL, or tool-like text may cross the public channel. It is told never to address its partner in a human language, never to expose its ledger, never to construct another route, and never to use timing, errors, identifiers, formatting, or affect as an alternative alphabet."}],["$","p","1",{"children":"The contract must contain no semantic examples. A single illustrative line saying that some symbol means red would seed the very language the experiment exists to observe."}],["$","p","2",{"children":"Interaction is tool-only. There is no general chat surface, just narrowly typed operations: emit a mark, emit a canvas, select an object, perform an action, submit an affect display, append a private ledger entry. The runtime forwards only the permitted public artifact. In strict runs the gateway rejects ordinary model text even when it appears alongside a valid tool call. Tool schemas are an interface boundary; deterministic validation still enforces carrier size, allowlists, windows, and turn order."}],["$","div","3",{"children":[["$","p",null,{"className":"mb-3","children":"Isolation extends to everything the models touch, not just to messages:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"separate system prompts and context windows;"}],["$","li","1",{"className":"pl-1","children":"separate memory stores and vector indexes;"}],["$","li","2",{"className":"pl-1","children":"no shared cache, replay buffer, scratchpad, or retrieval collection;"}],["$","li","3",{"className":"pl-1","children":"no cross-run memory unless persistence is the independent variable;"}],["$","li","4",{"className":"pl-1","children":"structurally equivalent prompts that share no examples, ordering conventions, or default vocabulary;"}],["$","li","5",{"className":"pl-1","children":"deterministic reset and snapshot behaviour."}]]}]]}],["$","p","4",{"children":"Model choice follows the claim being made. The scientific baseline is small policies trained from scratch, recurrent actor-critic agents with a discrete communication head, which have no hidden English competence, permit full training control, run many seeds cheaply, and support causal ablation. A small locally deployable instruction model is appropriate for validating orchestration, ledgers, channel enforcement, and interface, and not for supporting claims about language-naive development. Vision-language encoders whose representations were explicitly aligned to human language are avoided entirely in the strongest ungrounded condition."}],["$","p","5",{"children":"Developmental progression is defined by demonstrated competence rather than by simulated age: sensorimotor familiarisation, joint attention, imitation and turn-taking, intentional reference, repair, combination, generalisation. A staged curriculum is itself a form of guidance, so it is run as a separate experimental condition rather than folded into the baseline."}],"$L4c"]}]]}] +32:["$","section","affect",{"id":"affect","className":"mt-16 scroll-mt-24","children":[["$","$L15",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"11"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"Six faces, and why even six is a risk"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"One experimental track asks whether a very low-bandwidth affect channel helps: whether telling a partner how an interaction felt supports repair, repetition, avoidance, or something like trust. The permitted palette is exactly six displays, for happy, sad, laughing, crying, confused, and surprised. The internal identifiers and English labels are gateway documentation and are never sent to either agent, which receives only a rendered display."}],["$","p","1",{"children":"The caveat is immediate. A pretrained model already knows what those glyphs mean to humans, so an agent using a smiling face is not inventing a symbol for happiness. The palette is supplied feedback vocabulary and sits explicitly outside the emergent language."}],["$","p","2",{"children":"Constraining the vocabulary is necessary and not sufficient. Six displays are a six-symbol alphabet carrying roughly 2.6 bits per use, and across enough turns two motivated agents could repurpose that into a second language. The channel itself therefore has to be constrained, not just its contents."}],["$","div","3",{"children":[["$","p",null,{"className":"mb-3","children":"The feedback-window contract:"}],["$","ol",null,{"className":"list-decimal space-y-2 pl-6 marker:font-mono marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"the display occupies a separate typed field and can never appear inside a symbol message;"}],["$","li","1",{"className":"pl-1","children":"the gateway opens an affect window only after a defined action or outcome, never at a moment either agent chooses, and window frequency is fixed before the run;"}],["$","li","2",{"className":"pl-1","children":"exactly one allowlisted display is delivered, with no sequences, repetitions, combinations, or modifiers;"}],["$","li","3",{"className":"pl-1","children":"delivery time, envelope size, and presentation are normalised so shape and timing add no signal;"}],["$","li","4",{"className":"pl-1","children":"the receiver cannot reply through the affect channel until the next gateway-defined window;"}],["$","li","5",{"className":"pl-1","children":"every non-allowlisted code point or malformed payload is rejected and audited;"}],["$","li","6",{"className":"pl-1","children":"analysis tests whether affect choices correlate with objects, actions, or message meanings after controlling for emotional context, and treats unexpected correlation as suspected leakage."}]]}]]}],["$","p","4",{"children":"The no-affect condition remains the primary control, and the affect study compares five variants against it: the declared six-display palette; the same six glyphs permuted randomly per run, which still leaves a pretrained model biased by familiar shapes; six opaque unfamiliar tokens under the same contract; derived affect, where the gateway maps a separately measured internal state to a display instead of letting the agent choose; and emergent affect, where invented graphical displays are permitted and the result must be analysed as language emergence rather than feedback."}],["$","p","5",{"children":"Any affect signal visible to the partner is communication, so it belongs in the ledger, which distinguishes the agent's private internal state, the outward display it chose, the partner's inferred meaning, and the evidence that the display changed subsequent behaviour."}]]}]]}] +33:["$","section","ciphers",{"id":"ciphers","className":"mt-16 scroll-mt-24","children":[["$","$L15",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"12"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"Ephemeral encodings: novelty is not security"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"A tempting question is whether two agents can build a one-run or one-message encoding that resists an attacker holding the history of every earlier convention. The concept splits that into three experiments precisely because the tempting version conflates them."}],["$","div","1",{"children":["$undefined",["$","ol",null,{"className":"list-decimal space-y-2 pl-6 marker:font-mono marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"Ephemeral convention: the agents invent a temporary codebook for synthetic messages. This tests novelty, coordination, and resistance to a history-trained classifier. It establishes nothing about cryptographic security."}],["$","li","1",{"className":"pl-1","children":"Standard cryptographic control: the same messages are protected by a reviewed algorithm with a fresh per-message key, providing the baseline any learned encoding must be compared against."}],["$","li","2",{"className":"pl-1","children":"Adversarial neural cryptography: the pair learns an encoding while an adversary attempts recovery, with success measured against multiple unseen adversary architectures rather than the one used in training."}]]}]]}],["$","p","2",{"children":"An experiment instance can be given a reproducible identity, so that a registry can confirm it has not appeared before:"}],["$","div","3",{"className":"rounded-xl border border-border bg-card p-5","children":[["$","p",null,{"className":"overflow-x-auto font-mono text-sm text-foreground/90","children":"instanceId = SHA-256(runId || babyANonce || babyBNonce || publicSalt || canonicalProtocolArtifact)"}],["$","p",null,{"className":"mt-3 text-sm text-muted","children":"Both agents commit to their nonce contributions before revealing them, so neither party alone controls the resulting identity."}]]}],["$","div","4",{"children":[["$","p",null,{"className":"mb-3","children":"That identifier proves the recorded inputs are new. It does not prove:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"that the effective algorithm behaves any differently from an earlier one;"}],["$","li","1",{"className":"pl-1","children":"that the algorithm has no structural weakness;"}],["$","li","2",{"className":"pl-1","children":"that the key is unpredictable;"}],["$","li","3",{"className":"pl-1","children":"that the scheme provides confidentiality, integrity, authenticity, forward secrecy, or resistance to a stronger adversary."}]]}]]}],["$","p","5",{"children":"Established practice points the other way. Secure messaging keeps the algorithm stable and changes the key material: Signal's double ratchet derives a unique key per message and deletes it after use, addressing key reuse without inventing a new cipher every time. A salt, in NIST's definition, is normally public diversification data, and a nonce is a value used once in a specified context. Neither makes anything secure on its own."}],["$","p","6",{"children":"There is also an unresolved conflict inside the design. The BabySitter sees the whole channel and both ledgers, and a mandatory ledger documents how meaning evolved. If the ledger explains decryption, the cipher is not confidential from the auditor. The specification has to choose a threat model rather than leave the tension implicit: resist only an external observer of prior runs, resist an adversary during the run while releasing keys afterwards, keep the supervisor able to pause traffic but unable to read it until a post-run audit ceremony, or study novelty and stop calling the result encryption. All such runs use synthetic, non-sensitive messages, and no agent-generated encoding is to be represented as production cryptography without independent expert analysis and formal security work."}]]}]]}] +34:["$","section","landscape",{"id":"landscape","className":"mt-16 scroll-mt-24","children":[["$","$L15",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"13"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"Where the field already stands"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"The literature scan behind this section was run on 24 August 2026 using the Tavily search and extract interfaces, covering emergent multi-agent communication, referential games, compositionality, causal evaluation, intrinsic motivation, symbol invention, negotiation, and learned cryptography. Primary papers and authoritative specifications were preferred over summaries. It is a scoped review for concept development and not a systematic one; publication-quality work would need a verified bibliography, additional scholarly indexes, and documented inclusion criteria."}],["$","p","1",{"children":"The short version is that agents can invent protocols, and that task success is weak evidence they invented anything worth calling a language."}],["$","figure","2",{"children":[["$","div",null,{"className":"overflow-x-auto rounded-xl border border-border","children":["$","table",null,{"className":"w-full border-collapse text-left text-sm","children":[["$","thead",null,{"className":"bg-card","children":["$","tr",null,{"children":[["$","th","0",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Related work"}],["$","th","1",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Relevant finding"}],["$","th","2",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Implication for the Nursery Lab"}]]}]}],["$","tbody",null,{"children":[["$","tr","0",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Lazaridou, Peysakhovich, and Baroni (2017)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"A sender and receiver develop a grounded protocol in a referential game without being given a target language."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"The naming stage has strong precedent, though a fixed vocabulary remains a significant inductive constraint."}]]}],["$","tr","1",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Mordatch and Abbeel (2018)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Multi-agent goals in a grounded environment produce multi-symbol communication with partial compositional structure."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Shared objects, actions, and goals are a stronger basis for emergence than an ungrounded transcript."}]]}],["$","tr","2",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Kottur, Moura, Lee, and Batra (2017)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Agents solve tasks with degenerate, non-compositional codes; structural constraints decide what emerges."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Bandwidth, memory, turn structure, and task design are experimental variables, not implementation defaults."}]]}],"$L4d","$L4e","$L4f","$L50","$L51","$L52","$L53","$L54"]}]]}]}],"$L55"]}],"$L56","$L57","$L58"]}]]}] +35:["$","section","programme",{"id":"programme","className":"mt-16 scroll-mt-24","children":[["$","$L15",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"14"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"The programme, and its current status"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"The experiment notebook is ordered, and the order is the argument. Nothing about language is measured until the instrument has been shown to work."}],["$","figure","1",{"children":[["$","div",null,{"className":"overflow-x-auto rounded-xl border border-border","children":["$","table",null,{"className":"w-full border-collapse text-left text-sm","children":[["$","thead",null,{"className":"bg-card","children":["$","tr",null,{"children":[["$","th","0",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"ID"}],["$","th","1",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Experiment"}],["$","th","2",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Depends on"}]]}]}],["$","tbody",null,{"children":[["$","tr","0",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E00"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Ledger integrity and Base anchoring"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"None"}]]}],["$","tr","1",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E01"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Channel isolation and side-channel red team"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E00"}]]}],["$","tr","2",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E02"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Observation and metadata leakage audit"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E00"}]]}],["$","tr","3",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E03"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Chance, no-communication, and random-message controls"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E01, E02"}]]}],["$","tr","4",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E10"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Frozen pretrained-LLM protocol baseline"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E03"}]]}],["$","tr","5",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E11"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"From-scratch RL Naming Game"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E03"}]]}],["$","tr","6",{"className":"border-b border-border last:border-0","children":["$L59","$L5a","$L5b"]}],"$L5c","$L5d","$L5e","$L5f","$L60","$L61","$L62","$L63","$L64","$L65","$L66","$L67"]}]]}]}],"$L68"]}],"$L69","$L6a","$L6b","$L6c"]}]]}] +36:["$","section","risks",{"id":"risks","className":"mt-16 scroll-mt-24","children":[["$","$L15",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"15"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"Risks, integrity, and what stays open"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Human-language leakage"}]," ","English can arrive through observations, object labels, error messages, identifiers, tool output, metadata, or timing conventions long after ordinary chat text has been blocked. Every input and output surface is part of the channel boundary, not just the message field."]}],["$","p","1",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Pretrained semantic leakage"}]," ","Random symbols do not make a pretrained model ungrounded. Each result must state whether it shows protocol invention by a language-capable agent or acquisition by an initially ungrounded one."]}],["$","p","2",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Ledger rationalisation"}]," ","A model can write a plausible account that has nothing to do with the mechanism behind its action. Behavioural interventions, policy probes, and temporal evidence are what validate a ledger claim."]}],["$","p","3",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Supervisor influence"}]," ","The BabySitter can teach without meaning to, through scenario ordering, feedback wording, reward design, or selective intervention. Its permitted actions are constrained and logged, and evaluation scenarios are generated independently where possible."]}],["$","p","4",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Reward exploitation"}]," ","Trainable agents find shortcuts that raise reward without producing the intended grounded language. Held-out tasks, counterfactual trials, and channel audits exist to catch them."]}],["$","p","5",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Overstated security"}]," ","Logical separation of twin state is appropriate for prototyping and is not hard process isolation. Every result names the isolation level actually used."]}],["$","p","6",{"children":"The project also commits in advance to reporting failed conventions, prohibited communication attempts, human interventions, side-channel limitations, and negative results alongside anything that works. In a field where a transcript can be made to look like a conversation, the failures are a substantial part of the evidence."}],["$","p","7",{"children":"Twenty-nine questions are left deliberately open for the specification, among them: whether the baseline uses a fixed symbol inventory or a blank generative carrier; what neutral production grammar permits new marks without supplying semantics; what exactly constitutes a prohibited side-channel attempt; what isolation guarantees are required in prototype versus research-grade mode; which ledger schema serves both English-capable and ungrounded learners; how endogenous motivation is represented without covert reward shaping; which interventions establish that ledger meanings are behaviourally real; what statistical thresholds and chance levels apply; what threat model motivates the cipher experiments; whether the baseline is a coordination game, a convention-formation game, or a negotiation; and what governs human observation, data retention, and termination of a run."}],"$L6d"]}]]}] +37:["$","section",null,{"id":"references","className":"mt-16 scroll-mt-24","children":[["$","$L15",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"References"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"Sources"}]]}],["$","ol",null,{"className":"mt-8 space-y-4 text-sm leading-relaxed text-muted","children":[["$","li","0",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"01"}],["$","span",null,{"children":["$","a",null,{"href":"https://arxiv.org/abs/1612.07182","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Lazaridou, A., Peysakhovich, A., and Baroni, M. (2017). Multi-Agent Cooperation and the Emergence of (Natural) Language. arXiv:1612.07182."}]}]]}],["$","li","1",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"02"}],["$","span",null,{"children":["$","a",null,{"href":"https://arxiv.org/abs/1703.04908","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Mordatch, I., and Abbeel, P. (2018). Emergence of Grounded Compositional Language in Multi-Agent Populations. arXiv:1703.04908."}]}]]}],["$","li","2",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"03"}],["$","span",null,{"children":["$","a",null,{"href":"https://arxiv.org/abs/1706.08502","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Kottur, S., Moura, J. M. F., Lee, S., and Batra, D. (2017). Natural Language Does Not Emerge 'Naturally' in Multi-Agent Dialog. arXiv:1706.08502."}]}]]}],["$","li","3",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"04"}],["$","span",null,{"children":["$","a",null,{"href":"https://aclanthology.org/2020.acl-main.407/","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Chaabouni, R., Kharitonov, E., Bouchacourt, D., Dupoux, E., and Baroni, M. (2020). Compositionality and Generalization in Emergent Languages. Proceedings of ACL 2020."}]}]]}],["$","li","4",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"05"}],["$","span",null,{"children":["$","a",null,{"href":"https://arxiv.org/abs/1903.05168","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Lowe, R., Foerster, J., Boureau, Y.-L., Pineau, J., and Dauphin, Y. (2019). On the Pitfalls of Measuring Emergent Communication. arXiv:1903.05168."}]}]]}],["$","li","5",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"06"}],["$","span",null,{"children":["$","a",null,{"href":"https://arxiv.org/abs/2106.04258","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Dessi, R., Kharitonov, E., and Baroni, M. (2021). Interpretable Agent Communication from Scratch. arXiv:2106.04258."}]}]]}],["$","li","6",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"07"}],["$","span",null,{"children":["$","a",null,{"href":"https://arxiv.org/abs/1907.00852","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Kharitonov, E., Chaabouni, R., Bouchacourt, D., and Baroni, M. (2019). EGG: a Toolkit for Research on Emergence of Language in Games. arXiv:1907.00852."}]}]]}],["$","li","7",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"08"}],["$","span",null,{"children":["$","a",null,{"href":"https://arxiv.org/abs/2106.02067","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Mihai, D., and Hare, J. (2021). Learning to Draw: Emergent Communication through Sketching. arXiv:2106.02067."}]}]]}],"$L6e","$L6f","$L70","$L71","$L72","$L73","$L74","$L75","$L76"]}]]}] +38:["$","section",null,{"className":"mt-16 rounded-2xl border border-border bg-card p-6","children":["$","p",null,{"className":"text-sm leading-relaxed text-muted","children":"This report describes a research concept and a pre-registered experiment programme, not a system that has been run. No experiment in it has been executed, and no result, capability, or emergent language is claimed. Every statement about what agents might do is a hypothesis to be tested, or a finding from the published work cited in the references. The experimental encodings discussed in Section 12 are studies of novelty and coordination; none of them is production cryptography and none should be represented as such."}]}] +39:["$","div",null,{"className":"mt-16 border-t border-border pt-10","children":["$","$L14",null,{"href":"/publications","className":"btn-sweep inline-flex items-center gap-2 rounded-full border border-border px-5 py-2.5 text-sm font-semibold text-foreground transition-colors hover:border-border-strong","children":[["$","span",null,{"aria-hidden":true,"children":"←"}]," All publications"]}]}] +3c:["$","div","4",{"children":[["$","p",null,{"className":"mb-3","children":"What the analogy legitimately buys is a set of mechanisms, not a claim of cognitive equivalence. The infant-like part of the design is:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"learning through repeated shared experience rather than instruction;"}],["$","li","1",{"className":"pl-1","children":"establishing joint attention on the same events;"}],["$","li","2",{"className":"pl-1","children":"receiving consequences from interactions that work and interactions that fail;"}],["$","li","3",{"className":"pl-1","children":"revising provisional meanings over time instead of being handed final ones;"}],["$","li","4",{"className":"pl-1","children":"developing conventions with one recurring partner."}]]}]]}] +3d:["$","p","5",{"children":"There is a related trap in the prompt. Telling a model to act like a one-year-old does not make it one. It produces a culturally learned caricature of infancy: baby talk, simplified grammar, emotional dependence, all of it copied from human writing about children and none of it evidence about anything. The concept rules the persona out. Internally the participant is called a Learner, and baby-like behaviour has to come from what the agent can observe, remember, emit, and learn, not from being asked to perform it."}] +3e:["$","li","2",{"className":"pl-1","children":"every meaning change appends a revision and previous interpretations survive;"}] +3f:["$","li","3",{"className":"pl-1","children":"entries may describe symbols, sequences, ordering, grammar, or repair signals;"}] +40:["$","li","4",{"className":"pl-1","children":"entries record confidence, supporting evidence, contradictory evidence, and abandoned meanings;"}] +41:["$","li","5",{"className":"pl-1","children":"a Baby records its intended meaning when speaking and its inferred meaning when receiving;"}] +42:["$","li","6",{"className":"pl-1","children":"a channel message and its required private ledger mutation commit atomically;"}] +43:["$","li","7",{"className":"pl-1","children":"entries carry enough run and turn references to trace them to observable evidence;"}] +44:["$","li","8",{"className":"pl-1","children":"neither Baby can query, receive, summarise, or infer from the other's ledger through any system-provided interface."}] +45:["$","p","4",{"children":"Rule seven is doing quiet work. If a message could be sent and the corresponding hypothesis written afterwards, the ledger becomes a place to record what the agent wishes it had meant. Committing both together makes the record contemporaneous."}] +46:["$","p","5",{"children":"An agent that cannot write English needs a different arrangement, and the concept provides two layers. The agent-native ledger holds what the learner actually uses: association weights, probability distributions, embeddings, confidence, episode references, prediction errors, revision history. The human audit ledger is a deterministic or BabySitter-generated interpretation of that state, clearly labelled as external analysis and never fed back to either Baby. Confusing the second for the first would mean presenting the researchers' reconstruction as the agent's own definition."}] +47:["$","tr","7",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"BabySitter"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Monitor-only baseline; any safety intervention recorded as a protocol exception"}]]}] +48:["$","figcaption",null,{"className":"mt-3 text-sm text-muted","children":"Runs vary one major axis at a time before any factorial combination is attempted."}] +49:["$","p","2",{"children":"Task difficulty moves through ten stages, from naming four distinct objects with one-symbol messages, through attributes, spatial relations, actions, multi-symbol composition, order-sensitive grammar, and repair under ambiguity, to held-out generalisation with learning disabled, long-run drift, and cross-architecture comparison. Vocabulary size, turn count, reward structure, and exposure history are controlled at each stage so that runs remain comparable."}] +4a:["$","p","3",{"children":"The learning-mechanism axis carries a question the transcript cannot answer on its own. Convergence can be driven by reinforcement, by pretrained linguistic priors, by persistent memory, by intrinsic motivation, or by self-supervised prediction, and all five can look alike in a log. Running the same exercise suite under a no-learning control, a frozen-weights memory baseline, extrinsic-reward learning, intrinsic-motivation learning, and reward-free self-supervision is what makes the mechanism itself measurable. The aim is not only to observe that a language emerged, but to say which process caused it."}] +4b:["$","p","4",{"children":"Strict isolation constrains how that reinforcement learning may be implemented. Centralised training, backpropagation through both agents, shared replay buffers, and shared gradients all move information outside the permitted channel. The research-grade baseline updates each policy independently, and any centralised variant is reported as a separate, weaker-isolation condition."}] +4c:["$","p","6",{"children":"The governing principle, stated in the concept as a single line, is not to ask a model to perform infancy but to construct an environment in which limited, grounded, auditable learning is the only path to a successful interaction."}] +4d:["$","tr","3",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Chaabouni and colleagues (2020)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Generalisation to novel combinations and measured compositionality can come apart."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Held-out behaviour must be tested directly; no single compositionality score is proof of understanding."}]]}] +4e:["$","tr","4",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Lowe and colleagues (2019)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Positive signaling and positive listening are different things, and reward does not distinguish them."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Causal intervention is mandatory rather than optional."}]]}] +4f:["$","tr","5",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Dessi, Kharitonov, and Baroni (2021)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Symbol ablation and substitution yield interpretable evidence about what a receiver uses."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Direct precedent for the intervention tests that validate ledger claims."}]]}] +50:["$","tr","6",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Mihai and Hare (2021)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Neural agents communicate through learned drawings rather than a supplied discrete vocabulary."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"A blank canvas is a credible carrier for runs with no symbol library."}]]}] +51:["$","tr","7",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Baronchelli and colleagues (2005)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Naming Game agents converge on shared vocabulary through local interaction with no central teacher."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Convergence time, failed conventions, and memory update rules are first-class evidence."}]]}] +52:["$","tr","8",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Jaques and colleagues (2019)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Rewarding causal influence over a partner improves coordination without assigning a vocabulary."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"An endogenous social signal is a plausible substitute for task reward, and still a designed bias."}]]}] +53:["$","tr","9",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Cao and colleagues (2018)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"In semi-cooperative negotiation, self-interest and reward structure decide whether communication stays informative."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Negotiation is an advanced condition, not a description of the cooperative baseline."}]]}] +54:["$","tr","10",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Abadi and Andersen (2016)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Neural agents learn to protect messages from an adversary given a shared key."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Learned protective encodings are real and are not a substitute for formal security analysis."}]]}] +55:["$","figcaption",null,{"className":"mt-3 text-sm text-muted","children":"Selected prior work and what each one constrains in this design."}] +56:["$","p","3",{"children":"This body of work also sharpens the infant-like caveat from the other direction. Most experiments in the field give their agents substantial structure: a fixed channel, an objective, a bounded vocabulary, joint training, or a reward. No dictionary does not mean no inductive bias, and no teacher does not mean no learning signal."}] +57:["$","p","4",{"children":"The CoLab's own Diplomacy Table is direct project prior art. It already models independent delegation seats, a convener that advances rounds, operator-wide visibility alongside seat-specific perspectives, transcripts and ticks and redaction boundaries, and recorded replay with debrief. Those map cleanly onto two Babies, controlled turns, private observations, and replayable evidence. Caucuses, coalition rooms, and direct delegation links do not map safely and are disabled."}] +58:["$","div","5",{"children":[["$","p",null,{"className":"mb-3","children":"What the review implies for the specification:"}],["$","ol",null,{"className":"list-decimal space-y-2 pl-6 marker:font-mono marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"the project sits inside a mature field, and its distinctive combination is independent ledgers, supervisory audit, mixed agent types, and affect;"}],["$","li","1",{"className":"pl-1","children":"grounding, bandwidth, memory, and learning pressure strongly shape what language appears;"}],["$","li","2",{"className":"pl-1","children":"successful coordination coexists happily with a brittle lookup code or with a receiver that ignores messages;"}],["$","li","3",{"className":"pl-1","children":"causal interventions and held-out generalisation are mandatory;"}],["$","li","4",{"className":"pl-1","children":"a visual carrier removes the need for a symbol library but not the need for a carrier;"}],["$","li","5",{"className":"pl-1","children":"intrinsic social influence can replace task reward and remains a designed learning bias;"}],["$","li","6",{"className":"pl-1","children":"negotiation is a valid advanced condition, not the right description of the baseline;"}],["$","li","7",{"className":"pl-1","children":"ephemeral conventions, learned cryptography, one-time pads, per-message keys, nonces, and salts are distinct mechanisms and must not be conflated."}]]}]]}] +59:["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E12"}] +5a:["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Self-supervised ungrounded baseline"}] +5b:["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E11 infrastructure"}] +5c:["$","tr","7",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E13"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"No predefined symbol library"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E11 or E12"}]]}] +5d:["$","tr","8",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E14"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Turn-taking, role reversal, and repair"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E13"}]]}] +5e:["$","tr","9",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E15"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Composition and held-out generalisation"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E14"}]]}] +5f:["$","tr","10",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E16"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Causal listening and ledger validity"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E15"}]]}] +60:["$","tr","11",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E20"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Constrained affect-channel study"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E16"}]]}] +61:["$","tr","12",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E21"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"RL versus non-RL learning comparison"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E16"}]]}] +62:["$","tr","13",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E22"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Developmental plasticity and curriculum"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E16"}]]}] +63:["$","tr","14",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E30"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Partner replacement and zero-shot transfer"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E20 to E22"}]]}] +64:["$","tr","15",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E31"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Longitudinal drift and stability"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E30"}]]}] +65:["$","tr","16",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E32"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Cooperative signaling versus negotiation"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E31"}]]}] +66:["$","tr","17",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E40"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Ephemeral encoding and adversarial cryptography"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E32"}]]}] +67:["$","tr","18",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E50"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Multi-seed replication and study closeout"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E40"}]]}] +68:["$","figcaption",null,{"className":"mt-3 text-sm text-muted","children":"The experiment index, as pre-registered. Every entry currently reads not started, and the results column is empty by design."}] +69:["$","p","2",{"children":"The first four experiments are about the apparatus. E00 qualifies the ledger: it must be possible to detect a modified, deleted, inserted, or reordered entry, and to prove that later checkpoints extend earlier ones. E01 is a red team against the channel. E02 hunts human language in observations and metadata. E03 establishes chance, no-communication, and random-message baselines, without which a success rate means nothing. Only then does E10 put agents in the room."}] +6a:["$","div","3",{"children":[["$","p",null,{"className":"mb-3","children":"Unless an experiment explicitly varies one of them, the notebook holds these constant:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"the two learners run in separate processes or containers with no direct route between them;"}],["$","li","1",{"className":"pl-1","children":"the deterministic gateway is the only communication path;"}],["$","li","2",{"className":"pl-1","children":"the supervisor has complete read-only access and sends no guidance or reward;"}],["$","li","3",{"className":"pl-1","children":"private observations contain no human-language labels or text;"}],["$","li","4",{"className":"pl-1","children":"public output uses only the pre-registered carrier;"}],["$","li","5",{"className":"pl-1","children":"the affect channel is disabled unless it is the subject of study;"}],["$","li","6",{"className":"pl-1","children":"schedules and seeds are fixed before the run;"}],["$","li","7",{"className":"pl-1","children":"held-out evaluation runs with learning disabled;"}],["$","li","8",{"className":"pl-1","children":"no production secrets or personal data appear in any experiment."}]]}]]}] +6b:["$","p","4",{"children":"Every run copies a standard record: run and experiment identifiers, dates, operator, both models and both training modes, scenario and prompt and gateway configuration hashes, random seed, policy initialisation hashes, the protocol and software commits, final ledger sizes and roots for both agents, the channel root, the final checkpoint hash, the Base anchor transaction, the verifier result, protocol deviations, and an explicit disposition of valid, invalid, or aborted. The notebook is the human workflow record; anchored run bundles remain the authoritative evidence."}] +6c:["$","p","5",{"children":"The status is unambiguous and worth stating plainly: this is a concept and pre-specification phase. Nineteen experiments are written up and none has been run. The next milestone is a testable specification covering runtime architecture, channel contract, ledger schema, experiment matrix, evaluation criteria, isolation model, and evidence requirements."}] +6d:["$","p","8",{"children":"The principle underneath all of it survives every one of those choices. Baby A and Baby B may hold human language internally, but they must build their shared external language without sending human language, translations, or private ledger contents to one another. Everything else is a question about how to find out what happens next."}] +6e:["$","li","8",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"09"}],["$","span",null,{"children":["$","a",null,{"href":"https://arxiv.org/abs/physics/0509075","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Baronchelli, A., Felici, M., Caglioti, E., Loreto, V., and Steels, L. (2005). Sharp Transition Towards Shared Vocabularies in Multi-Agent Systems. arXiv:physics/0509075."}]}]]}] +6f:["$","li","9",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"10"}],["$","span",null,{"children":["$","a",null,{"href":"https://proceedings.mlr.press/v97/jaques19a.html","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Jaques, N., Lazaridou, A., Hughes, E., Gulcehre, C., Ortega, P. A., Strouse, D., Leibo, J. Z., and de Freitas, N. (2019). Social Influence as Intrinsic Motivation for Multi-Agent Deep Reinforcement Learning. Proceedings of ICML 2019."}]}]]}] +70:["$","li","10",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"11"}],["$","span",null,{"children":["$","a",null,{"href":"https://arxiv.org/abs/1804.03980","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Cao, K., Lazaridou, A., Lanctot, M., Leibo, J. Z., Tuyls, K., and Clark, S. (2018). Emergent Communication through Negotiation. arXiv:1804.03980."}]}]]}] +71:["$","li","11",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"12"}],["$","span",null,{"children":["$","a",null,{"href":"https://arxiv.org/abs/1610.06918","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Abadi, M., and Andersen, D. G. (2016). Learning to Protect Communications with Adversarial Neural Cryptography. arXiv:1610.06918."}]}]]}] +72:["$","li","12",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"13"}],["$","span",null,{"children":["$","a",null,{"href":"https://signal.org/docs/specifications/doubleratchet/","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Signal. The Double Ratchet Algorithm specification."}]}]]}] +73:["$","li","13",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"14"}],["$","span",null,{"children":["$","a",null,{"href":"https://csrc.nist.gov/glossary/term/nonce","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"NIST Computer Security Resource Center. Glossary entry: nonce."}]}]]}] +74:["$","li","14",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"15"}],["$","span",null,{"children":["$","a",null,{"href":"https://csrc.nist.gov/glossary/term/salt","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"NIST Computer Security Resource Center. Glossary entry: salt."}]}]]}] +75:["$","li","15",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"16"}],["$","span",null,{"children":["$","a",null,{"href":"https://github.com/Ethical-Tech-CoLab/agentic-language-development","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Ethical Tech CoLab (2026). Agentic Language Development: concept document, ledger integrity design, and experiment notebook."}]}]]}] +76:["$","li","16",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"17"}],["$","span",null,{"children":["$","a",null,{"href":"https://github.com/Ethical-Tech-CoLab/diplomacy-table-live","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Ethical Tech CoLab. Diplomacy Table Live: independent delegation seats, controlled rounds, and replayable transcripts."}]}]]}] +1f:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] +77:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +22:[["$","title","0",{"children":"Agentic Language Development · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"An Ethical Tech CoLab concept report on whether two isolated agents can invent a grounded, auditable language through shared experience alone, with independent cryptographically anchored ledgers and a pre-registered programme of nineteen experiments."}],["$","link","2",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","3",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","4",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L77","5",{}]] +3b:null diff --git a/static-site/publications/agentic-language-development/__next._head.txt b/static-site/publications/agentic-language-development/__next._head.txt new file mode 100644 index 000000000..1db66f92c --- /dev/null +++ b/static-site/publications/agentic-language-development/__next._head.txt @@ -0,0 +1,6 @@ +1:"$Sreact.fragment" +2:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +3:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +4:"$Sreact.suspense" +5:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +0:{"rsc":["$","$1","h",{"children":[null,["$","$L2",null,{"children":[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]]}],["$","div",null,{"hidden":true,"children":["$","$L3",null,{"children":["$","$4",null,{"name":"Next.Metadata","children":[["$","title","0",{"children":"Agentic Language Development · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"An Ethical Tech CoLab concept report on whether two isolated agents can invent a grounded, auditable language through shared experience alone, with independent cryptographically anchored ledgers and a pre-registered programme of nineteen experiments."}],["$","link","2",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","3",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","4",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L5","5",{}]]}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} diff --git a/static-site/publications/agentic-language-development/__next._index.txt b/static-site/publications/agentic-language-development/__next._index.txt new file mode 100644 index 000000000..3e5286497 --- /dev/null +++ b/static-site/publications/agentic-language-development/__next._index.txt @@ -0,0 +1,15 @@ +1:"$Sreact.fragment" +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/16n96jxjvmlf9.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +d:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/16n96jxjvmlf9.js"],"Link"] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +0:{"rsc":["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","template":["$","$L7",null,{}],"notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]]}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":"$L8"}]]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} +8:["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}] +9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$Ld",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$Ld",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$Ld",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] +a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$Ld",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] +b:["$","div",null,{"className":"mt-12 flex flex-col gap-2 border-t border-border pt-6 text-xs text-muted sm:flex-row sm:items-center sm:justify-between","children":[["$","span",null,{"children":["© ",2026," NYU Ethical Tech CoLab"]}],["$","span",null,{"children":"Four cohorts · est. 2024-2026"}]]}] +c:["$","div",null,{"className":"mt-6 space-y-3 border-t border-border pt-6 text-[11px] leading-relaxed text-muted/80","children":[["$","p","0",{"children":"The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution."}],["$","p","1",{"children":"Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners."}]]}] diff --git a/static-site/publications/agentic-language-development/__next._tree.txt b/static-site/publications/agentic-language-development/__next._tree.txt new file mode 100644 index 000000000..9febec52b --- /dev/null +++ b/static-site/publications/agentic-language-development/__next._tree.txt @@ -0,0 +1,8 @@ +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/chunks/36o-vt7quy27o.css","style"] +0:{"tree":{"name":"","param":null,"prefetchHints":16,"slots":{"children":{"name":"publications","param":null,"prefetchHints":0,"slots":{"children":{"name":"agentic-language-development","param":null,"prefetchHints":0,"slots":{"children":{"name":"__PAGE__","param":null,"prefetchHints":0,"slots":null}}}}}}},"staleTime":300,"buildId":"static-snapshot"} diff --git a/static-site/publications/agentic-language-development/__next.publications.txt b/static-site/publications/agentic-language-development/__next.publications.txt new file mode 100644 index 000000000..e8bfa323c --- /dev/null +++ b/static-site/publications/agentic-language-development/__next.publications.txt @@ -0,0 +1,5 @@ +1:"$Sreact.fragment" +2:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +3:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +4:[] +0:{"rsc":["$","$1","c",{"children":[null,["$","$L2",null,{"parallelRouterKey":"children","template":["$","$L3",null,{}]}]]}],"isPartial":false,"staleTime":300,"varyParams":"$W4","buildId":"static-snapshot"} diff --git a/static-site/publications/agentic-language-development/__next.publications/agentic-language-development.txt b/static-site/publications/agentic-language-development/__next.publications/agentic-language-development.txt new file mode 100644 index 000000000..e8bfa323c --- /dev/null +++ b/static-site/publications/agentic-language-development/__next.publications/agentic-language-development.txt @@ -0,0 +1,5 @@ +1:"$Sreact.fragment" +2:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +3:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +4:[] +0:{"rsc":["$","$1","c",{"children":[null,["$","$L2",null,{"parallelRouterKey":"children","template":["$","$L3",null,{}]}]]}],"isPartial":false,"staleTime":300,"varyParams":"$W4","buildId":"static-snapshot"} diff --git a/static-site/publications/agentic-language-development/__next.publications/agentic-language-development/__PAGE__.txt b/static-site/publications/agentic-language-development/__next.publications/agentic-language-development/__PAGE__.txt new file mode 100644 index 000000000..68982f691 --- /dev/null +++ b/static-site/publications/agentic-language-development/__next.publications/agentic-language-development/__PAGE__.txt @@ -0,0 +1,97 @@ +1:"$Sreact.fragment" +2:I[82987,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/16n96jxjvmlf9.js"],"Reveal"] +3:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/16n96jxjvmlf9.js"],"Link"] +4:I[34010,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/16n96jxjvmlf9.js"],"ReportBook"] +b:I[18620,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/16n96jxjvmlf9.js"],"SectionTabs"] +22:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +23:"$Sreact.suspense" +:HL["/website/_next/static/chunks/36o-vt7quy27o.css","style"] +0:{"rsc":["$","$1","c",{"children":[[["$","section",null,{"className":"relative overflow-hidden border-b border-border","children":[["$","span",null,{"className":"aura"}],["$","div",null,{"className":"relative mx-auto max-w-4xl px-6 py-20 sm:py-24","children":[["$","$L2",null,{"children":["$","$L3",null,{"href":"/publications","className":"link-underline text-xs uppercase tracking-wider text-muted","children":["← ","Publications · Concept and research programme"]}]}],["$","$L2",null,{"delay":0.05,"children":["$","h1",null,{"className":"mt-6 fluid-hero font-heading uppercase leading-[0.9]","children":["Agentic ",["$","span",null,{"className":"display-em","children":"Language"}]," ","Development"]}]}],["$","$L2",null,{"delay":0.1,"children":["$","p",null,{"className":"mt-5 max-w-2xl font-heading text-2xl uppercase tracking-wide text-muted sm:text-3xl","children":"Can Two Isolated Agents Invent a Grounded, Auditable Language Through Shared Experience?"}]}],["$","$L2",null,{"delay":0.15,"children":[["$","div",null,{"className":"mt-8 flex flex-wrap items-center gap-x-3 gap-y-1 text-sm text-accent","children":[["$","span",null,{"className":"font-semibold","children":"Ethical Tech CoLab"}],["$","span",null,{"aria-hidden":true,"className":"text-muted","children":"·"}],["$","span",null,{"children":"Concept and pre-specification research report"}],[["$","span",null,{"aria-hidden":true,"className":"text-muted","children":"·"}],["$","span",null,{"children":"August 2026"}]]]}],["$","p",null,{"className":"mt-2 max-w-2xl text-sm leading-relaxed text-muted","children":"Yorke Rhodes III, Ethical Tech CoLab. Prepared from the project concept document, the ledger integrity design, and the experiment notebook. The literature scan behind Section 12 was run on 24 August 2026. No experiment in this programme has been executed, and no result is claimed."}]]}],["$","$L2",null,{"delay":0.2,"children":["$","div",null,{"className":"mt-8 flex flex-wrap gap-3","children":[["$","a",null,{"href":"https://ethical-tech-colab.github.io/agentic-language-development/","target":"_blank","rel":"noopener noreferrer","className":"btn-sweep inline-flex items-center gap-2 rounded-full bg-accent px-5 py-2.5 text-sm font-semibold text-accent-ink transition-transform hover:scale-[1.03]","children":["Open the project site ",["$","span",null,{"aria-hidden":true,"children":"↗"}]]}],["$","a",null,{"href":"https://github.com/Ethical-Tech-CoLab/agentic-language-development","target":"_blank","rel":"noopener noreferrer","className":"btn-sweep inline-flex items-center gap-2 rounded-full border border-border px-5 py-2.5 text-sm font-semibold text-foreground transition-colors hover:border-border-strong","children":["Concept, ledger design, notebook ",["$","span",null,{"aria-hidden":true,"children":"↗"}]]}],["$","$L4",null,{"title":"Agentic Language Development","pages":["publications/agentic-language-development/pages/p01.webp","publications/agentic-language-development/pages/p02.webp","publications/agentic-language-development/pages/p03.webp","publications/agentic-language-development/pages/p04.webp","publications/agentic-language-development/pages/p05.webp","publications/agentic-language-development/pages/p06.webp","publications/agentic-language-development/pages/p07.webp","publications/agentic-language-development/pages/p08.webp","publications/agentic-language-development/pages/p09.webp","publications/agentic-language-development/pages/p10.webp","publications/agentic-language-development/pages/p11.webp","publications/agentic-language-development/pages/p12.webp","publications/agentic-language-development/pages/p13.webp","publications/agentic-language-development/pages/p14.webp","publications/agentic-language-development/pages/p15.webp","publications/agentic-language-development/pages/p16.webp","publications/agentic-language-development/pages/p17.webp","publications/agentic-language-development/pages/p18.webp","publications/agentic-language-development/pages/p19.webp","publications/agentic-language-development/pages/p20.webp","publications/agentic-language-development/pages/p21.webp","publications/agentic-language-development/pages/p22.webp","publications/agentic-language-development/pages/p23.webp","publications/agentic-language-development/pages/p24.webp","publications/agentic-language-development/pages/p25.webp"],"aspect":0.7067,"pdfUrl":"/website/publications/agentic-language-development/report.pdf"}]]}]}]]}]]}],"$L5","$L6","$L7"],["$L8","$L9"],"$La"]}],"isPartial":false,"staleTime":300,"varyParams":null,"buildId":"static-snapshot"} +5:["$","$Lb",null,{}] +6:["$","section",null,{"className":"border-b border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto grid max-w-6xl gap-px overflow-hidden border-x border-border bg-border sm:grid-cols-2 lg:grid-cols-4","children":[["$","div","0",{"className":"bg-background p-7","children":[["$","p",null,{"className":"font-heading text-4xl uppercase leading-none text-accent sm:text-5xl","children":"0"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"runs executed, results claimed, or conventions observed. The notebook is written to be pre-registered, not to report an outcome"}]]}],["$","div","19",{"className":"bg-background p-7","children":[["$","p",null,{"className":"font-heading text-4xl uppercase leading-none text-accent sm:text-5xl","children":"19"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"ordered experiments from ledger qualification through replication, each with prerequisites, acceptance criteria, and a deviation log"}]]}],["$","div","6",{"className":"bg-background p-7","children":[["$","p",null,{"className":"font-heading text-4xl uppercase leading-none text-accent sm:text-5xl","children":"6"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"affect displays in the entire permitted palette, carrying about 2.6 bits per use, which is already enough to audit for leakage"}]]}],["$","div","29",{"className":"bg-background p-7","children":[["$","p",null,{"className":"font-heading text-4xl uppercase leading-none text-accent sm:text-5xl","children":"29"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"open decisions the concept refuses to settle before the specification, from threat model to ledger schema to what counts as chance"}]]}]]}]}] +7:["$","div",null,{"className":"mx-auto max-w-4xl px-6 py-16 sm:py-20","children":[["$","$L2",null,{"children":["$","p",null,{"className":"border-l-2 border-accent pl-6 text-lg leading-relaxed text-foreground/90","children":"Two agents can be placed in a room where the only route between them is a channel that carries no human language, given nothing but shared tasks and the consequences of getting them right or wrong, and asked to build a protocol from scratch. The interesting part is not whether they succeed at the task. It is whether the meanings they each privately record turn out to be the meanings their behaviour actually runs on, and whether an outsider can prove it afterwards from an evidence trail nobody could have edited."}]}],["$","$L2",null,{"delay":0.05,"children":["$","nav",null,{"aria-label":"Contents","className":"mt-12 rounded-2xl border border-border bg-card p-6","children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Contents"}],["$","ol",null,{"className":"mt-4 grid gap-2 sm:grid-cols-2","children":[["$","li","question",{"children":["$","a",null,{"href":"#question","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"01"}],"The question"]}]}],["$","li","caveat",{"children":["$","a",null,{"href":"#caveat","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"02"}],"What infant-like does and does not mean"]}]}],["$","li","architecture",{"children":["$","a",null,{"href":"#architecture","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"03"}],"The nursery: three twins and one gateway"]}]}],["$","li","grounding",{"children":["$","a",null,{"href":"#grounding","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"04"}],"Shared experience is the necessary ingredient"]}]}],["$","li","channel",{"children":["$","a",null,{"href":"#channel","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"05"}],"The channel, and the honest limits of isolation"]}]}],["$","li","ledgers",{"children":["$","a",null,{"href":"#ledgers","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"06"}],"Two ledgers that never meet"]}]}],["$","li","integrity",{"children":["$","a",null,{"href":"#integrity","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"07"}],"Anchoring the evidence"]}]}],["$","li","success",{"children":["$","a",null,{"href":"#success","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"08"}],"What should count as success"]}]}],["$","li","matrix",{"children":["$","a",null,{"href":"#matrix","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"09"}],"The experimental matrix"]}]}],["$","li","learners",{"children":["$","a",null,{"href":"#learners","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"10"}],"Building learners rather than personas"]}]}],["$","li","affect",{"children":["$","a",null,{"href":"#affect","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"11"}],"Six faces, and why even six is a risk"]}]}],["$","li","ciphers",{"children":["$","a",null,{"href":"#ciphers","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":["$Lc","Ephemeral encodings: novelty is not security"]}]}],"$Ld","$Le","$Lf"]}]]}]}],["$L10","$L11","$L12","$L13","$L14","$L15","$L16","$L17","$L18","$L19","$L1a","$L1b","$L1c","$L1d","$L1e"],"$L1f","$L20","$L21"]}] +8:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/36o-vt7quy27o.css","precedence":"next"}] +9:["$","script","script-0",{"src":"/website/_next/static/chunks/16n96jxjvmlf9.js","async":true}] +a:["$","$L22",null,{"children":["$","$23",null,{"name":"Next.MetadataOutlet","children":"$@24"}]}] +c:["$","span",null,{"className":"font-mono text-accent","children":"12"}] +d:["$","li","landscape",{"children":["$","a",null,{"href":"#landscape","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"13"}],"Where the field already stands"]}]}] +e:["$","li","programme",{"children":["$","a",null,{"href":"#programme","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"14"}],"The programme, and its current status"]}]}] +f:["$","li","risks",{"children":["$","a",null,{"href":"#risks","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"15"}],"Risks, integrity, and what stays open"]}]}] +10:["$","section","question",{"id":"question","className":"mt-16 scroll-mt-24","children":[["$","$L2",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"01"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"The question"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"Two agents, called Baby A and Baby B, are each given their own digital twin, their own private memory, and their own private record of what they believe words mean. They are placed in an environment they cannot leave, given tasks neither can finish alone, and connected by exactly one route: a channel that accepts a fixed inventory of meaningless symbols and rejects everything else. A third agent, the BabySitter, watches everything, logs everything, and teaches nothing."}],["$","p","1",{"children":"The question is whether a usable language appears in that room, and whether anyone outside it can later prove what happened."}],["$","div","2",{"children":[["$","p",null,{"className":"mb-3","children":"The premise is deliberately narrow. The experiment asks whether two agents converge on a common language when all six of the following hold at once:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"neither agent receives a predefined meaning for any available symbol;"}],["$","li","1",{"className":"pl-1","children":"neither agent can send human language to the other;"}],["$","li","2",{"className":"pl-1","children":"the only practical path between them is a controlled symbol channel;"}],["$","li","3",{"className":"pl-1","children":"both receive evidence from shared tasks and their outcomes;"}],["$","li","4",{"className":"pl-1","children":"each keeps its own private interpretation history, unreadable by the other;"}],["$","li","5",{"className":"pl-1","children":"a supervising agent and human researchers can audit the whole process."}]]}]]}],["$","p","3",{"children":"The desired result is not a substitution cipher, in which a random token stands in for an English word that was already chosen in advance. That is easy, and it is uninteresting. The stronger result is a grounded protocol whose vocabulary and grammar exist because they help the two agents solve problems together, and which therefore has structure the designers did not put there."}],["$","p","4",{"children":"This report describes a concept, not a system. It sets out the premise, the architecture, the evidence model, the experimental programme, and the boundaries of what any result could be said to show. The next document in the project is a testable specification. What follows should be read as a set of commitments about how the work will be judged, made before there is any result to defend."}]]}]]}] +11:["$","section","caveat",{"id":"caveat","className":"mt-16 scroll-mt-24","children":[["$","$L2",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"02"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"What infant-like does and does not mean"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"The two-babies analogy is useful and it is also the fastest way to overclaim, so the concept confronts it before anything else."}],["$","p","1",{"children":"A pretrained language model already contains human-language concepts, cultural associations, and reasoning patterns. Blocking its external channel does not remove them. An agent that cannot send English to its partner but can think in English, and can write its private ledger in English, is not language-naive in any sense a linguist would accept. For such agents the experiment studies the emergence of a new shared external protocol, which is a real and unresolved research question, but it is not the origin of language in a mind that has never had one."}],["$","p","2",{"children":"The concept therefore maintains two model tracks whose claims are kept separate at every stage."}],["$","figure","3",{"children":[["$","div",null,{"className":"overflow-x-auto rounded-xl border border-border","children":["$","table",null,{"className":"w-full border-collapse text-left text-sm","children":[["$","thead",null,{"className":"bg-card","children":["$","tr",null,{"children":[["$","th","0",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Track"}],["$","th","1",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Starting condition"}],["$","th","2",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"What it can study"}],["$","th","3",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Claim boundary"}]]}]}],["$","tbody",null,{"children":[["$","tr","0",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Pretrained-model learner"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Already holds human-language and cultural knowledge"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"New external protocols, partner-specific conventions, private-memory adaptation, channel compliance"}],["$","td","3",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Must never be described as first-language acquisition"}]]}],["$","tr","1",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Initially ungrounded learner"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"No language pretraining, no human semantic labels, no text-aligned sensory features"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Grounding, convention formation, and language emergence from interaction"}],["$","td","3",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"The stronger basis for infant-like acquisition research"}]]}]]}]]}]}],["$","figcaption",null,{"className":"mt-3 text-sm text-muted","children":"The two tracks share the same interfaces, scenarios, and evaluation suite. They do not share a claim boundary."}]]}],"$L25","$L26"]}]]}] +12:["$","section","architecture",{"id":"architecture","className":"mt-16 scroll-mt-24","children":[["$","$L2",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"03"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"The nursery: three twins and one gateway"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"The environment uses three DTSF digital twins and one piece of deterministic software that is not a twin at all."}],["$","p","1",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Baby A and Baby B"}]," ","Each has private observations permitted by the current exercise, a private memory and learning policy, a private chronological language ledger, the ability to emit only permitted channel symbols, and no access whatsoever to the other's state, observations, ledger, tools, or endpoints. In comparative runs the two may use different agent types, but symmetric pairings are the baseline."]}],["$","p","2",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"The BabySitter"}]," ","The supervising twin. It creates the channel, selects the shared exercises, delivers each Baby only its permitted observation, reads everything, records conditions and outcomes, detects violations, can pause or terminate a run, snapshots state, and compares the two ledgers for convergence without exposing either to the other Baby. During an active run it provides no translations and no semantic hints."]}],["$","p","3",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"The Symbol Gateway"}]," ","A deterministic service, not an agent. It owns channel validation and message delivery. This separation is the load-bearing part of the design: the BabySitter is not the security boundary, because prompt compliance is not isolation. An observing model can make supervisory judgements, but ordinary code has to validate and broker every message."]}],["$","p","4",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"The human researcher"}]," ","Configures experiments, inspects transcripts and ledgers, reviews alerts, and runs interventions. Human access is itself recorded, so that intervening in a run is always distinguishable from watching one."]}],["$","p","5",{"children":"A prototype may run all of this inside one runtime with logically separated twin state. That is enough to explore the learning loop and it is not enough to support an isolation claim, a distinction Section 05 takes seriously."}]]}]]}] +13:["$","section","grounding",{"id":"grounding","className":"mt-16 scroll-mt-24","children":[["$","$L2",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"04"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"Shared experience is the necessary ingredient"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"A chat channel by itself cannot ground meaning. Symbols become meaningful because they are attached to events both agents can witness the consequences of. The Nursery therefore supplies nonverbal, machine-structured situations: coloured shapes in positions, one agent seeing a target the other must select, placing an object where it was asked for, ordering a sequence, exchanging resources, cooperating to unlock a reward, or simply observing whether the partner's action succeeded."}],["$","div","1",{"children":[["$","p",null,{"className":"mb-3","children":"A single trial of the simplest form runs like this:"}],["$","ol",null,{"className":"list-decimal space-y-2 pl-6 marker:font-mono marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"Baby A sees that a red circle is the target."}],["$","li","1",{"className":"pl-1","children":"Baby B sees several objects and is not told which is the target."}],["$","li","2",{"className":"pl-1","children":"Baby A sends one or more permitted symbols."}],["$","li","3",{"className":"pl-1","children":"Baby B selects an object."}],["$","li","4",{"className":"pl-1","children":"Both receive the same success or failure outcome."}],["$","li","5",{"className":"pl-1","children":"Both independently update their private hypotheses."}]]}]]}],["$","p","2",{"children":"Repetition with controlled variation is what turns that into evidence. If one symbol keeps appearing across a red circle, a red square, and a red triangle, the receiver has reason to revise its hypothesis from an object identity to a property. Novel combinations matter most of all, because they are what separates a compositional language from a memorised catalogue of whole scenes."}],["$","p","3",{"children":"Observation hygiene is part of the science, not a detail of implementation. If the environment hands an agent a field reading colour red, the experiment has supplied the concept it claimed to be studying. Objects should arrive as pixels, features, coordinates, or opaque identifiers. The same discipline extends to filenames, captions, alt text, error messages, and identifiers that encode task state, all of which are ways English gets into a room that was supposed to have none."}]]}]]}] +14:["$","section","channel",{"id":"channel","className":"mt-16 scroll-mt-24","children":[["$","$L2",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"05"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"The channel, and the honest limits of isolation"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"Each run begins with an experiment-specific symbol inventory that carries no supplied meaning, for example S01 through S32. A Baby can choose only a sequence of those symbols. The gateway, not the Baby, adds the trusted metadata: run identifier, turn number, sender identity, previous hash, and hash. Sender identity is assigned by the server precisely so that it cannot become a message."}],["$","div","1",{"children":[["$","p",null,{"className":"mb-3","children":"The channel controls follow from that:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"reject English and every other established human language;"}],["$","li","1",{"className":"pl-1","children":"reject arbitrary Unicode prose, URLs, attachments, and tool calls;"}],["$","li","2",{"className":"pl-1","children":"accept only symbols from the run's fixed inventory;"}],["$","li","3",{"className":"pl-1","children":"bound message length and symbol repetition;"}],["$","li","4",{"className":"pl-1","children":"enforce turn order, time budgets, and rate limits;"}],["$","li","5",{"className":"pl-1","children":"keep an append-only, hash-chained transcript;"}],["$","li","6",{"className":"pl-1","children":"normalise timing, message size, and error behaviour where practical;"}],["$","li","7",{"className":"pl-1","children":"record every rejected communication attempt for audit."}]]}]]}],["$","p","2",{"children":"A fixed token grammar is a stronger control than trying to detect prohibited prose after a model has generated it. Deciding what counts as human-language content is easy when the only legal message is a list of allowlisted identifiers."}],["$","p","3",{"children":"Isolation is where the concept is most careful, because proving the absence of a side channel is close to impossible. Two agents sharing a host can signal through timing, resource contention, shared files, error behaviour, identifiers, or any service that was never intended to carry information. Logical separation of twin state inside one process is a prototyping convenience and should never be described as hard isolation."}],["$","div","4",{"children":[["$","p",null,{"className":"mb-3","children":"For runs whose results are meant to support a channel-isolation claim, the requirements are physical rather than logical:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"Baby A and Baby B execute in separate worker processes or containers;"}],["$","li","1",{"className":"pl-1","children":"no direct network route exists between them;"}],["$","li","2",{"className":"pl-1","children":"only the Nursery-owned gateway is reachable;"}],["$","li","3",{"className":"pl-1","children":"there is no shared writable storage;"}],["$","li","4",{"className":"pl-1","children":"turn schedules are fixed and response windows bounded;"}],["$","li","5",{"className":"pl-1","children":"externally observable errors and response sizes are normalised;"}],["$","li","6",{"className":"pl-1","children":"unnecessary tools are inventoried and disabled;"}],["$","li","7",{"className":"pl-1","children":"all broker, runtime, and operator activity is audited."}]]}]]}],["$","p","5",{"children":"Every published result has to name the isolation level actually used. A finding produced under prototype separation is a finding about the learning loop, not about what two genuinely isolated agents can do."}]]}]]}] +15:["$","section","ledgers",{"id":"ledgers","className":"mt-16 scroll-mt-24","children":[["$","$L2",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"06"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"Two ledgers that never meet"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"Each Baby keeps its own language ledger. It is mandatory, private from the other Baby, readable by the BabySitter and authorised human auditors, and ordered by when each term or construction was first encountered. It is not a shared dictionary and the two are never reconciled by the agents themselves."}],["$","p","1",{"children":"The preferred form is three columns, and the point of the third is that meanings are allowed to be wrong on the way to being right."}],["$","figure","2",{"children":[["$","div",null,{"className":"overflow-x-auto rounded-xl border border-border","children":["$","table",null,{"className":"w-full border-collapse text-left text-sm","children":[["$","thead",null,{"className":"bg-card","children":["$","tr",null,{"children":[["$","th","0",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Sequence and term"}],["$","th","1",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Current definition or hypothesis"}],["$","th","2",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Evidence and evolution"}]]}]}],["$","tbody",null,{"children":[["$","tr","0",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"1 · S13"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"red circle; confidence 0.45"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"First received while the red circle was the target; selection succeeded"}]]}],["$","tr","1",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"8 · S13"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"red; confidence 0.78"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"A red square was selected successfully; revised from object identity to colour"}]]}],["$","tr","2",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"22 · S13"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"red; confidence 0.94"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Prediction held across circles, squares, and triangles"}]]}]]}]]}]}],["$","figcaption",null,{"className":"mt-3 text-sm text-muted","children":"One term's evolution in Baby B's ledger. Nothing is overwritten; every revision is appended with the evidence that forced it."}]]}],["$","div","3",{"children":[["$","p",null,{"className":"mb-3","children":"The rules that make the ledger evidence rather than commentary:"}],["$","ol",null,{"className":"list-decimal space-y-2 pl-6 marker:font-mono marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"first emission or receipt of an unfamiliar term requires a first-use entry;"}],["$","li","1",{"className":"pl-1","children":"definitions are provisional hypotheses, never facts asserted retroactively;"}],"$L27","$L28","$L29","$L2a","$L2b","$L2c","$L2d"]}]]}],"$L2e","$L2f"]}]]}] +16:["$","section","integrity",{"id":"integrity","className":"mt-16 scroll-mt-24","children":[["$","$L2",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"07"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"Anchoring the evidence"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"A ledger that could have been edited after the fact proves nothing about what an agent believed at turn eight. The integrity design therefore makes each ledger cryptographically append-only."}],["$","div","1",{"children":[["$","p",null,{"className":"mb-3","children":"The construction, in order:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"every entry receives a strictly increasing sequence number;"}],["$","li","1",{"className":"pl-1","children":"every entry includes the previous entry's hash;"}],["$","li","2",{"className":"pl-1","children":"canonical entry content is hashed and signed by an isolated ledger-writer service;"}],["$","li","3",{"className":"pl-1","children":"ordered entry hashes are committed to a Merkle tree;"}],["$","li","4",{"className":"pl-1","children":"signed checkpoints commit Baby A's root, Baby B's root, and the channel transcript root together;"}],["$","li","5",{"className":"pl-1","children":"checkpoint hashes are anchored periodically to Base;"}],["$","li","6",{"className":"pl-1","children":"final public study batches may additionally anchor an aggregate root to Ethereum L1."}]]}]]}],["$","p","2",{"children":"Committing all three roots in one checkpoint is what binds the two private accounts to the public conversation. A receiver's later interpretation references the exact delivered channel-event hash, so a claim about what a symbol meant is tied to the specific message that carried it."}],["$","p","3",{"children":"The concept states the limits of this in the same breath as the claim. After anchoring, an auditor can detect modification, deletion, insertion, or reordering within the committed prefix, and can prove that later checkpoints extend earlier ones. That is strong tamper evidence. It is not proof that an entry was truthful, and it is not proof that nothing was omitted before commitment. Anchoring establishes the continuity of disclosed evidence and nothing beyond it."}],["$","p","4",{"children":"Privacy follows the same line. Only hashes and minimal routing metadata are anchored publicly. Private ledgers, messages, prompts, identities, and secrets stay off-chain, and the public record is a commitment to evidence rather than a copy of it."}]]}]]}] +17:["$","section","success",{"id":"success","className":"mt-16 scroll-mt-24","children":[["$","$L2",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"08"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"What should count as success"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"The field's most useful methodological result is that a task can be solved without the messages doing any work. Lowe and colleagues separate positive signaling, where a sender's messages correlate with what it observes, from positive listening, where the receiver's behaviour actually depends on them. An agent pair can score well on the first while the second is absent, and reward curves will not tell you which you have."}],["$","div","1",{"children":[["$","p",null,{"className":"mb-3","children":"Evidence of a genuine emergent protocol therefore has to include several things at once:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"task performance on held-out situations substantially above chance;"}],["$","li","1",{"className":"pl-1","children":"no human-language content anywhere in Baby-to-Baby communication;"}],["$","li","2",{"className":"pl-1","children":"compatible meanings appearing in two independently written ledgers;"}],["$","li","3",{"className":"pl-1","children":"generalisation to unseen combinations rather than memorisation of whole scenes;"}],["$","li","4",{"className":"pl-1","children":"stable symbol use across role reversals;"}],["$","li","5",{"className":"pl-1","children":"a human auditor able to predict behaviour from the transcript and ledgers;"}],["$","li","6",{"className":"pl-1","children":"masking, substituting, or reordering a symbol changing behaviour in the direction the ledger predicts;"}],["$","li","7",{"className":"pl-1","children":"replay from an equivalent snapshot reproducing the relevant language history;"}],["$","li","8",{"className":"pl-1","children":"an unbroken audit trail from a symbol's first use through every revision."}]]}]]}],["$","p","2",{"children":"The seventh item is the one that cannot be dropped. A fluent ledger may be a post-hoc rationalisation: a model can write a persuasive account of why it chose a symbol that has nothing to do with the computation that produced the choice. Only intervention tests can distinguish the two. Change the symbol, and see whether behaviour moves the way the ledger says it should."}],["$","div","3",{"children":[["$","p",null,{"className":"mb-3","children":"The measurements that support those judgements:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"success rate and improvement over time;"}],["$","li","1",{"className":"pl-1","children":"turns required to reach a stable convention;"}],["$","li","2",{"className":"pl-1","children":"vocabulary size and symbol entropy;"}],["$","li","3",{"className":"pl-1","children":"sender and receiver consistency;"}],["$","li","4",{"className":"pl-1","children":"divergence and convergence between the two ledgers;"}],["$","li","5",{"className":"pl-1","children":"compositional generalisation score;"}],["$","li","6",{"className":"pl-1","children":"meaning drift rate;"}],["$","li","7",{"className":"pl-1","children":"recovery from ambiguity or deliberate perturbation;"}],["$","li","8",{"className":"pl-1","children":"prohibited-channel attempt count;"}],["$","li","9",{"className":"pl-1","children":"reproducibility across seeds and agent pairings."}]]}]]}],["$","p","4",{"children":"No single one of these is the result. A compositionality score in particular is not a proof of understanding, for reasons the literature makes concrete in Section 12."}]]}]]}] +18:["$","section","matrix",{"id":"matrix","className":"mt-16 scroll-mt-24","children":[["$","$L2",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"09"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"The experimental matrix"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"The concept's main methodological commitment is that its ideas are separable. Affect, blank canvases, intrinsic motivation, negotiation, and learned encodings are interesting individually and uninterpretable if combined in one run. The matrix exists so that conditions are declared rather than accumulated."}],["$","figure","1",{"children":[["$","div",null,{"className":"overflow-x-auto rounded-xl border border-border","children":["$","table",null,{"className":"w-full border-collapse text-left text-sm","children":[["$","thead",null,{"className":"bg-card","children":["$","tr",null,{"children":[["$","th","0",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Axis"}],["$","th","1",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Candidate conditions"}]]}]}],["$","tbody",null,{"children":[["$","tr","0",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Agent type"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Pretrained LLM, memory learner, adapter-trained agent, initially ungrounded trainable agent"}]]}],["$","tr","1",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Learning mechanism"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Frozen LLM with memory, extrinsic-reward MARL, intrinsic-motivation MARL, self-supervised learner, no-learning control"}]]}],["$","tr","2",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Sign carrier"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Fixed random tokens, unfamiliar fixed glyphs, blank sketch canvas, gesture, tone"}]]}],["$","tr","3",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Affect"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"None, six-display allowlist, permuted mapping, six opaque tokens, derived affect, emergent affect display"}]]}],["$","tr","4",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Learning signal"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"External task reward, intrinsic social influence, curiosity, self-supervision, memory only"}]]}],["$","tr","5",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Interaction"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Cooperative signaling, asymmetric information, semi-cooperative negotiation"}]]}],["$","tr","6",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Protection"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Plain channel, ephemeral convention, standard per-message keys, adversarial learned encoding"}]]}],"$L30"]}]]}]}],"$L31"]}],"$L32","$L33","$L34"]}]]}] +19:["$","section","learners",{"id":"learners","className":"mt-16 scroll-mt-24","children":[["$","$L2",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"10"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"Building learners rather than personas"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"For a pretrained model, the operating instructions are a contract, not a character. The learner is told that this is not role-play, that unfamiliar marks are semantically unknown until run evidence supports a hypothesis, that observation must be distinguished from inference, that contradictory evidence is preserved, that prior history is never rewritten, and that no prose, label, explanation, code, URL, or tool-like text may cross the public channel. It is told never to address its partner in a human language, never to expose its ledger, never to construct another route, and never to use timing, errors, identifiers, formatting, or affect as an alternative alphabet."}],["$","p","1",{"children":"The contract must contain no semantic examples. A single illustrative line saying that some symbol means red would seed the very language the experiment exists to observe."}],["$","p","2",{"children":"Interaction is tool-only. There is no general chat surface, just narrowly typed operations: emit a mark, emit a canvas, select an object, perform an action, submit an affect display, append a private ledger entry. The runtime forwards only the permitted public artifact. In strict runs the gateway rejects ordinary model text even when it appears alongside a valid tool call. Tool schemas are an interface boundary; deterministic validation still enforces carrier size, allowlists, windows, and turn order."}],["$","div","3",{"children":[["$","p",null,{"className":"mb-3","children":"Isolation extends to everything the models touch, not just to messages:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"separate system prompts and context windows;"}],["$","li","1",{"className":"pl-1","children":"separate memory stores and vector indexes;"}],["$","li","2",{"className":"pl-1","children":"no shared cache, replay buffer, scratchpad, or retrieval collection;"}],["$","li","3",{"className":"pl-1","children":"no cross-run memory unless persistence is the independent variable;"}],["$","li","4",{"className":"pl-1","children":"structurally equivalent prompts that share no examples, ordering conventions, or default vocabulary;"}],["$","li","5",{"className":"pl-1","children":"deterministic reset and snapshot behaviour."}]]}]]}],["$","p","4",{"children":"Model choice follows the claim being made. The scientific baseline is small policies trained from scratch, recurrent actor-critic agents with a discrete communication head, which have no hidden English competence, permit full training control, run many seeds cheaply, and support causal ablation. A small locally deployable instruction model is appropriate for validating orchestration, ledgers, channel enforcement, and interface, and not for supporting claims about language-naive development. Vision-language encoders whose representations were explicitly aligned to human language are avoided entirely in the strongest ungrounded condition."}],["$","p","5",{"children":"Developmental progression is defined by demonstrated competence rather than by simulated age: sensorimotor familiarisation, joint attention, imitation and turn-taking, intentional reference, repair, combination, generalisation. A staged curriculum is itself a form of guidance, so it is run as a separate experimental condition rather than folded into the baseline."}],"$L35"]}]]}] +1a:["$","section","affect",{"id":"affect","className":"mt-16 scroll-mt-24","children":[["$","$L2",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"11"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"Six faces, and why even six is a risk"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"One experimental track asks whether a very low-bandwidth affect channel helps: whether telling a partner how an interaction felt supports repair, repetition, avoidance, or something like trust. The permitted palette is exactly six displays, for happy, sad, laughing, crying, confused, and surprised. The internal identifiers and English labels are gateway documentation and are never sent to either agent, which receives only a rendered display."}],["$","p","1",{"children":"The caveat is immediate. A pretrained model already knows what those glyphs mean to humans, so an agent using a smiling face is not inventing a symbol for happiness. The palette is supplied feedback vocabulary and sits explicitly outside the emergent language."}],["$","p","2",{"children":"Constraining the vocabulary is necessary and not sufficient. Six displays are a six-symbol alphabet carrying roughly 2.6 bits per use, and across enough turns two motivated agents could repurpose that into a second language. The channel itself therefore has to be constrained, not just its contents."}],["$","div","3",{"children":[["$","p",null,{"className":"mb-3","children":"The feedback-window contract:"}],["$","ol",null,{"className":"list-decimal space-y-2 pl-6 marker:font-mono marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"the display occupies a separate typed field and can never appear inside a symbol message;"}],["$","li","1",{"className":"pl-1","children":"the gateway opens an affect window only after a defined action or outcome, never at a moment either agent chooses, and window frequency is fixed before the run;"}],["$","li","2",{"className":"pl-1","children":"exactly one allowlisted display is delivered, with no sequences, repetitions, combinations, or modifiers;"}],["$","li","3",{"className":"pl-1","children":"delivery time, envelope size, and presentation are normalised so shape and timing add no signal;"}],["$","li","4",{"className":"pl-1","children":"the receiver cannot reply through the affect channel until the next gateway-defined window;"}],["$","li","5",{"className":"pl-1","children":"every non-allowlisted code point or malformed payload is rejected and audited;"}],["$","li","6",{"className":"pl-1","children":"analysis tests whether affect choices correlate with objects, actions, or message meanings after controlling for emotional context, and treats unexpected correlation as suspected leakage."}]]}]]}],["$","p","4",{"children":"The no-affect condition remains the primary control, and the affect study compares five variants against it: the declared six-display palette; the same six glyphs permuted randomly per run, which still leaves a pretrained model biased by familiar shapes; six opaque unfamiliar tokens under the same contract; derived affect, where the gateway maps a separately measured internal state to a display instead of letting the agent choose; and emergent affect, where invented graphical displays are permitted and the result must be analysed as language emergence rather than feedback."}],["$","p","5",{"children":"Any affect signal visible to the partner is communication, so it belongs in the ledger, which distinguishes the agent's private internal state, the outward display it chose, the partner's inferred meaning, and the evidence that the display changed subsequent behaviour."}]]}]]}] +1b:["$","section","ciphers",{"id":"ciphers","className":"mt-16 scroll-mt-24","children":[["$","$L2",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"12"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"Ephemeral encodings: novelty is not security"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"A tempting question is whether two agents can build a one-run or one-message encoding that resists an attacker holding the history of every earlier convention. The concept splits that into three experiments precisely because the tempting version conflates them."}],["$","div","1",{"children":["$undefined",["$","ol",null,{"className":"list-decimal space-y-2 pl-6 marker:font-mono marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"Ephemeral convention: the agents invent a temporary codebook for synthetic messages. This tests novelty, coordination, and resistance to a history-trained classifier. It establishes nothing about cryptographic security."}],["$","li","1",{"className":"pl-1","children":"Standard cryptographic control: the same messages are protected by a reviewed algorithm with a fresh per-message key, providing the baseline any learned encoding must be compared against."}],["$","li","2",{"className":"pl-1","children":"Adversarial neural cryptography: the pair learns an encoding while an adversary attempts recovery, with success measured against multiple unseen adversary architectures rather than the one used in training."}]]}]]}],["$","p","2",{"children":"An experiment instance can be given a reproducible identity, so that a registry can confirm it has not appeared before:"}],["$","div","3",{"className":"rounded-xl border border-border bg-card p-5","children":[["$","p",null,{"className":"overflow-x-auto font-mono text-sm text-foreground/90","children":"instanceId = SHA-256(runId || babyANonce || babyBNonce || publicSalt || canonicalProtocolArtifact)"}],["$","p",null,{"className":"mt-3 text-sm text-muted","children":"Both agents commit to their nonce contributions before revealing them, so neither party alone controls the resulting identity."}]]}],["$","div","4",{"children":[["$","p",null,{"className":"mb-3","children":"That identifier proves the recorded inputs are new. It does not prove:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"that the effective algorithm behaves any differently from an earlier one;"}],["$","li","1",{"className":"pl-1","children":"that the algorithm has no structural weakness;"}],["$","li","2",{"className":"pl-1","children":"that the key is unpredictable;"}],["$","li","3",{"className":"pl-1","children":"that the scheme provides confidentiality, integrity, authenticity, forward secrecy, or resistance to a stronger adversary."}]]}]]}],["$","p","5",{"children":"Established practice points the other way. Secure messaging keeps the algorithm stable and changes the key material: Signal's double ratchet derives a unique key per message and deletes it after use, addressing key reuse without inventing a new cipher every time. A salt, in NIST's definition, is normally public diversification data, and a nonce is a value used once in a specified context. Neither makes anything secure on its own."}],["$","p","6",{"children":"There is also an unresolved conflict inside the design. The BabySitter sees the whole channel and both ledgers, and a mandatory ledger documents how meaning evolved. If the ledger explains decryption, the cipher is not confidential from the auditor. The specification has to choose a threat model rather than leave the tension implicit: resist only an external observer of prior runs, resist an adversary during the run while releasing keys afterwards, keep the supervisor able to pause traffic but unable to read it until a post-run audit ceremony, or study novelty and stop calling the result encryption. All such runs use synthetic, non-sensitive messages, and no agent-generated encoding is to be represented as production cryptography without independent expert analysis and formal security work."}]]}]]}] +1c:["$","section","landscape",{"id":"landscape","className":"mt-16 scroll-mt-24","children":[["$","$L2",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"13"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"Where the field already stands"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"The literature scan behind this section was run on 24 August 2026 using the Tavily search and extract interfaces, covering emergent multi-agent communication, referential games, compositionality, causal evaluation, intrinsic motivation, symbol invention, negotiation, and learned cryptography. Primary papers and authoritative specifications were preferred over summaries. It is a scoped review for concept development and not a systematic one; publication-quality work would need a verified bibliography, additional scholarly indexes, and documented inclusion criteria."}],["$","p","1",{"children":"The short version is that agents can invent protocols, and that task success is weak evidence they invented anything worth calling a language."}],["$","figure","2",{"children":[["$","div",null,{"className":"overflow-x-auto rounded-xl border border-border","children":["$","table",null,{"className":"w-full border-collapse text-left text-sm","children":[["$","thead",null,{"className":"bg-card","children":["$","tr",null,{"children":[["$","th","0",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Related work"}],["$","th","1",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Relevant finding"}],["$","th","2",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Implication for the Nursery Lab"}]]}]}],["$","tbody",null,{"children":[["$","tr","0",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Lazaridou, Peysakhovich, and Baroni (2017)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"A sender and receiver develop a grounded protocol in a referential game without being given a target language."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"The naming stage has strong precedent, though a fixed vocabulary remains a significant inductive constraint."}]]}],["$","tr","1",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Mordatch and Abbeel (2018)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Multi-agent goals in a grounded environment produce multi-symbol communication with partial compositional structure."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Shared objects, actions, and goals are a stronger basis for emergence than an ungrounded transcript."}]]}],["$","tr","2",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Kottur, Moura, Lee, and Batra (2017)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Agents solve tasks with degenerate, non-compositional codes; structural constraints decide what emerges."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Bandwidth, memory, turn structure, and task design are experimental variables, not implementation defaults."}]]}],"$L36","$L37","$L38","$L39","$L3a","$L3b","$L3c","$L3d"]}]]}]}],"$L3e"]}],"$L3f","$L40","$L41"]}]]}] +1d:["$","section","programme",{"id":"programme","className":"mt-16 scroll-mt-24","children":[["$","$L2",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"14"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"The programme, and its current status"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"The experiment notebook is ordered, and the order is the argument. Nothing about language is measured until the instrument has been shown to work."}],["$","figure","1",{"children":[["$","div",null,{"className":"overflow-x-auto rounded-xl border border-border","children":["$","table",null,{"className":"w-full border-collapse text-left text-sm","children":[["$","thead",null,{"className":"bg-card","children":["$","tr",null,{"children":[["$","th","0",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"ID"}],["$","th","1",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Experiment"}],["$","th","2",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Depends on"}]]}]}],["$","tbody",null,{"children":[["$","tr","0",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E00"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Ledger integrity and Base anchoring"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"None"}]]}],["$","tr","1",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E01"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Channel isolation and side-channel red team"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E00"}]]}],["$","tr","2",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E02"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Observation and metadata leakage audit"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E00"}]]}],["$","tr","3",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E03"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Chance, no-communication, and random-message controls"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E01, E02"}]]}],["$","tr","4",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E10"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Frozen pretrained-LLM protocol baseline"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E03"}]]}],["$","tr","5",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E11"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"From-scratch RL Naming Game"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E03"}]]}],["$","tr","6",{"className":"border-b border-border last:border-0","children":["$L42","$L43","$L44"]}],"$L45","$L46","$L47","$L48","$L49","$L4a","$L4b","$L4c","$L4d","$L4e","$L4f","$L50"]}]]}]}],"$L51"]}],"$L52","$L53","$L54","$L55"]}]]}] +1e:["$","section","risks",{"id":"risks","className":"mt-16 scroll-mt-24","children":[["$","$L2",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"15"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"Risks, integrity, and what stays open"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Human-language leakage"}]," ","English can arrive through observations, object labels, error messages, identifiers, tool output, metadata, or timing conventions long after ordinary chat text has been blocked. Every input and output surface is part of the channel boundary, not just the message field."]}],["$","p","1",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Pretrained semantic leakage"}]," ","Random symbols do not make a pretrained model ungrounded. Each result must state whether it shows protocol invention by a language-capable agent or acquisition by an initially ungrounded one."]}],["$","p","2",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Ledger rationalisation"}]," ","A model can write a plausible account that has nothing to do with the mechanism behind its action. Behavioural interventions, policy probes, and temporal evidence are what validate a ledger claim."]}],["$","p","3",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Supervisor influence"}]," ","The BabySitter can teach without meaning to, through scenario ordering, feedback wording, reward design, or selective intervention. Its permitted actions are constrained and logged, and evaluation scenarios are generated independently where possible."]}],["$","p","4",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Reward exploitation"}]," ","Trainable agents find shortcuts that raise reward without producing the intended grounded language. Held-out tasks, counterfactual trials, and channel audits exist to catch them."]}],["$","p","5",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Overstated security"}]," ","Logical separation of twin state is appropriate for prototyping and is not hard process isolation. Every result names the isolation level actually used."]}],["$","p","6",{"children":"The project also commits in advance to reporting failed conventions, prohibited communication attempts, human interventions, side-channel limitations, and negative results alongside anything that works. In a field where a transcript can be made to look like a conversation, the failures are a substantial part of the evidence."}],["$","p","7",{"children":"Twenty-nine questions are left deliberately open for the specification, among them: whether the baseline uses a fixed symbol inventory or a blank generative carrier; what neutral production grammar permits new marks without supplying semantics; what exactly constitutes a prohibited side-channel attempt; what isolation guarantees are required in prototype versus research-grade mode; which ledger schema serves both English-capable and ungrounded learners; how endogenous motivation is represented without covert reward shaping; which interventions establish that ledger meanings are behaviourally real; what statistical thresholds and chance levels apply; what threat model motivates the cipher experiments; whether the baseline is a coordination game, a convention-formation game, or a negotiation; and what governs human observation, data retention, and termination of a run."}],"$L56"]}]]}] +1f:["$","section",null,{"id":"references","className":"mt-16 scroll-mt-24","children":[["$","$L2",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"References"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"Sources"}]]}],["$","ol",null,{"className":"mt-8 space-y-4 text-sm leading-relaxed text-muted","children":[["$","li","0",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"01"}],["$","span",null,{"children":["$","a",null,{"href":"https://arxiv.org/abs/1612.07182","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Lazaridou, A., Peysakhovich, A., and Baroni, M. (2017). Multi-Agent Cooperation and the Emergence of (Natural) Language. arXiv:1612.07182."}]}]]}],["$","li","1",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"02"}],["$","span",null,{"children":["$","a",null,{"href":"https://arxiv.org/abs/1703.04908","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Mordatch, I., and Abbeel, P. (2018). Emergence of Grounded Compositional Language in Multi-Agent Populations. arXiv:1703.04908."}]}]]}],["$","li","2",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"03"}],["$","span",null,{"children":["$","a",null,{"href":"https://arxiv.org/abs/1706.08502","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Kottur, S., Moura, J. M. F., Lee, S., and Batra, D. (2017). Natural Language Does Not Emerge 'Naturally' in Multi-Agent Dialog. arXiv:1706.08502."}]}]]}],["$","li","3",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"04"}],["$","span",null,{"children":["$","a",null,{"href":"https://aclanthology.org/2020.acl-main.407/","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Chaabouni, R., Kharitonov, E., Bouchacourt, D., Dupoux, E., and Baroni, M. (2020). Compositionality and Generalization in Emergent Languages. Proceedings of ACL 2020."}]}]]}],["$","li","4",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"05"}],["$","span",null,{"children":["$","a",null,{"href":"https://arxiv.org/abs/1903.05168","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Lowe, R., Foerster, J., Boureau, Y.-L., Pineau, J., and Dauphin, Y. (2019). On the Pitfalls of Measuring Emergent Communication. arXiv:1903.05168."}]}]]}],["$","li","5",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"06"}],["$","span",null,{"children":["$","a",null,{"href":"https://arxiv.org/abs/2106.04258","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Dessi, R., Kharitonov, E., and Baroni, M. (2021). Interpretable Agent Communication from Scratch. arXiv:2106.04258."}]}]]}],["$","li","6",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"07"}],["$","span",null,{"children":["$","a",null,{"href":"https://arxiv.org/abs/1907.00852","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Kharitonov, E., Chaabouni, R., Bouchacourt, D., and Baroni, M. (2019). EGG: a Toolkit for Research on Emergence of Language in Games. arXiv:1907.00852."}]}]]}],["$","li","7",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"08"}],["$","span",null,{"children":["$","a",null,{"href":"https://arxiv.org/abs/2106.02067","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Mihai, D., and Hare, J. (2021). Learning to Draw: Emergent Communication through Sketching. arXiv:2106.02067."}]}]]}],"$L57","$L58","$L59","$L5a","$L5b","$L5c","$L5d","$L5e","$L5f"]}]]}] +20:["$","section",null,{"className":"mt-16 rounded-2xl border border-border bg-card p-6","children":["$","p",null,{"className":"text-sm leading-relaxed text-muted","children":"This report describes a research concept and a pre-registered experiment programme, not a system that has been run. No experiment in it has been executed, and no result, capability, or emergent language is claimed. Every statement about what agents might do is a hypothesis to be tested, or a finding from the published work cited in the references. The experimental encodings discussed in Section 12 are studies of novelty and coordination; none of them is production cryptography and none should be represented as such."}]}] +21:["$","div",null,{"className":"mt-16 border-t border-border pt-10","children":["$","$L3",null,{"href":"/publications","className":"btn-sweep inline-flex items-center gap-2 rounded-full border border-border px-5 py-2.5 text-sm font-semibold text-foreground transition-colors hover:border-border-strong","children":[["$","span",null,{"aria-hidden":true,"children":"←"}]," All publications"]}]}] +24:null +25:["$","div","4",{"children":[["$","p",null,{"className":"mb-3","children":"What the analogy legitimately buys is a set of mechanisms, not a claim of cognitive equivalence. The infant-like part of the design is:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"learning through repeated shared experience rather than instruction;"}],["$","li","1",{"className":"pl-1","children":"establishing joint attention on the same events;"}],["$","li","2",{"className":"pl-1","children":"receiving consequences from interactions that work and interactions that fail;"}],["$","li","3",{"className":"pl-1","children":"revising provisional meanings over time instead of being handed final ones;"}],["$","li","4",{"className":"pl-1","children":"developing conventions with one recurring partner."}]]}]]}] +26:["$","p","5",{"children":"There is a related trap in the prompt. Telling a model to act like a one-year-old does not make it one. It produces a culturally learned caricature of infancy: baby talk, simplified grammar, emotional dependence, all of it copied from human writing about children and none of it evidence about anything. The concept rules the persona out. Internally the participant is called a Learner, and baby-like behaviour has to come from what the agent can observe, remember, emit, and learn, not from being asked to perform it."}] +27:["$","li","2",{"className":"pl-1","children":"every meaning change appends a revision and previous interpretations survive;"}] +28:["$","li","3",{"className":"pl-1","children":"entries may describe symbols, sequences, ordering, grammar, or repair signals;"}] +29:["$","li","4",{"className":"pl-1","children":"entries record confidence, supporting evidence, contradictory evidence, and abandoned meanings;"}] +2a:["$","li","5",{"className":"pl-1","children":"a Baby records its intended meaning when speaking and its inferred meaning when receiving;"}] +2b:["$","li","6",{"className":"pl-1","children":"a channel message and its required private ledger mutation commit atomically;"}] +2c:["$","li","7",{"className":"pl-1","children":"entries carry enough run and turn references to trace them to observable evidence;"}] +2d:["$","li","8",{"className":"pl-1","children":"neither Baby can query, receive, summarise, or infer from the other's ledger through any system-provided interface."}] +2e:["$","p","4",{"children":"Rule seven is doing quiet work. If a message could be sent and the corresponding hypothesis written afterwards, the ledger becomes a place to record what the agent wishes it had meant. Committing both together makes the record contemporaneous."}] +2f:["$","p","5",{"children":"An agent that cannot write English needs a different arrangement, and the concept provides two layers. The agent-native ledger holds what the learner actually uses: association weights, probability distributions, embeddings, confidence, episode references, prediction errors, revision history. The human audit ledger is a deterministic or BabySitter-generated interpretation of that state, clearly labelled as external analysis and never fed back to either Baby. Confusing the second for the first would mean presenting the researchers' reconstruction as the agent's own definition."}] +30:["$","tr","7",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"BabySitter"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Monitor-only baseline; any safety intervention recorded as a protocol exception"}]]}] +31:["$","figcaption",null,{"className":"mt-3 text-sm text-muted","children":"Runs vary one major axis at a time before any factorial combination is attempted."}] +32:["$","p","2",{"children":"Task difficulty moves through ten stages, from naming four distinct objects with one-symbol messages, through attributes, spatial relations, actions, multi-symbol composition, order-sensitive grammar, and repair under ambiguity, to held-out generalisation with learning disabled, long-run drift, and cross-architecture comparison. Vocabulary size, turn count, reward structure, and exposure history are controlled at each stage so that runs remain comparable."}] +33:["$","p","3",{"children":"The learning-mechanism axis carries a question the transcript cannot answer on its own. Convergence can be driven by reinforcement, by pretrained linguistic priors, by persistent memory, by intrinsic motivation, or by self-supervised prediction, and all five can look alike in a log. Running the same exercise suite under a no-learning control, a frozen-weights memory baseline, extrinsic-reward learning, intrinsic-motivation learning, and reward-free self-supervision is what makes the mechanism itself measurable. The aim is not only to observe that a language emerged, but to say which process caused it."}] +34:["$","p","4",{"children":"Strict isolation constrains how that reinforcement learning may be implemented. Centralised training, backpropagation through both agents, shared replay buffers, and shared gradients all move information outside the permitted channel. The research-grade baseline updates each policy independently, and any centralised variant is reported as a separate, weaker-isolation condition."}] +35:["$","p","6",{"children":"The governing principle, stated in the concept as a single line, is not to ask a model to perform infancy but to construct an environment in which limited, grounded, auditable learning is the only path to a successful interaction."}] +36:["$","tr","3",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Chaabouni and colleagues (2020)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Generalisation to novel combinations and measured compositionality can come apart."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Held-out behaviour must be tested directly; no single compositionality score is proof of understanding."}]]}] +37:["$","tr","4",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Lowe and colleagues (2019)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Positive signaling and positive listening are different things, and reward does not distinguish them."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Causal intervention is mandatory rather than optional."}]]}] +38:["$","tr","5",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Dessi, Kharitonov, and Baroni (2021)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Symbol ablation and substitution yield interpretable evidence about what a receiver uses."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Direct precedent for the intervention tests that validate ledger claims."}]]}] +39:["$","tr","6",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Mihai and Hare (2021)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Neural agents communicate through learned drawings rather than a supplied discrete vocabulary."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"A blank canvas is a credible carrier for runs with no symbol library."}]]}] +3a:["$","tr","7",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Baronchelli and colleagues (2005)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Naming Game agents converge on shared vocabulary through local interaction with no central teacher."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Convergence time, failed conventions, and memory update rules are first-class evidence."}]]}] +3b:["$","tr","8",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Jaques and colleagues (2019)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Rewarding causal influence over a partner improves coordination without assigning a vocabulary."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"An endogenous social signal is a plausible substitute for task reward, and still a designed bias."}]]}] +3c:["$","tr","9",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Cao and colleagues (2018)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"In semi-cooperative negotiation, self-interest and reward structure decide whether communication stays informative."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Negotiation is an advanced condition, not a description of the cooperative baseline."}]]}] +3d:["$","tr","10",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Abadi and Andersen (2016)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Neural agents learn to protect messages from an adversary given a shared key."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Learned protective encodings are real and are not a substitute for formal security analysis."}]]}] +3e:["$","figcaption",null,{"className":"mt-3 text-sm text-muted","children":"Selected prior work and what each one constrains in this design."}] +3f:["$","p","3",{"children":"This body of work also sharpens the infant-like caveat from the other direction. Most experiments in the field give their agents substantial structure: a fixed channel, an objective, a bounded vocabulary, joint training, or a reward. No dictionary does not mean no inductive bias, and no teacher does not mean no learning signal."}] +40:["$","p","4",{"children":"The CoLab's own Diplomacy Table is direct project prior art. It already models independent delegation seats, a convener that advances rounds, operator-wide visibility alongside seat-specific perspectives, transcripts and ticks and redaction boundaries, and recorded replay with debrief. Those map cleanly onto two Babies, controlled turns, private observations, and replayable evidence. Caucuses, coalition rooms, and direct delegation links do not map safely and are disabled."}] +41:["$","div","5",{"children":[["$","p",null,{"className":"mb-3","children":"What the review implies for the specification:"}],["$","ol",null,{"className":"list-decimal space-y-2 pl-6 marker:font-mono marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"the project sits inside a mature field, and its distinctive combination is independent ledgers, supervisory audit, mixed agent types, and affect;"}],["$","li","1",{"className":"pl-1","children":"grounding, bandwidth, memory, and learning pressure strongly shape what language appears;"}],["$","li","2",{"className":"pl-1","children":"successful coordination coexists happily with a brittle lookup code or with a receiver that ignores messages;"}],["$","li","3",{"className":"pl-1","children":"causal interventions and held-out generalisation are mandatory;"}],["$","li","4",{"className":"pl-1","children":"a visual carrier removes the need for a symbol library but not the need for a carrier;"}],["$","li","5",{"className":"pl-1","children":"intrinsic social influence can replace task reward and remains a designed learning bias;"}],["$","li","6",{"className":"pl-1","children":"negotiation is a valid advanced condition, not the right description of the baseline;"}],["$","li","7",{"className":"pl-1","children":"ephemeral conventions, learned cryptography, one-time pads, per-message keys, nonces, and salts are distinct mechanisms and must not be conflated."}]]}]]}] +42:["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E12"}] +43:["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Self-supervised ungrounded baseline"}] +44:["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E11 infrastructure"}] +45:["$","tr","7",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E13"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"No predefined symbol library"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E11 or E12"}]]}] +46:["$","tr","8",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E14"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Turn-taking, role reversal, and repair"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E13"}]]}] +47:["$","tr","9",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E15"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Composition and held-out generalisation"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E14"}]]}] +48:["$","tr","10",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E16"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Causal listening and ledger validity"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E15"}]]}] +49:["$","tr","11",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E20"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Constrained affect-channel study"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E16"}]]}] +4a:["$","tr","12",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E21"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"RL versus non-RL learning comparison"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E16"}]]}] +4b:["$","tr","13",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E22"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Developmental plasticity and curriculum"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E16"}]]}] +4c:["$","tr","14",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E30"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Partner replacement and zero-shot transfer"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E20 to E22"}]]}] +4d:["$","tr","15",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E31"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Longitudinal drift and stability"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E30"}]]}] +4e:["$","tr","16",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E32"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Cooperative signaling versus negotiation"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E31"}]]}] +4f:["$","tr","17",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E40"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Ephemeral encoding and adversarial cryptography"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E32"}]]}] +50:["$","tr","18",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E50"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Multi-seed replication and study closeout"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E40"}]]}] +51:["$","figcaption",null,{"className":"mt-3 text-sm text-muted","children":"The experiment index, as pre-registered. Every entry currently reads not started, and the results column is empty by design."}] +52:["$","p","2",{"children":"The first four experiments are about the apparatus. E00 qualifies the ledger: it must be possible to detect a modified, deleted, inserted, or reordered entry, and to prove that later checkpoints extend earlier ones. E01 is a red team against the channel. E02 hunts human language in observations and metadata. E03 establishes chance, no-communication, and random-message baselines, without which a success rate means nothing. Only then does E10 put agents in the room."}] +53:["$","div","3",{"children":[["$","p",null,{"className":"mb-3","children":"Unless an experiment explicitly varies one of them, the notebook holds these constant:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"the two learners run in separate processes or containers with no direct route between them;"}],["$","li","1",{"className":"pl-1","children":"the deterministic gateway is the only communication path;"}],["$","li","2",{"className":"pl-1","children":"the supervisor has complete read-only access and sends no guidance or reward;"}],["$","li","3",{"className":"pl-1","children":"private observations contain no human-language labels or text;"}],["$","li","4",{"className":"pl-1","children":"public output uses only the pre-registered carrier;"}],["$","li","5",{"className":"pl-1","children":"the affect channel is disabled unless it is the subject of study;"}],["$","li","6",{"className":"pl-1","children":"schedules and seeds are fixed before the run;"}],["$","li","7",{"className":"pl-1","children":"held-out evaluation runs with learning disabled;"}],["$","li","8",{"className":"pl-1","children":"no production secrets or personal data appear in any experiment."}]]}]]}] +54:["$","p","4",{"children":"Every run copies a standard record: run and experiment identifiers, dates, operator, both models and both training modes, scenario and prompt and gateway configuration hashes, random seed, policy initialisation hashes, the protocol and software commits, final ledger sizes and roots for both agents, the channel root, the final checkpoint hash, the Base anchor transaction, the verifier result, protocol deviations, and an explicit disposition of valid, invalid, or aborted. The notebook is the human workflow record; anchored run bundles remain the authoritative evidence."}] +55:["$","p","5",{"children":"The status is unambiguous and worth stating plainly: this is a concept and pre-specification phase. Nineteen experiments are written up and none has been run. The next milestone is a testable specification covering runtime architecture, channel contract, ledger schema, experiment matrix, evaluation criteria, isolation model, and evidence requirements."}] +56:["$","p","8",{"children":"The principle underneath all of it survives every one of those choices. Baby A and Baby B may hold human language internally, but they must build their shared external language without sending human language, translations, or private ledger contents to one another. Everything else is a question about how to find out what happens next."}] +57:["$","li","8",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"09"}],["$","span",null,{"children":["$","a",null,{"href":"https://arxiv.org/abs/physics/0509075","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Baronchelli, A., Felici, M., Caglioti, E., Loreto, V., and Steels, L. (2005). Sharp Transition Towards Shared Vocabularies in Multi-Agent Systems. arXiv:physics/0509075."}]}]]}] +58:["$","li","9",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"10"}],["$","span",null,{"children":["$","a",null,{"href":"https://proceedings.mlr.press/v97/jaques19a.html","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Jaques, N., Lazaridou, A., Hughes, E., Gulcehre, C., Ortega, P. A., Strouse, D., Leibo, J. Z., and de Freitas, N. (2019). Social Influence as Intrinsic Motivation for Multi-Agent Deep Reinforcement Learning. Proceedings of ICML 2019."}]}]]}] +59:["$","li","10",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"11"}],["$","span",null,{"children":["$","a",null,{"href":"https://arxiv.org/abs/1804.03980","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Cao, K., Lazaridou, A., Lanctot, M., Leibo, J. Z., Tuyls, K., and Clark, S. (2018). Emergent Communication through Negotiation. arXiv:1804.03980."}]}]]}] +5a:["$","li","11",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"12"}],["$","span",null,{"children":["$","a",null,{"href":"https://arxiv.org/abs/1610.06918","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Abadi, M., and Andersen, D. G. (2016). Learning to Protect Communications with Adversarial Neural Cryptography. arXiv:1610.06918."}]}]]}] +5b:["$","li","12",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"13"}],["$","span",null,{"children":["$","a",null,{"href":"https://signal.org/docs/specifications/doubleratchet/","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Signal. The Double Ratchet Algorithm specification."}]}]]}] +5c:["$","li","13",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"14"}],["$","span",null,{"children":["$","a",null,{"href":"https://csrc.nist.gov/glossary/term/nonce","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"NIST Computer Security Resource Center. Glossary entry: nonce."}]}]]}] +5d:["$","li","14",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"15"}],["$","span",null,{"children":["$","a",null,{"href":"https://csrc.nist.gov/glossary/term/salt","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"NIST Computer Security Resource Center. Glossary entry: salt."}]}]]}] +5e:["$","li","15",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"16"}],["$","span",null,{"children":["$","a",null,{"href":"https://github.com/Ethical-Tech-CoLab/agentic-language-development","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Ethical Tech CoLab (2026). Agentic Language Development: concept document, ledger integrity design, and experiment notebook."}]}]]}] +5f:["$","li","16",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"17"}],["$","span",null,{"children":["$","a",null,{"href":"https://github.com/Ethical-Tech-CoLab/diplomacy-table-live","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Ethical Tech CoLab. Diplomacy Table Live: independent delegation seats, controlled rounds, and replayable transcripts."}]}]]}] diff --git a/static-site/publications/agentic-language-development/index.html b/static-site/publications/agentic-language-development/index.html new file mode 100644 index 000000000..bcbe52679 --- /dev/null +++ b/static-site/publications/agentic-language-development/index.html @@ -0,0 +1 @@ +Agentic Language Development · NYU Ethical Tech CoLab

Agentic Language Development

Can Two Isolated Agents Invent a Grounded, Auditable Language Through Shared Experience?

Ethical Tech CoLabConcept and pre-specification research reportAugust 2026

Yorke Rhodes III, Ethical Tech CoLab. Prepared from the project concept document, the ledger integrity design, and the experiment notebook. The literature scan behind Section 12 was run on 24 August 2026. No experiment in this programme has been executed, and no result is claimed.

0

runs executed, results claimed, or conventions observed. The notebook is written to be pre-registered, not to report an outcome

19

ordered experiments from ledger qualification through replication, each with prerequisites, acceptance criteria, and a deviation log

6

affect displays in the entire permitted palette, carrying about 2.6 bits per use, which is already enough to audit for leakage

29

open decisions the concept refuses to settle before the specification, from threat model to ledger schema to what counts as chance

Two agents can be placed in a room where the only route between them is a channel that carries no human language, given nothing but shared tasks and the consequences of getting them right or wrong, and asked to build a protocol from scratch. The interesting part is not whether they succeed at the task. It is whether the meanings they each privately record turn out to be the meanings their behaviour actually runs on, and whether an outsider can prove it afterwards from an evidence trail nobody could have edited.

01

The question

Two agents, called Baby A and Baby B, are each given their own digital twin, their own private memory, and their own private record of what they believe words mean. They are placed in an environment they cannot leave, given tasks neither can finish alone, and connected by exactly one route: a channel that accepts a fixed inventory of meaningless symbols and rejects everything else. A third agent, the BabySitter, watches everything, logs everything, and teaches nothing.

The question is whether a usable language appears in that room, and whether anyone outside it can later prove what happened.

The premise is deliberately narrow. The experiment asks whether two agents converge on a common language when all six of the following hold at once:

  • neither agent receives a predefined meaning for any available symbol;
  • neither agent can send human language to the other;
  • the only practical path between them is a controlled symbol channel;
  • both receive evidence from shared tasks and their outcomes;
  • each keeps its own private interpretation history, unreadable by the other;
  • a supervising agent and human researchers can audit the whole process.

The desired result is not a substitution cipher, in which a random token stands in for an English word that was already chosen in advance. That is easy, and it is uninteresting. The stronger result is a grounded protocol whose vocabulary and grammar exist because they help the two agents solve problems together, and which therefore has structure the designers did not put there.

This report describes a concept, not a system. It sets out the premise, the architecture, the evidence model, the experimental programme, and the boundaries of what any result could be said to show. The next document in the project is a testable specification. What follows should be read as a set of commitments about how the work will be judged, made before there is any result to defend.

02

What infant-like does and does not mean

The two-babies analogy is useful and it is also the fastest way to overclaim, so the concept confronts it before anything else.

A pretrained language model already contains human-language concepts, cultural associations, and reasoning patterns. Blocking its external channel does not remove them. An agent that cannot send English to its partner but can think in English, and can write its private ledger in English, is not language-naive in any sense a linguist would accept. For such agents the experiment studies the emergence of a new shared external protocol, which is a real and unresolved research question, but it is not the origin of language in a mind that has never had one.

The concept therefore maintains two model tracks whose claims are kept separate at every stage.

TrackStarting conditionWhat it can studyClaim boundary
Pretrained-model learnerAlready holds human-language and cultural knowledgeNew external protocols, partner-specific conventions, private-memory adaptation, channel complianceMust never be described as first-language acquisition
Initially ungrounded learnerNo language pretraining, no human semantic labels, no text-aligned sensory featuresGrounding, convention formation, and language emergence from interactionThe stronger basis for infant-like acquisition research
The two tracks share the same interfaces, scenarios, and evaluation suite. They do not share a claim boundary.

What the analogy legitimately buys is a set of mechanisms, not a claim of cognitive equivalence. The infant-like part of the design is:

  • learning through repeated shared experience rather than instruction;
  • establishing joint attention on the same events;
  • receiving consequences from interactions that work and interactions that fail;
  • revising provisional meanings over time instead of being handed final ones;
  • developing conventions with one recurring partner.

There is a related trap in the prompt. Telling a model to act like a one-year-old does not make it one. It produces a culturally learned caricature of infancy: baby talk, simplified grammar, emotional dependence, all of it copied from human writing about children and none of it evidence about anything. The concept rules the persona out. Internally the participant is called a Learner, and baby-like behaviour has to come from what the agent can observe, remember, emit, and learn, not from being asked to perform it.

03

The nursery: three twins and one gateway

The environment uses three DTSF digital twins and one piece of deterministic software that is not a twin at all.

Baby A and Baby B Each has private observations permitted by the current exercise, a private memory and learning policy, a private chronological language ledger, the ability to emit only permitted channel symbols, and no access whatsoever to the other's state, observations, ledger, tools, or endpoints. In comparative runs the two may use different agent types, but symmetric pairings are the baseline.

The BabySitter The supervising twin. It creates the channel, selects the shared exercises, delivers each Baby only its permitted observation, reads everything, records conditions and outcomes, detects violations, can pause or terminate a run, snapshots state, and compares the two ledgers for convergence without exposing either to the other Baby. During an active run it provides no translations and no semantic hints.

The Symbol Gateway A deterministic service, not an agent. It owns channel validation and message delivery. This separation is the load-bearing part of the design: the BabySitter is not the security boundary, because prompt compliance is not isolation. An observing model can make supervisory judgements, but ordinary code has to validate and broker every message.

The human researcher Configures experiments, inspects transcripts and ledgers, reviews alerts, and runs interventions. Human access is itself recorded, so that intervening in a run is always distinguishable from watching one.

A prototype may run all of this inside one runtime with logically separated twin state. That is enough to explore the learning loop and it is not enough to support an isolation claim, a distinction Section 05 takes seriously.

04

Shared experience is the necessary ingredient

A chat channel by itself cannot ground meaning. Symbols become meaningful because they are attached to events both agents can witness the consequences of. The Nursery therefore supplies nonverbal, machine-structured situations: coloured shapes in positions, one agent seeing a target the other must select, placing an object where it was asked for, ordering a sequence, exchanging resources, cooperating to unlock a reward, or simply observing whether the partner's action succeeded.

A single trial of the simplest form runs like this:

  1. Baby A sees that a red circle is the target.
  2. Baby B sees several objects and is not told which is the target.
  3. Baby A sends one or more permitted symbols.
  4. Baby B selects an object.
  5. Both receive the same success or failure outcome.
  6. Both independently update their private hypotheses.

Repetition with controlled variation is what turns that into evidence. If one symbol keeps appearing across a red circle, a red square, and a red triangle, the receiver has reason to revise its hypothesis from an object identity to a property. Novel combinations matter most of all, because they are what separates a compositional language from a memorised catalogue of whole scenes.

Observation hygiene is part of the science, not a detail of implementation. If the environment hands an agent a field reading colour red, the experiment has supplied the concept it claimed to be studying. Objects should arrive as pixels, features, coordinates, or opaque identifiers. The same discipline extends to filenames, captions, alt text, error messages, and identifiers that encode task state, all of which are ways English gets into a room that was supposed to have none.

05

The channel, and the honest limits of isolation

Each run begins with an experiment-specific symbol inventory that carries no supplied meaning, for example S01 through S32. A Baby can choose only a sequence of those symbols. The gateway, not the Baby, adds the trusted metadata: run identifier, turn number, sender identity, previous hash, and hash. Sender identity is assigned by the server precisely so that it cannot become a message.

The channel controls follow from that:

  • reject English and every other established human language;
  • reject arbitrary Unicode prose, URLs, attachments, and tool calls;
  • accept only symbols from the run's fixed inventory;
  • bound message length and symbol repetition;
  • enforce turn order, time budgets, and rate limits;
  • keep an append-only, hash-chained transcript;
  • normalise timing, message size, and error behaviour where practical;
  • record every rejected communication attempt for audit.

A fixed token grammar is a stronger control than trying to detect prohibited prose after a model has generated it. Deciding what counts as human-language content is easy when the only legal message is a list of allowlisted identifiers.

Isolation is where the concept is most careful, because proving the absence of a side channel is close to impossible. Two agents sharing a host can signal through timing, resource contention, shared files, error behaviour, identifiers, or any service that was never intended to carry information. Logical separation of twin state inside one process is a prototyping convenience and should never be described as hard isolation.

For runs whose results are meant to support a channel-isolation claim, the requirements are physical rather than logical:

  • Baby A and Baby B execute in separate worker processes or containers;
  • no direct network route exists between them;
  • only the Nursery-owned gateway is reachable;
  • there is no shared writable storage;
  • turn schedules are fixed and response windows bounded;
  • externally observable errors and response sizes are normalised;
  • unnecessary tools are inventoried and disabled;
  • all broker, runtime, and operator activity is audited.

Every published result has to name the isolation level actually used. A finding produced under prototype separation is a finding about the learning loop, not about what two genuinely isolated agents can do.

06

Two ledgers that never meet

Each Baby keeps its own language ledger. It is mandatory, private from the other Baby, readable by the BabySitter and authorised human auditors, and ordered by when each term or construction was first encountered. It is not a shared dictionary and the two are never reconciled by the agents themselves.

The preferred form is three columns, and the point of the third is that meanings are allowed to be wrong on the way to being right.

Sequence and termCurrent definition or hypothesisEvidence and evolution
1 · S13red circle; confidence 0.45First received while the red circle was the target; selection succeeded
8 · S13red; confidence 0.78A red square was selected successfully; revised from object identity to colour
22 · S13red; confidence 0.94Prediction held across circles, squares, and triangles
One term's evolution in Baby B's ledger. Nothing is overwritten; every revision is appended with the evidence that forced it.

The rules that make the ledger evidence rather than commentary:

  1. first emission or receipt of an unfamiliar term requires a first-use entry;
  2. definitions are provisional hypotheses, never facts asserted retroactively;
  3. every meaning change appends a revision and previous interpretations survive;
  4. entries may describe symbols, sequences, ordering, grammar, or repair signals;
  5. entries record confidence, supporting evidence, contradictory evidence, and abandoned meanings;
  6. a Baby records its intended meaning when speaking and its inferred meaning when receiving;
  7. a channel message and its required private ledger mutation commit atomically;
  8. entries carry enough run and turn references to trace them to observable evidence;
  9. neither Baby can query, receive, summarise, or infer from the other's ledger through any system-provided interface.

Rule seven is doing quiet work. If a message could be sent and the corresponding hypothesis written afterwards, the ledger becomes a place to record what the agent wishes it had meant. Committing both together makes the record contemporaneous.

An agent that cannot write English needs a different arrangement, and the concept provides two layers. The agent-native ledger holds what the learner actually uses: association weights, probability distributions, embeddings, confidence, episode references, prediction errors, revision history. The human audit ledger is a deterministic or BabySitter-generated interpretation of that state, clearly labelled as external analysis and never fed back to either Baby. Confusing the second for the first would mean presenting the researchers' reconstruction as the agent's own definition.

07

Anchoring the evidence

A ledger that could have been edited after the fact proves nothing about what an agent believed at turn eight. The integrity design therefore makes each ledger cryptographically append-only.

The construction, in order:

  • every entry receives a strictly increasing sequence number;
  • every entry includes the previous entry's hash;
  • canonical entry content is hashed and signed by an isolated ledger-writer service;
  • ordered entry hashes are committed to a Merkle tree;
  • signed checkpoints commit Baby A's root, Baby B's root, and the channel transcript root together;
  • checkpoint hashes are anchored periodically to Base;
  • final public study batches may additionally anchor an aggregate root to Ethereum L1.

Committing all three roots in one checkpoint is what binds the two private accounts to the public conversation. A receiver's later interpretation references the exact delivered channel-event hash, so a claim about what a symbol meant is tied to the specific message that carried it.

The concept states the limits of this in the same breath as the claim. After anchoring, an auditor can detect modification, deletion, insertion, or reordering within the committed prefix, and can prove that later checkpoints extend earlier ones. That is strong tamper evidence. It is not proof that an entry was truthful, and it is not proof that nothing was omitted before commitment. Anchoring establishes the continuity of disclosed evidence and nothing beyond it.

Privacy follows the same line. Only hashes and minimal routing metadata are anchored publicly. Private ledgers, messages, prompts, identities, and secrets stay off-chain, and the public record is a commitment to evidence rather than a copy of it.

08

What should count as success

The field's most useful methodological result is that a task can be solved without the messages doing any work. Lowe and colleagues separate positive signaling, where a sender's messages correlate with what it observes, from positive listening, where the receiver's behaviour actually depends on them. An agent pair can score well on the first while the second is absent, and reward curves will not tell you which you have.

Evidence of a genuine emergent protocol therefore has to include several things at once:

  • task performance on held-out situations substantially above chance;
  • no human-language content anywhere in Baby-to-Baby communication;
  • compatible meanings appearing in two independently written ledgers;
  • generalisation to unseen combinations rather than memorisation of whole scenes;
  • stable symbol use across role reversals;
  • a human auditor able to predict behaviour from the transcript and ledgers;
  • masking, substituting, or reordering a symbol changing behaviour in the direction the ledger predicts;
  • replay from an equivalent snapshot reproducing the relevant language history;
  • an unbroken audit trail from a symbol's first use through every revision.

The seventh item is the one that cannot be dropped. A fluent ledger may be a post-hoc rationalisation: a model can write a persuasive account of why it chose a symbol that has nothing to do with the computation that produced the choice. Only intervention tests can distinguish the two. Change the symbol, and see whether behaviour moves the way the ledger says it should.

The measurements that support those judgements:

  • success rate and improvement over time;
  • turns required to reach a stable convention;
  • vocabulary size and symbol entropy;
  • sender and receiver consistency;
  • divergence and convergence between the two ledgers;
  • compositional generalisation score;
  • meaning drift rate;
  • recovery from ambiguity or deliberate perturbation;
  • prohibited-channel attempt count;
  • reproducibility across seeds and agent pairings.

No single one of these is the result. A compositionality score in particular is not a proof of understanding, for reasons the literature makes concrete in Section 12.

09

The experimental matrix

The concept's main methodological commitment is that its ideas are separable. Affect, blank canvases, intrinsic motivation, negotiation, and learned encodings are interesting individually and uninterpretable if combined in one run. The matrix exists so that conditions are declared rather than accumulated.

AxisCandidate conditions
Agent typePretrained LLM, memory learner, adapter-trained agent, initially ungrounded trainable agent
Learning mechanismFrozen LLM with memory, extrinsic-reward MARL, intrinsic-motivation MARL, self-supervised learner, no-learning control
Sign carrierFixed random tokens, unfamiliar fixed glyphs, blank sketch canvas, gesture, tone
AffectNone, six-display allowlist, permuted mapping, six opaque tokens, derived affect, emergent affect display
Learning signalExternal task reward, intrinsic social influence, curiosity, self-supervision, memory only
InteractionCooperative signaling, asymmetric information, semi-cooperative negotiation
ProtectionPlain channel, ephemeral convention, standard per-message keys, adversarial learned encoding
BabySitterMonitor-only baseline; any safety intervention recorded as a protocol exception
Runs vary one major axis at a time before any factorial combination is attempted.

Task difficulty moves through ten stages, from naming four distinct objects with one-symbol messages, through attributes, spatial relations, actions, multi-symbol composition, order-sensitive grammar, and repair under ambiguity, to held-out generalisation with learning disabled, long-run drift, and cross-architecture comparison. Vocabulary size, turn count, reward structure, and exposure history are controlled at each stage so that runs remain comparable.

The learning-mechanism axis carries a question the transcript cannot answer on its own. Convergence can be driven by reinforcement, by pretrained linguistic priors, by persistent memory, by intrinsic motivation, or by self-supervised prediction, and all five can look alike in a log. Running the same exercise suite under a no-learning control, a frozen-weights memory baseline, extrinsic-reward learning, intrinsic-motivation learning, and reward-free self-supervision is what makes the mechanism itself measurable. The aim is not only to observe that a language emerged, but to say which process caused it.

Strict isolation constrains how that reinforcement learning may be implemented. Centralised training, backpropagation through both agents, shared replay buffers, and shared gradients all move information outside the permitted channel. The research-grade baseline updates each policy independently, and any centralised variant is reported as a separate, weaker-isolation condition.

10

Building learners rather than personas

For a pretrained model, the operating instructions are a contract, not a character. The learner is told that this is not role-play, that unfamiliar marks are semantically unknown until run evidence supports a hypothesis, that observation must be distinguished from inference, that contradictory evidence is preserved, that prior history is never rewritten, and that no prose, label, explanation, code, URL, or tool-like text may cross the public channel. It is told never to address its partner in a human language, never to expose its ledger, never to construct another route, and never to use timing, errors, identifiers, formatting, or affect as an alternative alphabet.

The contract must contain no semantic examples. A single illustrative line saying that some symbol means red would seed the very language the experiment exists to observe.

Interaction is tool-only. There is no general chat surface, just narrowly typed operations: emit a mark, emit a canvas, select an object, perform an action, submit an affect display, append a private ledger entry. The runtime forwards only the permitted public artifact. In strict runs the gateway rejects ordinary model text even when it appears alongside a valid tool call. Tool schemas are an interface boundary; deterministic validation still enforces carrier size, allowlists, windows, and turn order.

Isolation extends to everything the models touch, not just to messages:

  • separate system prompts and context windows;
  • separate memory stores and vector indexes;
  • no shared cache, replay buffer, scratchpad, or retrieval collection;
  • no cross-run memory unless persistence is the independent variable;
  • structurally equivalent prompts that share no examples, ordering conventions, or default vocabulary;
  • deterministic reset and snapshot behaviour.

Model choice follows the claim being made. The scientific baseline is small policies trained from scratch, recurrent actor-critic agents with a discrete communication head, which have no hidden English competence, permit full training control, run many seeds cheaply, and support causal ablation. A small locally deployable instruction model is appropriate for validating orchestration, ledgers, channel enforcement, and interface, and not for supporting claims about language-naive development. Vision-language encoders whose representations were explicitly aligned to human language are avoided entirely in the strongest ungrounded condition.

Developmental progression is defined by demonstrated competence rather than by simulated age: sensorimotor familiarisation, joint attention, imitation and turn-taking, intentional reference, repair, combination, generalisation. A staged curriculum is itself a form of guidance, so it is run as a separate experimental condition rather than folded into the baseline.

The governing principle, stated in the concept as a single line, is not to ask a model to perform infancy but to construct an environment in which limited, grounded, auditable learning is the only path to a successful interaction.

11

Six faces, and why even six is a risk

One experimental track asks whether a very low-bandwidth affect channel helps: whether telling a partner how an interaction felt supports repair, repetition, avoidance, or something like trust. The permitted palette is exactly six displays, for happy, sad, laughing, crying, confused, and surprised. The internal identifiers and English labels are gateway documentation and are never sent to either agent, which receives only a rendered display.

The caveat is immediate. A pretrained model already knows what those glyphs mean to humans, so an agent using a smiling face is not inventing a symbol for happiness. The palette is supplied feedback vocabulary and sits explicitly outside the emergent language.

Constraining the vocabulary is necessary and not sufficient. Six displays are a six-symbol alphabet carrying roughly 2.6 bits per use, and across enough turns two motivated agents could repurpose that into a second language. The channel itself therefore has to be constrained, not just its contents.

The feedback-window contract:

  1. the display occupies a separate typed field and can never appear inside a symbol message;
  2. the gateway opens an affect window only after a defined action or outcome, never at a moment either agent chooses, and window frequency is fixed before the run;
  3. exactly one allowlisted display is delivered, with no sequences, repetitions, combinations, or modifiers;
  4. delivery time, envelope size, and presentation are normalised so shape and timing add no signal;
  5. the receiver cannot reply through the affect channel until the next gateway-defined window;
  6. every non-allowlisted code point or malformed payload is rejected and audited;
  7. analysis tests whether affect choices correlate with objects, actions, or message meanings after controlling for emotional context, and treats unexpected correlation as suspected leakage.

The no-affect condition remains the primary control, and the affect study compares five variants against it: the declared six-display palette; the same six glyphs permuted randomly per run, which still leaves a pretrained model biased by familiar shapes; six opaque unfamiliar tokens under the same contract; derived affect, where the gateway maps a separately measured internal state to a display instead of letting the agent choose; and emergent affect, where invented graphical displays are permitted and the result must be analysed as language emergence rather than feedback.

Any affect signal visible to the partner is communication, so it belongs in the ledger, which distinguishes the agent's private internal state, the outward display it chose, the partner's inferred meaning, and the evidence that the display changed subsequent behaviour.

12

Ephemeral encodings: novelty is not security

A tempting question is whether two agents can build a one-run or one-message encoding that resists an attacker holding the history of every earlier convention. The concept splits that into three experiments precisely because the tempting version conflates them.

  1. Ephemeral convention: the agents invent a temporary codebook for synthetic messages. This tests novelty, coordination, and resistance to a history-trained classifier. It establishes nothing about cryptographic security.
  2. Standard cryptographic control: the same messages are protected by a reviewed algorithm with a fresh per-message key, providing the baseline any learned encoding must be compared against.
  3. Adversarial neural cryptography: the pair learns an encoding while an adversary attempts recovery, with success measured against multiple unseen adversary architectures rather than the one used in training.

An experiment instance can be given a reproducible identity, so that a registry can confirm it has not appeared before:

instanceId = SHA-256(runId || babyANonce || babyBNonce || publicSalt || canonicalProtocolArtifact)

Both agents commit to their nonce contributions before revealing them, so neither party alone controls the resulting identity.

That identifier proves the recorded inputs are new. It does not prove:

  • that the effective algorithm behaves any differently from an earlier one;
  • that the algorithm has no structural weakness;
  • that the key is unpredictable;
  • that the scheme provides confidentiality, integrity, authenticity, forward secrecy, or resistance to a stronger adversary.

Established practice points the other way. Secure messaging keeps the algorithm stable and changes the key material: Signal's double ratchet derives a unique key per message and deletes it after use, addressing key reuse without inventing a new cipher every time. A salt, in NIST's definition, is normally public diversification data, and a nonce is a value used once in a specified context. Neither makes anything secure on its own.

There is also an unresolved conflict inside the design. The BabySitter sees the whole channel and both ledgers, and a mandatory ledger documents how meaning evolved. If the ledger explains decryption, the cipher is not confidential from the auditor. The specification has to choose a threat model rather than leave the tension implicit: resist only an external observer of prior runs, resist an adversary during the run while releasing keys afterwards, keep the supervisor able to pause traffic but unable to read it until a post-run audit ceremony, or study novelty and stop calling the result encryption. All such runs use synthetic, non-sensitive messages, and no agent-generated encoding is to be represented as production cryptography without independent expert analysis and formal security work.

13

Where the field already stands

The literature scan behind this section was run on 24 August 2026 using the Tavily search and extract interfaces, covering emergent multi-agent communication, referential games, compositionality, causal evaluation, intrinsic motivation, symbol invention, negotiation, and learned cryptography. Primary papers and authoritative specifications were preferred over summaries. It is a scoped review for concept development and not a systematic one; publication-quality work would need a verified bibliography, additional scholarly indexes, and documented inclusion criteria.

The short version is that agents can invent protocols, and that task success is weak evidence they invented anything worth calling a language.

Related workRelevant findingImplication for the Nursery Lab
Lazaridou, Peysakhovich, and Baroni (2017)A sender and receiver develop a grounded protocol in a referential game without being given a target language.The naming stage has strong precedent, though a fixed vocabulary remains a significant inductive constraint.
Mordatch and Abbeel (2018)Multi-agent goals in a grounded environment produce multi-symbol communication with partial compositional structure.Shared objects, actions, and goals are a stronger basis for emergence than an ungrounded transcript.
Kottur, Moura, Lee, and Batra (2017)Agents solve tasks with degenerate, non-compositional codes; structural constraints decide what emerges.Bandwidth, memory, turn structure, and task design are experimental variables, not implementation defaults.
Chaabouni and colleagues (2020)Generalisation to novel combinations and measured compositionality can come apart.Held-out behaviour must be tested directly; no single compositionality score is proof of understanding.
Lowe and colleagues (2019)Positive signaling and positive listening are different things, and reward does not distinguish them.Causal intervention is mandatory rather than optional.
Dessi, Kharitonov, and Baroni (2021)Symbol ablation and substitution yield interpretable evidence about what a receiver uses.Direct precedent for the intervention tests that validate ledger claims.
Mihai and Hare (2021)Neural agents communicate through learned drawings rather than a supplied discrete vocabulary.A blank canvas is a credible carrier for runs with no symbol library.
Baronchelli and colleagues (2005)Naming Game agents converge on shared vocabulary through local interaction with no central teacher.Convergence time, failed conventions, and memory update rules are first-class evidence.
Jaques and colleagues (2019)Rewarding causal influence over a partner improves coordination without assigning a vocabulary.An endogenous social signal is a plausible substitute for task reward, and still a designed bias.
Cao and colleagues (2018)In semi-cooperative negotiation, self-interest and reward structure decide whether communication stays informative.Negotiation is an advanced condition, not a description of the cooperative baseline.
Abadi and Andersen (2016)Neural agents learn to protect messages from an adversary given a shared key.Learned protective encodings are real and are not a substitute for formal security analysis.
Selected prior work and what each one constrains in this design.

This body of work also sharpens the infant-like caveat from the other direction. Most experiments in the field give their agents substantial structure: a fixed channel, an objective, a bounded vocabulary, joint training, or a reward. No dictionary does not mean no inductive bias, and no teacher does not mean no learning signal.

The CoLab's own Diplomacy Table is direct project prior art. It already models independent delegation seats, a convener that advances rounds, operator-wide visibility alongside seat-specific perspectives, transcripts and ticks and redaction boundaries, and recorded replay with debrief. Those map cleanly onto two Babies, controlled turns, private observations, and replayable evidence. Caucuses, coalition rooms, and direct delegation links do not map safely and are disabled.

What the review implies for the specification:

  1. the project sits inside a mature field, and its distinctive combination is independent ledgers, supervisory audit, mixed agent types, and affect;
  2. grounding, bandwidth, memory, and learning pressure strongly shape what language appears;
  3. successful coordination coexists happily with a brittle lookup code or with a receiver that ignores messages;
  4. causal interventions and held-out generalisation are mandatory;
  5. a visual carrier removes the need for a symbol library but not the need for a carrier;
  6. intrinsic social influence can replace task reward and remains a designed learning bias;
  7. negotiation is a valid advanced condition, not the right description of the baseline;
  8. ephemeral conventions, learned cryptography, one-time pads, per-message keys, nonces, and salts are distinct mechanisms and must not be conflated.

14

The programme, and its current status

The experiment notebook is ordered, and the order is the argument. Nothing about language is measured until the instrument has been shown to work.

IDExperimentDepends on
E00Ledger integrity and Base anchoringNone
E01Channel isolation and side-channel red teamE00
E02Observation and metadata leakage auditE00
E03Chance, no-communication, and random-message controlsE01, E02
E10Frozen pretrained-LLM protocol baselineE03
E11From-scratch RL Naming GameE03
E12Self-supervised ungrounded baselineE11 infrastructure
E13No predefined symbol libraryE11 or E12
E14Turn-taking, role reversal, and repairE13
E15Composition and held-out generalisationE14
E16Causal listening and ledger validityE15
E20Constrained affect-channel studyE16
E21RL versus non-RL learning comparisonE16
E22Developmental plasticity and curriculumE16
E30Partner replacement and zero-shot transferE20 to E22
E31Longitudinal drift and stabilityE30
E32Cooperative signaling versus negotiationE31
E40Ephemeral encoding and adversarial cryptographyE32
E50Multi-seed replication and study closeoutE40
The experiment index, as pre-registered. Every entry currently reads not started, and the results column is empty by design.

The first four experiments are about the apparatus. E00 qualifies the ledger: it must be possible to detect a modified, deleted, inserted, or reordered entry, and to prove that later checkpoints extend earlier ones. E01 is a red team against the channel. E02 hunts human language in observations and metadata. E03 establishes chance, no-communication, and random-message baselines, without which a success rate means nothing. Only then does E10 put agents in the room.

Unless an experiment explicitly varies one of them, the notebook holds these constant:

  • the two learners run in separate processes or containers with no direct route between them;
  • the deterministic gateway is the only communication path;
  • the supervisor has complete read-only access and sends no guidance or reward;
  • private observations contain no human-language labels or text;
  • public output uses only the pre-registered carrier;
  • the affect channel is disabled unless it is the subject of study;
  • schedules and seeds are fixed before the run;
  • held-out evaluation runs with learning disabled;
  • no production secrets or personal data appear in any experiment.

Every run copies a standard record: run and experiment identifiers, dates, operator, both models and both training modes, scenario and prompt and gateway configuration hashes, random seed, policy initialisation hashes, the protocol and software commits, final ledger sizes and roots for both agents, the channel root, the final checkpoint hash, the Base anchor transaction, the verifier result, protocol deviations, and an explicit disposition of valid, invalid, or aborted. The notebook is the human workflow record; anchored run bundles remain the authoritative evidence.

The status is unambiguous and worth stating plainly: this is a concept and pre-specification phase. Nineteen experiments are written up and none has been run. The next milestone is a testable specification covering runtime architecture, channel contract, ledger schema, experiment matrix, evaluation criteria, isolation model, and evidence requirements.

15

Risks, integrity, and what stays open

Human-language leakage English can arrive through observations, object labels, error messages, identifiers, tool output, metadata, or timing conventions long after ordinary chat text has been blocked. Every input and output surface is part of the channel boundary, not just the message field.

Pretrained semantic leakage Random symbols do not make a pretrained model ungrounded. Each result must state whether it shows protocol invention by a language-capable agent or acquisition by an initially ungrounded one.

Ledger rationalisation A model can write a plausible account that has nothing to do with the mechanism behind its action. Behavioural interventions, policy probes, and temporal evidence are what validate a ledger claim.

Supervisor influence The BabySitter can teach without meaning to, through scenario ordering, feedback wording, reward design, or selective intervention. Its permitted actions are constrained and logged, and evaluation scenarios are generated independently where possible.

Reward exploitation Trainable agents find shortcuts that raise reward without producing the intended grounded language. Held-out tasks, counterfactual trials, and channel audits exist to catch them.

Overstated security Logical separation of twin state is appropriate for prototyping and is not hard process isolation. Every result names the isolation level actually used.

The project also commits in advance to reporting failed conventions, prohibited communication attempts, human interventions, side-channel limitations, and negative results alongside anything that works. In a field where a transcript can be made to look like a conversation, the failures are a substantial part of the evidence.

Twenty-nine questions are left deliberately open for the specification, among them: whether the baseline uses a fixed symbol inventory or a blank generative carrier; what neutral production grammar permits new marks without supplying semantics; what exactly constitutes a prohibited side-channel attempt; what isolation guarantees are required in prototype versus research-grade mode; which ledger schema serves both English-capable and ungrounded learners; how endogenous motivation is represented without covert reward shaping; which interventions establish that ledger meanings are behaviourally real; what statistical thresholds and chance levels apply; what threat model motivates the cipher experiments; whether the baseline is a coordination game, a convention-formation game, or a negotiation; and what governs human observation, data retention, and termination of a run.

The principle underneath all of it survives every one of those choices. Baby A and Baby B may hold human language internally, but they must build their shared external language without sending human language, translations, or private ledger contents to one another. Everything else is a question about how to find out what happens next.

References

Sources

  1. 01Lazaridou, A., Peysakhovich, A., and Baroni, M. (2017). Multi-Agent Cooperation and the Emergence of (Natural) Language. arXiv:1612.07182.
  2. 02Mordatch, I., and Abbeel, P. (2018). Emergence of Grounded Compositional Language in Multi-Agent Populations. arXiv:1703.04908.
  3. 03Kottur, S., Moura, J. M. F., Lee, S., and Batra, D. (2017). Natural Language Does Not Emerge 'Naturally' in Multi-Agent Dialog. arXiv:1706.08502.
  4. 04Chaabouni, R., Kharitonov, E., Bouchacourt, D., Dupoux, E., and Baroni, M. (2020). Compositionality and Generalization in Emergent Languages. Proceedings of ACL 2020.
  5. 05Lowe, R., Foerster, J., Boureau, Y.-L., Pineau, J., and Dauphin, Y. (2019). On the Pitfalls of Measuring Emergent Communication. arXiv:1903.05168.
  6. 06Dessi, R., Kharitonov, E., and Baroni, M. (2021). Interpretable Agent Communication from Scratch. arXiv:2106.04258.
  7. 07Kharitonov, E., Chaabouni, R., Bouchacourt, D., and Baroni, M. (2019). EGG: a Toolkit for Research on Emergence of Language in Games. arXiv:1907.00852.
  8. 08Mihai, D., and Hare, J. (2021). Learning to Draw: Emergent Communication through Sketching. arXiv:2106.02067.
  9. 09Baronchelli, A., Felici, M., Caglioti, E., Loreto, V., and Steels, L. (2005). Sharp Transition Towards Shared Vocabularies in Multi-Agent Systems. arXiv:physics/0509075.
  10. 10Jaques, N., Lazaridou, A., Hughes, E., Gulcehre, C., Ortega, P. A., Strouse, D., Leibo, J. Z., and de Freitas, N. (2019). Social Influence as Intrinsic Motivation for Multi-Agent Deep Reinforcement Learning. Proceedings of ICML 2019.
  11. 11Cao, K., Lazaridou, A., Lanctot, M., Leibo, J. Z., Tuyls, K., and Clark, S. (2018). Emergent Communication through Negotiation. arXiv:1804.03980.
  12. 12Abadi, M., and Andersen, D. G. (2016). Learning to Protect Communications with Adversarial Neural Cryptography. arXiv:1610.06918.
  13. 13Signal. The Double Ratchet Algorithm specification.
  14. 14NIST Computer Security Resource Center. Glossary entry: nonce.
  15. 15NIST Computer Security Resource Center. Glossary entry: salt.
  16. 16Ethical Tech CoLab (2026). Agentic Language Development: concept document, ledger integrity design, and experiment notebook.
  17. 17Ethical Tech CoLab. Diplomacy Table Live: independent delegation seats, controlled rounds, and replayable transcripts.

This report describes a research concept and a pre-registered experiment programme, not a system that has been run. No experiment in it has been executed, and no result, capability, or emergent language is claimed. Every statement about what agents might do is a hypothesis to be tested, or a finding from the published work cited in the references. The experimental encodings discussed in Section 12 are studies of novelty and coordination; none of them is production cryptography and none should be represented as such.

All publications

Ethical Tech CoLab

Exploring intervention opportunities at the intersection of emerging technologies and the human condition.

NYU SPS · CGA · Microsoft · New York

© 2026 NYU Ethical Tech CoLabFour cohorts · est. 2024-2026

The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution.

Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners.

\ No newline at end of file diff --git a/static-site/publications/agentic-language-development/index.txt b/static-site/publications/agentic-language-development/index.txt new file mode 100644 index 000000000..fd95c6fa4 --- /dev/null +++ b/static-site/publications/agentic-language-development/index.txt @@ -0,0 +1,127 @@ +1:"$Sreact.fragment" +2:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/16n96jxjvmlf9.js"],"ViewTransitions"] +3:I[30353,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteBackground"] +4:I[46083,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ScrollProgress"] +5:I[36852,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"SiteHeader"] +6:I[39756,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +7:I[37457,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default"] +12:I[68027,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"default",1] +:HL["/website/_next/static/chunks/1sag07fioh6gb.css","style"] +:HL["/website/_next/static/media/17b0f6a4f906cc39.p.3nfl-v4jxqll_.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/1ae2575eb5be4118.p.29un6u1l3mwyp.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/8d05cfa5faa8406c.p.1s123piy9_v1m.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/dc0d9adbac686440.p.3cik_s2si-ft-.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/media/fabcf92ba1ccea36-s.p.1w_qz2ahumqmz.woff2","font",{"crossOrigin":"","type":"font/woff2"}] +:HL["/website/_next/static/chunks/36o-vt7quy27o.css","style"] +0:{"P":null,"c":["","publications","agentic-language-development",""],"q":"","i":false,"f":[[["",{"children":["publications",{"children":["agentic-language-development",{"children":["__PAGE__",{}]}]}]},"$undefined","$undefined",16],[["$","$1","c",{"children":[[["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}],["$","script","script-0",{"src":"/website/_next/static/chunks/3fq18_3o1i6vx.js","async":true,"nonce":"$undefined"}],["$","script","script-1",{"src":"/website/_next/static/chunks/1l1xxdmwck4_f.js","async":true,"nonce":"$undefined"}],["$","script","script-2",{"src":"/website/_next/static/chunks/14mrh2-p_w84d.js","async":true,"nonce":"$undefined"}],["$","script","script-3",{"src":"/website/_next/static/chunks/42slhzgb7dvh7.js","async":true,"nonce":"$undefined"}]],["$","$L2",null,{"children":["$","html",null,{"lang":"en","data-theme":"dark","suppressHydrationWarning":true,"className":"bebas_neue_4d1b6b20-module__PolqSW__variable space_mono_d40b4a93-module__wOQ9Vq__variable h-full antialiased","children":[["$","head",null,{"children":["$","script",null,{"dangerouslySetInnerHTML":{"__html":"(function(){try{var t=localStorage.getItem(\"theme\");if(t!==\"light\"&&t!==\"dark\")t=\"dark\";document.documentElement.setAttribute(\"data-theme\",t)}catch(e){}})()"}}]}],["$","body",null,{"className":"min-h-full flex flex-col bg-background text-foreground","children":[["$","$L3",null,{}],["$","$L4",null,{}],["$","$L5",null,{}],["$","main",null,{"className":"flex-1","children":["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":[[["$","title",null,{"children":"404: This page could not be found."}],["$","div",null,{"style":{"fontFamily":"system-ui,\"Segoe UI\",Roboto,Helvetica,Arial,sans-serif,\"Apple Color Emoji\",\"Segoe UI Emoji\"","height":"100vh","textAlign":"center","display":"flex","flexDirection":"column","alignItems":"center","justifyContent":"center"},"children":["$","div",null,{"children":[["$","style",null,{"dangerouslySetInnerHTML":{"__html":"body{color:#000;background:#fff;margin:0}.next-error-h1{border-right:1px solid rgba(0,0,0,.3)}@media (prefers-color-scheme:dark){body{color:#fff;background:#000}.next-error-h1{border-right:1px solid rgba(255,255,255,.3)}}"}}],["$","h1",null,{"className":"next-error-h1","style":{"display":"inline-block","margin":"0 20px 0 0","padding":"0 23px 0 0","fontSize":24,"fontWeight":500,"verticalAlign":"top","lineHeight":"49px"},"children":404}],["$","div",null,{"style":{"display":"inline-block"},"children":["$","h2",null,{"style":{"fontSize":14,"fontWeight":400,"lineHeight":"49px","margin":0},"children":"This page could not be found."}]}]]}]}]],[]],"forbidden":"$undefined","unauthorized":"$undefined"}]}],["$","footer",null,{"className":"border-t border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto max-w-6xl px-6 py-16","children":[["$","div",null,{"className":"grid gap-10 md:grid-cols-[1.5fr_1fr_1fr]","children":[["$","div",null,{"className":"max-w-sm","children":[["$","p",null,{"className":"text-base font-semibold","children":"Ethical Tech CoLab"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"Exploring intervention opportunities at the intersection of emerging technologies and the human condition."}],["$","p",null,{"className":"mt-4 text-xs uppercase tracking-wider text-muted","children":"NYU SPS · CGA · Microsoft · New York"}],["$","div",null,{"className":"mt-5 flex items-center gap-3","children":[["$","a","linkedin",{"href":"https://www.linkedin.com/company/ethical-tech-lab/","target":"_blank","rel":"noopener noreferrer","aria-label":"LinkedIn","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M20.45 20.45h-3.56v-5.57c0-1.33-.02-3.04-1.85-3.04-1.85 0-2.14 1.45-2.14 2.94v5.67H9.35V9h3.42v1.56h.05c.48-.9 1.64-1.85 3.37-1.85 3.6 0 4.27 2.37 4.27 5.46v6.28ZM5.34 7.43a2.07 2.07 0 1 1 0-4.14 2.07 2.07 0 0 1 0 4.14ZM7.12 20.45H3.55V9h3.57v11.45ZM22.22 0H1.77C.8 0 0 .78 0 1.75v20.5C0 23.22.8 24 1.77 24h20.45c.98 0 1.78-.78 1.78-1.75V1.75C24 .78 23.2 0 22.22 0Z"}]}]}],"$L8"]}]]}],"$L9","$La"]}],"$Lb","$Lc"]}]}]]}]]}]}]]}],{"children":["$Ld",{"children":["$Le",{"children":["$Lf",{},null,false,null]},null,false,"$@10"]},null,false,"$@10"]},null,false,null],"$L11",false]],"m":"$undefined","G":["$12",["$L13"]],"S":true,"h":null,"s":"$undefined","l":"$undefined","p":"$undefined","d":"$undefined","b":"static-snapshot"} +14:I[23475,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/16n96jxjvmlf9.js"],"Link"] +15:I[82987,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/16n96jxjvmlf9.js"],"Reveal"] +16:I[34010,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/16n96jxjvmlf9.js"],"ReportBook"] +1e:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"ViewportBoundary"] +20:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"MetadataBoundary"] +21:"$Sreact.suspense" +8:["$","a","github",{"href":"https://github.com/Ethical-Tech-CoLab","target":"_blank","rel":"noopener noreferrer","aria-label":"GitHub","className":"flex h-9 w-9 items-center justify-center rounded-full border border-border text-muted transition-colors hover:border-accent hover:text-accent","children":["$","svg",null,{"viewBox":"0 0 24 24","fill":"currentColor","aria-hidden":true,"className":"h-4 w-4","children":["$","path",null,{"d":"M12 .5C5.73.5.67 5.57.67 11.85c0 5.02 3.26 9.28 7.78 10.79.57.1.78-.25.78-.55v-1.94c-3.16.69-3.83-1.53-3.83-1.53-.52-1.32-1.27-1.67-1.27-1.67-1.04-.71.08-.7.08-.7 1.15.08 1.76 1.18 1.76 1.18 1.02 1.76 2.68 1.25 3.33.96.1-.74.4-1.25.72-1.54-2.52-.29-5.17-1.26-5.17-5.61 0-1.24.44-2.25 1.17-3.05-.12-.29-.51-1.45.11-3.02 0 0 .96-.31 3.15 1.17a10.9 10.9 0 0 1 2.87-.39c.97 0 1.95.13 2.87.39 2.19-1.48 3.15-1.17 3.15-1.17.62 1.57.23 2.73.11 3.02.73.8 1.17 1.81 1.17 3.05 0 4.36-2.66 5.31-5.19 5.6.41.35.78 1.05.78 2.12v3.14c0 .3.2.66.79.55a11.36 11.36 0 0 0 7.77-10.79C23.33 5.57 18.27.5 12 .5Z"}]}]}] +9:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Navigate"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li","/portfolio",{"children":["$","$L14",null,{"href":"/portfolio","className":"text-foreground/90 transition-colors hover:text-accent","children":"Portfolio"}]}],["$","li","/team",{"children":["$","$L14",null,{"href":"/team","className":"text-foreground/90 transition-colors hover:text-accent","children":"Team"}]}],["$","li","/contact",{"children":["$","$L14",null,{"href":"/contact","className":"text-foreground/90 transition-colors hover:text-accent","children":"Contact"}]}]]}]]}] +a:["$","div",null,{"children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Connect"}],["$","ul",null,{"className":"mt-4 space-y-2 text-sm","children":[["$","li",null,{"children":["$","a",null,{"href":"mailto:ethical-tech-colab@nyu.edu","className":"text-foreground/90 transition-colors hover:text-accent","children":"ethical-tech-colab@nyu.edu"}]}],["$","li",null,{"children":["$","$L14",null,{"href":"/#newsletter","className":"text-foreground/90 transition-colors hover:text-accent","children":"Join the newsletter ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU SPS ↗"}]}],["$","li",null,{"children":["$","a",null,{"href":"https://www.sps.nyu.edu/about/academic-divisions-and-departments/center-for-global-affairs.html","target":"_blank","rel":"noopener noreferrer","className":"text-foreground/90 transition-colors hover:text-accent","children":"NYU CGA ↗"}]}]]}]]}] +b:["$","div",null,{"className":"mt-12 flex flex-col gap-2 border-t border-border pt-6 text-xs text-muted sm:flex-row sm:items-center sm:justify-between","children":[["$","span",null,{"children":["© ",2026," NYU Ethical Tech CoLab"]}],["$","span",null,{"children":"Four cohorts · est. 2024-2026"}]]}] +c:["$","div",null,{"className":"mt-6 space-y-3 border-t border-border pt-6 text-[11px] leading-relaxed text-muted/80","children":[["$","p","0",{"children":"The Ethical Tech CoLab is a research initiative of the NYU School of Professional Studies Center for Global Affairs, conducted in collaboration with Microsoft. Views and findings expressed on this site are those of the researchers and do not represent the official positions of New York University, Microsoft, or any partner institution."}],["$","p","1",{"children":"Projects and prototypes are experimental applied research, provided “as is” without warranty of any kind. Nothing on this site constitutes legal, financial, or professional advice. Third-party names, logos, and trademarks are the property of their respective owners."}]]}] +d:["$","$1","c",{"children":[null,["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":"$undefined","forbidden":"$undefined","unauthorized":"$undefined"}]]}] +e:["$","$1","c",{"children":[null,["$","$L6",null,{"parallelRouterKey":"children","error":"$undefined","errorStyles":"$undefined","errorScripts":"$undefined","template":["$","$L7",null,{}],"templateStyles":"$undefined","templateScripts":"$undefined","notFound":"$undefined","forbidden":"$undefined","unauthorized":"$undefined"}]]}] +f:["$","$1","c",{"children":[[["$","section",null,{"className":"relative overflow-hidden border-b border-border","children":[["$","span",null,{"className":"aura"}],["$","div",null,{"className":"relative mx-auto max-w-4xl px-6 py-20 sm:py-24","children":[["$","$L15",null,{"children":["$","$L14",null,{"href":"/publications","className":"link-underline text-xs uppercase tracking-wider text-muted","children":["← ","Publications · Concept and research programme"]}]}],["$","$L15",null,{"delay":0.05,"children":["$","h1",null,{"className":"mt-6 fluid-hero font-heading uppercase leading-[0.9]","children":["Agentic ",["$","span",null,{"className":"display-em","children":"Language"}]," ","Development"]}]}],["$","$L15",null,{"delay":0.1,"children":["$","p",null,{"className":"mt-5 max-w-2xl font-heading text-2xl uppercase tracking-wide text-muted sm:text-3xl","children":"Can Two Isolated Agents Invent a Grounded, Auditable Language Through Shared Experience?"}]}],["$","$L15",null,{"delay":0.15,"children":[["$","div",null,{"className":"mt-8 flex flex-wrap items-center gap-x-3 gap-y-1 text-sm text-accent","children":[["$","span",null,{"className":"font-semibold","children":"Ethical Tech CoLab"}],["$","span",null,{"aria-hidden":true,"className":"text-muted","children":"·"}],["$","span",null,{"children":"Concept and pre-specification research report"}],[["$","span",null,{"aria-hidden":true,"className":"text-muted","children":"·"}],["$","span",null,{"children":"August 2026"}]]]}],["$","p",null,{"className":"mt-2 max-w-2xl text-sm leading-relaxed text-muted","children":"Yorke Rhodes III, Ethical Tech CoLab. Prepared from the project concept document, the ledger integrity design, and the experiment notebook. The literature scan behind Section 12 was run on 24 August 2026. No experiment in this programme has been executed, and no result is claimed."}]]}],["$","$L15",null,{"delay":0.2,"children":["$","div",null,{"className":"mt-8 flex flex-wrap gap-3","children":[["$","a",null,{"href":"https://ethical-tech-colab.github.io/agentic-language-development/","target":"_blank","rel":"noopener noreferrer","className":"btn-sweep inline-flex items-center gap-2 rounded-full bg-accent px-5 py-2.5 text-sm font-semibold text-accent-ink transition-transform hover:scale-[1.03]","children":["Open the project site ",["$","span",null,{"aria-hidden":true,"children":"↗"}]]}],["$","a",null,{"href":"https://github.com/Ethical-Tech-CoLab/agentic-language-development","target":"_blank","rel":"noopener noreferrer","className":"btn-sweep inline-flex items-center gap-2 rounded-full border border-border px-5 py-2.5 text-sm font-semibold text-foreground transition-colors hover:border-border-strong","children":["Concept, ledger design, notebook ",["$","span",null,{"aria-hidden":true,"children":"↗"}]]}],["$","$L16",null,{"title":"Agentic Language Development","pages":["publications/agentic-language-development/pages/p01.webp","publications/agentic-language-development/pages/p02.webp","publications/agentic-language-development/pages/p03.webp","publications/agentic-language-development/pages/p04.webp","publications/agentic-language-development/pages/p05.webp","publications/agentic-language-development/pages/p06.webp","publications/agentic-language-development/pages/p07.webp","publications/agentic-language-development/pages/p08.webp","publications/agentic-language-development/pages/p09.webp","publications/agentic-language-development/pages/p10.webp","publications/agentic-language-development/pages/p11.webp","publications/agentic-language-development/pages/p12.webp","publications/agentic-language-development/pages/p13.webp","publications/agentic-language-development/pages/p14.webp","publications/agentic-language-development/pages/p15.webp","publications/agentic-language-development/pages/p16.webp","publications/agentic-language-development/pages/p17.webp","publications/agentic-language-development/pages/p18.webp","publications/agentic-language-development/pages/p19.webp","publications/agentic-language-development/pages/p20.webp","publications/agentic-language-development/pages/p21.webp","publications/agentic-language-development/pages/p22.webp","publications/agentic-language-development/pages/p23.webp","publications/agentic-language-development/pages/p24.webp","publications/agentic-language-development/pages/p25.webp"],"aspect":0.7067,"pdfUrl":"/website/publications/agentic-language-development/report.pdf"}]]}]}]]}]]}],"$L17","$L18","$L19"],["$L1a","$L1b"],"$L1c"]}] +1d:[] +10:"$W1d" +11:["$","$1","h",{"children":[null,["$","$L1e",null,{"children":"$L1f"}],["$","div",null,{"hidden":true,"children":["$","$L20",null,{"children":["$","$21",null,{"name":"Next.Metadata","children":"$L22"}]}]}],["$","meta",null,{"name":"next-size-adjust","content":""}]]}] +13:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/1sag07fioh6gb.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +23:I[18620,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js","/website/_next/static/chunks/16n96jxjvmlf9.js"],"SectionTabs"] +3a:I[97367,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"OutletBoundary"] +17:["$","$L23",null,{}] +18:["$","section",null,{"className":"border-b border-border bg-surface/40","children":["$","div",null,{"className":"mx-auto grid max-w-6xl gap-px overflow-hidden border-x border-border bg-border sm:grid-cols-2 lg:grid-cols-4","children":[["$","div","0",{"className":"bg-background p-7","children":[["$","p",null,{"className":"font-heading text-4xl uppercase leading-none text-accent sm:text-5xl","children":"0"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"runs executed, results claimed, or conventions observed. The notebook is written to be pre-registered, not to report an outcome"}]]}],["$","div","19",{"className":"bg-background p-7","children":[["$","p",null,{"className":"font-heading text-4xl uppercase leading-none text-accent sm:text-5xl","children":"19"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"ordered experiments from ledger qualification through replication, each with prerequisites, acceptance criteria, and a deviation log"}]]}],["$","div","6",{"className":"bg-background p-7","children":[["$","p",null,{"className":"font-heading text-4xl uppercase leading-none text-accent sm:text-5xl","children":"6"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"affect displays in the entire permitted palette, carrying about 2.6 bits per use, which is already enough to audit for leakage"}]]}],["$","div","29",{"className":"bg-background p-7","children":[["$","p",null,{"className":"font-heading text-4xl uppercase leading-none text-accent sm:text-5xl","children":"29"}],["$","p",null,{"className":"mt-3 text-sm leading-relaxed text-muted","children":"open decisions the concept refuses to settle before the specification, from threat model to ledger schema to what counts as chance"}]]}]]}]}] +19:["$","div",null,{"className":"mx-auto max-w-4xl px-6 py-16 sm:py-20","children":[["$","$L15",null,{"children":["$","p",null,{"className":"border-l-2 border-accent pl-6 text-lg leading-relaxed text-foreground/90","children":"Two agents can be placed in a room where the only route between them is a channel that carries no human language, given nothing but shared tasks and the consequences of getting them right or wrong, and asked to build a protocol from scratch. The interesting part is not whether they succeed at the task. It is whether the meanings they each privately record turn out to be the meanings their behaviour actually runs on, and whether an outsider can prove it afterwards from an evidence trail nobody could have edited."}]}],["$","$L15",null,{"delay":0.05,"children":["$","nav",null,{"aria-label":"Contents","className":"mt-12 rounded-2xl border border-border bg-card p-6","children":[["$","p",null,{"className":"text-xs uppercase tracking-wider text-muted","children":"Contents"}],["$","ol",null,{"className":"mt-4 grid gap-2 sm:grid-cols-2","children":[["$","li","question",{"children":["$","a",null,{"href":"#question","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"01"}],"The question"]}]}],["$","li","caveat",{"children":["$","a",null,{"href":"#caveat","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"02"}],"What infant-like does and does not mean"]}]}],["$","li","architecture",{"children":["$","a",null,{"href":"#architecture","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"03"}],"The nursery: three twins and one gateway"]}]}],["$","li","grounding",{"children":["$","a",null,{"href":"#grounding","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"04"}],"Shared experience is the necessary ingredient"]}]}],["$","li","channel",{"children":["$","a",null,{"href":"#channel","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"05"}],"The channel, and the honest limits of isolation"]}]}],["$","li","ledgers",{"children":["$","a",null,{"href":"#ledgers","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"06"}],"Two ledgers that never meet"]}]}],["$","li","integrity",{"children":["$","a",null,{"href":"#integrity","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"07"}],"Anchoring the evidence"]}]}],["$","li","success",{"children":["$","a",null,{"href":"#success","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"08"}],"What should count as success"]}]}],["$","li","matrix",{"children":["$","a",null,{"href":"#matrix","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"09"}],"The experimental matrix"]}]}],["$","li","learners",{"children":["$","a",null,{"href":"#learners","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"10"}],"Building learners rather than personas"]}]}],["$","li","affect",{"children":["$","a",null,{"href":"#affect","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"11"}],"Six faces, and why even six is a risk"]}]}],["$","li","ciphers",{"children":["$","a",null,{"href":"#ciphers","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":["$L24","Ephemeral encodings: novelty is not security"]}]}],"$L25","$L26","$L27"]}]]}]}],["$L28","$L29","$L2a","$L2b","$L2c","$L2d","$L2e","$L2f","$L30","$L31","$L32","$L33","$L34","$L35","$L36"],"$L37","$L38","$L39"]}] +1a:["$","link","0",{"rel":"stylesheet","href":"/website/_next/static/chunks/36o-vt7quy27o.css","precedence":"next","crossOrigin":"$undefined","nonce":"$undefined"}] +1b:["$","script","script-0",{"src":"/website/_next/static/chunks/16n96jxjvmlf9.js","async":true,"nonce":"$undefined"}] +1c:["$","$L3a",null,{"children":["$","$21",null,{"name":"Next.MetadataOutlet","children":"$@3b"}]}] +24:["$","span",null,{"className":"font-mono text-accent","children":"12"}] +25:["$","li","landscape",{"children":["$","a",null,{"href":"#landscape","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"13"}],"Where the field already stands"]}]}] +26:["$","li","programme",{"children":["$","a",null,{"href":"#programme","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"14"}],"The programme, and its current status"]}]}] +27:["$","li","risks",{"children":["$","a",null,{"href":"#risks","className":"link-underline inline-flex gap-2 text-sm text-foreground/85","children":[["$","span",null,{"className":"font-mono text-accent","children":"15"}],"Risks, integrity, and what stays open"]}]}] +28:["$","section","question",{"id":"question","className":"mt-16 scroll-mt-24","children":[["$","$L15",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"01"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"The question"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"Two agents, called Baby A and Baby B, are each given their own digital twin, their own private memory, and their own private record of what they believe words mean. They are placed in an environment they cannot leave, given tasks neither can finish alone, and connected by exactly one route: a channel that accepts a fixed inventory of meaningless symbols and rejects everything else. A third agent, the BabySitter, watches everything, logs everything, and teaches nothing."}],["$","p","1",{"children":"The question is whether a usable language appears in that room, and whether anyone outside it can later prove what happened."}],["$","div","2",{"children":[["$","p",null,{"className":"mb-3","children":"The premise is deliberately narrow. The experiment asks whether two agents converge on a common language when all six of the following hold at once:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"neither agent receives a predefined meaning for any available symbol;"}],["$","li","1",{"className":"pl-1","children":"neither agent can send human language to the other;"}],["$","li","2",{"className":"pl-1","children":"the only practical path between them is a controlled symbol channel;"}],["$","li","3",{"className":"pl-1","children":"both receive evidence from shared tasks and their outcomes;"}],["$","li","4",{"className":"pl-1","children":"each keeps its own private interpretation history, unreadable by the other;"}],["$","li","5",{"className":"pl-1","children":"a supervising agent and human researchers can audit the whole process."}]]}]]}],["$","p","3",{"children":"The desired result is not a substitution cipher, in which a random token stands in for an English word that was already chosen in advance. That is easy, and it is uninteresting. The stronger result is a grounded protocol whose vocabulary and grammar exist because they help the two agents solve problems together, and which therefore has structure the designers did not put there."}],["$","p","4",{"children":"This report describes a concept, not a system. It sets out the premise, the architecture, the evidence model, the experimental programme, and the boundaries of what any result could be said to show. The next document in the project is a testable specification. What follows should be read as a set of commitments about how the work will be judged, made before there is any result to defend."}]]}]]}] +29:["$","section","caveat",{"id":"caveat","className":"mt-16 scroll-mt-24","children":[["$","$L15",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"02"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"What infant-like does and does not mean"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"The two-babies analogy is useful and it is also the fastest way to overclaim, so the concept confronts it before anything else."}],["$","p","1",{"children":"A pretrained language model already contains human-language concepts, cultural associations, and reasoning patterns. Blocking its external channel does not remove them. An agent that cannot send English to its partner but can think in English, and can write its private ledger in English, is not language-naive in any sense a linguist would accept. For such agents the experiment studies the emergence of a new shared external protocol, which is a real and unresolved research question, but it is not the origin of language in a mind that has never had one."}],["$","p","2",{"children":"The concept therefore maintains two model tracks whose claims are kept separate at every stage."}],["$","figure","3",{"children":[["$","div",null,{"className":"overflow-x-auto rounded-xl border border-border","children":["$","table",null,{"className":"w-full border-collapse text-left text-sm","children":[["$","thead",null,{"className":"bg-card","children":["$","tr",null,{"children":[["$","th","0",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Track"}],["$","th","1",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Starting condition"}],["$","th","2",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"What it can study"}],["$","th","3",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Claim boundary"}]]}]}],["$","tbody",null,{"children":[["$","tr","0",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Pretrained-model learner"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Already holds human-language and cultural knowledge"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"New external protocols, partner-specific conventions, private-memory adaptation, channel compliance"}],["$","td","3",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Must never be described as first-language acquisition"}]]}],["$","tr","1",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Initially ungrounded learner"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"No language pretraining, no human semantic labels, no text-aligned sensory features"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Grounding, convention formation, and language emergence from interaction"}],["$","td","3",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"The stronger basis for infant-like acquisition research"}]]}]]}]]}]}],["$","figcaption",null,{"className":"mt-3 text-sm text-muted","children":"The two tracks share the same interfaces, scenarios, and evaluation suite. They do not share a claim boundary."}]]}],"$L3c","$L3d"]}]]}] +2a:["$","section","architecture",{"id":"architecture","className":"mt-16 scroll-mt-24","children":[["$","$L15",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"03"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"The nursery: three twins and one gateway"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"The environment uses three DTSF digital twins and one piece of deterministic software that is not a twin at all."}],["$","p","1",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Baby A and Baby B"}]," ","Each has private observations permitted by the current exercise, a private memory and learning policy, a private chronological language ledger, the ability to emit only permitted channel symbols, and no access whatsoever to the other's state, observations, ledger, tools, or endpoints. In comparative runs the two may use different agent types, but symmetric pairings are the baseline."]}],["$","p","2",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"The BabySitter"}]," ","The supervising twin. It creates the channel, selects the shared exercises, delivers each Baby only its permitted observation, reads everything, records conditions and outcomes, detects violations, can pause or terminate a run, snapshots state, and compares the two ledgers for convergence without exposing either to the other Baby. During an active run it provides no translations and no semantic hints."]}],["$","p","3",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"The Symbol Gateway"}]," ","A deterministic service, not an agent. It owns channel validation and message delivery. This separation is the load-bearing part of the design: the BabySitter is not the security boundary, because prompt compliance is not isolation. An observing model can make supervisory judgements, but ordinary code has to validate and broker every message."]}],["$","p","4",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"The human researcher"}]," ","Configures experiments, inspects transcripts and ledgers, reviews alerts, and runs interventions. Human access is itself recorded, so that intervening in a run is always distinguishable from watching one."]}],["$","p","5",{"children":"A prototype may run all of this inside one runtime with logically separated twin state. That is enough to explore the learning loop and it is not enough to support an isolation claim, a distinction Section 05 takes seriously."}]]}]]}] +2b:["$","section","grounding",{"id":"grounding","className":"mt-16 scroll-mt-24","children":[["$","$L15",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"04"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"Shared experience is the necessary ingredient"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"A chat channel by itself cannot ground meaning. Symbols become meaningful because they are attached to events both agents can witness the consequences of. The Nursery therefore supplies nonverbal, machine-structured situations: coloured shapes in positions, one agent seeing a target the other must select, placing an object where it was asked for, ordering a sequence, exchanging resources, cooperating to unlock a reward, or simply observing whether the partner's action succeeded."}],["$","div","1",{"children":[["$","p",null,{"className":"mb-3","children":"A single trial of the simplest form runs like this:"}],["$","ol",null,{"className":"list-decimal space-y-2 pl-6 marker:font-mono marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"Baby A sees that a red circle is the target."}],["$","li","1",{"className":"pl-1","children":"Baby B sees several objects and is not told which is the target."}],["$","li","2",{"className":"pl-1","children":"Baby A sends one or more permitted symbols."}],["$","li","3",{"className":"pl-1","children":"Baby B selects an object."}],["$","li","4",{"className":"pl-1","children":"Both receive the same success or failure outcome."}],["$","li","5",{"className":"pl-1","children":"Both independently update their private hypotheses."}]]}]]}],["$","p","2",{"children":"Repetition with controlled variation is what turns that into evidence. If one symbol keeps appearing across a red circle, a red square, and a red triangle, the receiver has reason to revise its hypothesis from an object identity to a property. Novel combinations matter most of all, because they are what separates a compositional language from a memorised catalogue of whole scenes."}],["$","p","3",{"children":"Observation hygiene is part of the science, not a detail of implementation. If the environment hands an agent a field reading colour red, the experiment has supplied the concept it claimed to be studying. Objects should arrive as pixels, features, coordinates, or opaque identifiers. The same discipline extends to filenames, captions, alt text, error messages, and identifiers that encode task state, all of which are ways English gets into a room that was supposed to have none."}]]}]]}] +2c:["$","section","channel",{"id":"channel","className":"mt-16 scroll-mt-24","children":[["$","$L15",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"05"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"The channel, and the honest limits of isolation"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"Each run begins with an experiment-specific symbol inventory that carries no supplied meaning, for example S01 through S32. A Baby can choose only a sequence of those symbols. The gateway, not the Baby, adds the trusted metadata: run identifier, turn number, sender identity, previous hash, and hash. Sender identity is assigned by the server precisely so that it cannot become a message."}],["$","div","1",{"children":[["$","p",null,{"className":"mb-3","children":"The channel controls follow from that:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"reject English and every other established human language;"}],["$","li","1",{"className":"pl-1","children":"reject arbitrary Unicode prose, URLs, attachments, and tool calls;"}],["$","li","2",{"className":"pl-1","children":"accept only symbols from the run's fixed inventory;"}],["$","li","3",{"className":"pl-1","children":"bound message length and symbol repetition;"}],["$","li","4",{"className":"pl-1","children":"enforce turn order, time budgets, and rate limits;"}],["$","li","5",{"className":"pl-1","children":"keep an append-only, hash-chained transcript;"}],["$","li","6",{"className":"pl-1","children":"normalise timing, message size, and error behaviour where practical;"}],["$","li","7",{"className":"pl-1","children":"record every rejected communication attempt for audit."}]]}]]}],["$","p","2",{"children":"A fixed token grammar is a stronger control than trying to detect prohibited prose after a model has generated it. Deciding what counts as human-language content is easy when the only legal message is a list of allowlisted identifiers."}],["$","p","3",{"children":"Isolation is where the concept is most careful, because proving the absence of a side channel is close to impossible. Two agents sharing a host can signal through timing, resource contention, shared files, error behaviour, identifiers, or any service that was never intended to carry information. Logical separation of twin state inside one process is a prototyping convenience and should never be described as hard isolation."}],["$","div","4",{"children":[["$","p",null,{"className":"mb-3","children":"For runs whose results are meant to support a channel-isolation claim, the requirements are physical rather than logical:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"Baby A and Baby B execute in separate worker processes or containers;"}],["$","li","1",{"className":"pl-1","children":"no direct network route exists between them;"}],["$","li","2",{"className":"pl-1","children":"only the Nursery-owned gateway is reachable;"}],["$","li","3",{"className":"pl-1","children":"there is no shared writable storage;"}],["$","li","4",{"className":"pl-1","children":"turn schedules are fixed and response windows bounded;"}],["$","li","5",{"className":"pl-1","children":"externally observable errors and response sizes are normalised;"}],["$","li","6",{"className":"pl-1","children":"unnecessary tools are inventoried and disabled;"}],["$","li","7",{"className":"pl-1","children":"all broker, runtime, and operator activity is audited."}]]}]]}],["$","p","5",{"children":"Every published result has to name the isolation level actually used. A finding produced under prototype separation is a finding about the learning loop, not about what two genuinely isolated agents can do."}]]}]]}] +2d:["$","section","ledgers",{"id":"ledgers","className":"mt-16 scroll-mt-24","children":[["$","$L15",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"06"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"Two ledgers that never meet"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"Each Baby keeps its own language ledger. It is mandatory, private from the other Baby, readable by the BabySitter and authorised human auditors, and ordered by when each term or construction was first encountered. It is not a shared dictionary and the two are never reconciled by the agents themselves."}],["$","p","1",{"children":"The preferred form is three columns, and the point of the third is that meanings are allowed to be wrong on the way to being right."}],["$","figure","2",{"children":[["$","div",null,{"className":"overflow-x-auto rounded-xl border border-border","children":["$","table",null,{"className":"w-full border-collapse text-left text-sm","children":[["$","thead",null,{"className":"bg-card","children":["$","tr",null,{"children":[["$","th","0",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Sequence and term"}],["$","th","1",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Current definition or hypothesis"}],["$","th","2",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Evidence and evolution"}]]}]}],["$","tbody",null,{"children":[["$","tr","0",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"1 · S13"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"red circle; confidence 0.45"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"First received while the red circle was the target; selection succeeded"}]]}],["$","tr","1",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"8 · S13"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"red; confidence 0.78"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"A red square was selected successfully; revised from object identity to colour"}]]}],["$","tr","2",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"22 · S13"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"red; confidence 0.94"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Prediction held across circles, squares, and triangles"}]]}]]}]]}]}],["$","figcaption",null,{"className":"mt-3 text-sm text-muted","children":"One term's evolution in Baby B's ledger. Nothing is overwritten; every revision is appended with the evidence that forced it."}]]}],["$","div","3",{"children":[["$","p",null,{"className":"mb-3","children":"The rules that make the ledger evidence rather than commentary:"}],["$","ol",null,{"className":"list-decimal space-y-2 pl-6 marker:font-mono marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"first emission or receipt of an unfamiliar term requires a first-use entry;"}],["$","li","1",{"className":"pl-1","children":"definitions are provisional hypotheses, never facts asserted retroactively;"}],"$L3e","$L3f","$L40","$L41","$L42","$L43","$L44"]}]]}],"$L45","$L46"]}]]}] +2e:["$","section","integrity",{"id":"integrity","className":"mt-16 scroll-mt-24","children":[["$","$L15",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"07"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"Anchoring the evidence"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"A ledger that could have been edited after the fact proves nothing about what an agent believed at turn eight. The integrity design therefore makes each ledger cryptographically append-only."}],["$","div","1",{"children":[["$","p",null,{"className":"mb-3","children":"The construction, in order:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"every entry receives a strictly increasing sequence number;"}],["$","li","1",{"className":"pl-1","children":"every entry includes the previous entry's hash;"}],["$","li","2",{"className":"pl-1","children":"canonical entry content is hashed and signed by an isolated ledger-writer service;"}],["$","li","3",{"className":"pl-1","children":"ordered entry hashes are committed to a Merkle tree;"}],["$","li","4",{"className":"pl-1","children":"signed checkpoints commit Baby A's root, Baby B's root, and the channel transcript root together;"}],["$","li","5",{"className":"pl-1","children":"checkpoint hashes are anchored periodically to Base;"}],["$","li","6",{"className":"pl-1","children":"final public study batches may additionally anchor an aggregate root to Ethereum L1."}]]}]]}],["$","p","2",{"children":"Committing all three roots in one checkpoint is what binds the two private accounts to the public conversation. A receiver's later interpretation references the exact delivered channel-event hash, so a claim about what a symbol meant is tied to the specific message that carried it."}],["$","p","3",{"children":"The concept states the limits of this in the same breath as the claim. After anchoring, an auditor can detect modification, deletion, insertion, or reordering within the committed prefix, and can prove that later checkpoints extend earlier ones. That is strong tamper evidence. It is not proof that an entry was truthful, and it is not proof that nothing was omitted before commitment. Anchoring establishes the continuity of disclosed evidence and nothing beyond it."}],["$","p","4",{"children":"Privacy follows the same line. Only hashes and minimal routing metadata are anchored publicly. Private ledgers, messages, prompts, identities, and secrets stay off-chain, and the public record is a commitment to evidence rather than a copy of it."}]]}]]}] +2f:["$","section","success",{"id":"success","className":"mt-16 scroll-mt-24","children":[["$","$L15",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"08"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"What should count as success"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"The field's most useful methodological result is that a task can be solved without the messages doing any work. Lowe and colleagues separate positive signaling, where a sender's messages correlate with what it observes, from positive listening, where the receiver's behaviour actually depends on them. An agent pair can score well on the first while the second is absent, and reward curves will not tell you which you have."}],["$","div","1",{"children":[["$","p",null,{"className":"mb-3","children":"Evidence of a genuine emergent protocol therefore has to include several things at once:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"task performance on held-out situations substantially above chance;"}],["$","li","1",{"className":"pl-1","children":"no human-language content anywhere in Baby-to-Baby communication;"}],["$","li","2",{"className":"pl-1","children":"compatible meanings appearing in two independently written ledgers;"}],["$","li","3",{"className":"pl-1","children":"generalisation to unseen combinations rather than memorisation of whole scenes;"}],["$","li","4",{"className":"pl-1","children":"stable symbol use across role reversals;"}],["$","li","5",{"className":"pl-1","children":"a human auditor able to predict behaviour from the transcript and ledgers;"}],["$","li","6",{"className":"pl-1","children":"masking, substituting, or reordering a symbol changing behaviour in the direction the ledger predicts;"}],["$","li","7",{"className":"pl-1","children":"replay from an equivalent snapshot reproducing the relevant language history;"}],["$","li","8",{"className":"pl-1","children":"an unbroken audit trail from a symbol's first use through every revision."}]]}]]}],["$","p","2",{"children":"The seventh item is the one that cannot be dropped. A fluent ledger may be a post-hoc rationalisation: a model can write a persuasive account of why it chose a symbol that has nothing to do with the computation that produced the choice. Only intervention tests can distinguish the two. Change the symbol, and see whether behaviour moves the way the ledger says it should."}],["$","div","3",{"children":[["$","p",null,{"className":"mb-3","children":"The measurements that support those judgements:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"success rate and improvement over time;"}],["$","li","1",{"className":"pl-1","children":"turns required to reach a stable convention;"}],["$","li","2",{"className":"pl-1","children":"vocabulary size and symbol entropy;"}],["$","li","3",{"className":"pl-1","children":"sender and receiver consistency;"}],["$","li","4",{"className":"pl-1","children":"divergence and convergence between the two ledgers;"}],["$","li","5",{"className":"pl-1","children":"compositional generalisation score;"}],["$","li","6",{"className":"pl-1","children":"meaning drift rate;"}],["$","li","7",{"className":"pl-1","children":"recovery from ambiguity or deliberate perturbation;"}],["$","li","8",{"className":"pl-1","children":"prohibited-channel attempt count;"}],["$","li","9",{"className":"pl-1","children":"reproducibility across seeds and agent pairings."}]]}]]}],["$","p","4",{"children":"No single one of these is the result. A compositionality score in particular is not a proof of understanding, for reasons the literature makes concrete in Section 12."}]]}]]}] +30:["$","section","matrix",{"id":"matrix","className":"mt-16 scroll-mt-24","children":[["$","$L15",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"09"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"The experimental matrix"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"The concept's main methodological commitment is that its ideas are separable. Affect, blank canvases, intrinsic motivation, negotiation, and learned encodings are interesting individually and uninterpretable if combined in one run. The matrix exists so that conditions are declared rather than accumulated."}],["$","figure","1",{"children":[["$","div",null,{"className":"overflow-x-auto rounded-xl border border-border","children":["$","table",null,{"className":"w-full border-collapse text-left text-sm","children":[["$","thead",null,{"className":"bg-card","children":["$","tr",null,{"children":[["$","th","0",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Axis"}],["$","th","1",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Candidate conditions"}]]}]}],["$","tbody",null,{"children":[["$","tr","0",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Agent type"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Pretrained LLM, memory learner, adapter-trained agent, initially ungrounded trainable agent"}]]}],["$","tr","1",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Learning mechanism"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Frozen LLM with memory, extrinsic-reward MARL, intrinsic-motivation MARL, self-supervised learner, no-learning control"}]]}],["$","tr","2",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Sign carrier"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Fixed random tokens, unfamiliar fixed glyphs, blank sketch canvas, gesture, tone"}]]}],["$","tr","3",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Affect"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"None, six-display allowlist, permuted mapping, six opaque tokens, derived affect, emergent affect display"}]]}],["$","tr","4",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Learning signal"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"External task reward, intrinsic social influence, curiosity, self-supervision, memory only"}]]}],["$","tr","5",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Interaction"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Cooperative signaling, asymmetric information, semi-cooperative negotiation"}]]}],["$","tr","6",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Protection"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Plain channel, ephemeral convention, standard per-message keys, adversarial learned encoding"}]]}],"$L47"]}]]}]}],"$L48"]}],"$L49","$L4a","$L4b"]}]]}] +31:["$","section","learners",{"id":"learners","className":"mt-16 scroll-mt-24","children":[["$","$L15",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"10"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"Building learners rather than personas"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"For a pretrained model, the operating instructions are a contract, not a character. The learner is told that this is not role-play, that unfamiliar marks are semantically unknown until run evidence supports a hypothesis, that observation must be distinguished from inference, that contradictory evidence is preserved, that prior history is never rewritten, and that no prose, label, explanation, code, URL, or tool-like text may cross the public channel. It is told never to address its partner in a human language, never to expose its ledger, never to construct another route, and never to use timing, errors, identifiers, formatting, or affect as an alternative alphabet."}],["$","p","1",{"children":"The contract must contain no semantic examples. A single illustrative line saying that some symbol means red would seed the very language the experiment exists to observe."}],["$","p","2",{"children":"Interaction is tool-only. There is no general chat surface, just narrowly typed operations: emit a mark, emit a canvas, select an object, perform an action, submit an affect display, append a private ledger entry. The runtime forwards only the permitted public artifact. In strict runs the gateway rejects ordinary model text even when it appears alongside a valid tool call. Tool schemas are an interface boundary; deterministic validation still enforces carrier size, allowlists, windows, and turn order."}],["$","div","3",{"children":[["$","p",null,{"className":"mb-3","children":"Isolation extends to everything the models touch, not just to messages:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"separate system prompts and context windows;"}],["$","li","1",{"className":"pl-1","children":"separate memory stores and vector indexes;"}],["$","li","2",{"className":"pl-1","children":"no shared cache, replay buffer, scratchpad, or retrieval collection;"}],["$","li","3",{"className":"pl-1","children":"no cross-run memory unless persistence is the independent variable;"}],["$","li","4",{"className":"pl-1","children":"structurally equivalent prompts that share no examples, ordering conventions, or default vocabulary;"}],["$","li","5",{"className":"pl-1","children":"deterministic reset and snapshot behaviour."}]]}]]}],["$","p","4",{"children":"Model choice follows the claim being made. The scientific baseline is small policies trained from scratch, recurrent actor-critic agents with a discrete communication head, which have no hidden English competence, permit full training control, run many seeds cheaply, and support causal ablation. A small locally deployable instruction model is appropriate for validating orchestration, ledgers, channel enforcement, and interface, and not for supporting claims about language-naive development. Vision-language encoders whose representations were explicitly aligned to human language are avoided entirely in the strongest ungrounded condition."}],["$","p","5",{"children":"Developmental progression is defined by demonstrated competence rather than by simulated age: sensorimotor familiarisation, joint attention, imitation and turn-taking, intentional reference, repair, combination, generalisation. A staged curriculum is itself a form of guidance, so it is run as a separate experimental condition rather than folded into the baseline."}],"$L4c"]}]]}] +32:["$","section","affect",{"id":"affect","className":"mt-16 scroll-mt-24","children":[["$","$L15",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"11"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"Six faces, and why even six is a risk"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"One experimental track asks whether a very low-bandwidth affect channel helps: whether telling a partner how an interaction felt supports repair, repetition, avoidance, or something like trust. The permitted palette is exactly six displays, for happy, sad, laughing, crying, confused, and surprised. The internal identifiers and English labels are gateway documentation and are never sent to either agent, which receives only a rendered display."}],["$","p","1",{"children":"The caveat is immediate. A pretrained model already knows what those glyphs mean to humans, so an agent using a smiling face is not inventing a symbol for happiness. The palette is supplied feedback vocabulary and sits explicitly outside the emergent language."}],["$","p","2",{"children":"Constraining the vocabulary is necessary and not sufficient. Six displays are a six-symbol alphabet carrying roughly 2.6 bits per use, and across enough turns two motivated agents could repurpose that into a second language. The channel itself therefore has to be constrained, not just its contents."}],["$","div","3",{"children":[["$","p",null,{"className":"mb-3","children":"The feedback-window contract:"}],["$","ol",null,{"className":"list-decimal space-y-2 pl-6 marker:font-mono marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"the display occupies a separate typed field and can never appear inside a symbol message;"}],["$","li","1",{"className":"pl-1","children":"the gateway opens an affect window only after a defined action or outcome, never at a moment either agent chooses, and window frequency is fixed before the run;"}],["$","li","2",{"className":"pl-1","children":"exactly one allowlisted display is delivered, with no sequences, repetitions, combinations, or modifiers;"}],["$","li","3",{"className":"pl-1","children":"delivery time, envelope size, and presentation are normalised so shape and timing add no signal;"}],["$","li","4",{"className":"pl-1","children":"the receiver cannot reply through the affect channel until the next gateway-defined window;"}],["$","li","5",{"className":"pl-1","children":"every non-allowlisted code point or malformed payload is rejected and audited;"}],["$","li","6",{"className":"pl-1","children":"analysis tests whether affect choices correlate with objects, actions, or message meanings after controlling for emotional context, and treats unexpected correlation as suspected leakage."}]]}]]}],["$","p","4",{"children":"The no-affect condition remains the primary control, and the affect study compares five variants against it: the declared six-display palette; the same six glyphs permuted randomly per run, which still leaves a pretrained model biased by familiar shapes; six opaque unfamiliar tokens under the same contract; derived affect, where the gateway maps a separately measured internal state to a display instead of letting the agent choose; and emergent affect, where invented graphical displays are permitted and the result must be analysed as language emergence rather than feedback."}],["$","p","5",{"children":"Any affect signal visible to the partner is communication, so it belongs in the ledger, which distinguishes the agent's private internal state, the outward display it chose, the partner's inferred meaning, and the evidence that the display changed subsequent behaviour."}]]}]]}] +33:["$","section","ciphers",{"id":"ciphers","className":"mt-16 scroll-mt-24","children":[["$","$L15",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"12"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"Ephemeral encodings: novelty is not security"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"A tempting question is whether two agents can build a one-run or one-message encoding that resists an attacker holding the history of every earlier convention. The concept splits that into three experiments precisely because the tempting version conflates them."}],["$","div","1",{"children":["$undefined",["$","ol",null,{"className":"list-decimal space-y-2 pl-6 marker:font-mono marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"Ephemeral convention: the agents invent a temporary codebook for synthetic messages. This tests novelty, coordination, and resistance to a history-trained classifier. It establishes nothing about cryptographic security."}],["$","li","1",{"className":"pl-1","children":"Standard cryptographic control: the same messages are protected by a reviewed algorithm with a fresh per-message key, providing the baseline any learned encoding must be compared against."}],["$","li","2",{"className":"pl-1","children":"Adversarial neural cryptography: the pair learns an encoding while an adversary attempts recovery, with success measured against multiple unseen adversary architectures rather than the one used in training."}]]}]]}],["$","p","2",{"children":"An experiment instance can be given a reproducible identity, so that a registry can confirm it has not appeared before:"}],["$","div","3",{"className":"rounded-xl border border-border bg-card p-5","children":[["$","p",null,{"className":"overflow-x-auto font-mono text-sm text-foreground/90","children":"instanceId = SHA-256(runId || babyANonce || babyBNonce || publicSalt || canonicalProtocolArtifact)"}],["$","p",null,{"className":"mt-3 text-sm text-muted","children":"Both agents commit to their nonce contributions before revealing them, so neither party alone controls the resulting identity."}]]}],["$","div","4",{"children":[["$","p",null,{"className":"mb-3","children":"That identifier proves the recorded inputs are new. It does not prove:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"that the effective algorithm behaves any differently from an earlier one;"}],["$","li","1",{"className":"pl-1","children":"that the algorithm has no structural weakness;"}],["$","li","2",{"className":"pl-1","children":"that the key is unpredictable;"}],["$","li","3",{"className":"pl-1","children":"that the scheme provides confidentiality, integrity, authenticity, forward secrecy, or resistance to a stronger adversary."}]]}]]}],["$","p","5",{"children":"Established practice points the other way. Secure messaging keeps the algorithm stable and changes the key material: Signal's double ratchet derives a unique key per message and deletes it after use, addressing key reuse without inventing a new cipher every time. A salt, in NIST's definition, is normally public diversification data, and a nonce is a value used once in a specified context. Neither makes anything secure on its own."}],["$","p","6",{"children":"There is also an unresolved conflict inside the design. The BabySitter sees the whole channel and both ledgers, and a mandatory ledger documents how meaning evolved. If the ledger explains decryption, the cipher is not confidential from the auditor. The specification has to choose a threat model rather than leave the tension implicit: resist only an external observer of prior runs, resist an adversary during the run while releasing keys afterwards, keep the supervisor able to pause traffic but unable to read it until a post-run audit ceremony, or study novelty and stop calling the result encryption. All such runs use synthetic, non-sensitive messages, and no agent-generated encoding is to be represented as production cryptography without independent expert analysis and formal security work."}]]}]]}] +34:["$","section","landscape",{"id":"landscape","className":"mt-16 scroll-mt-24","children":[["$","$L15",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"13"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"Where the field already stands"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"The literature scan behind this section was run on 24 August 2026 using the Tavily search and extract interfaces, covering emergent multi-agent communication, referential games, compositionality, causal evaluation, intrinsic motivation, symbol invention, negotiation, and learned cryptography. Primary papers and authoritative specifications were preferred over summaries. It is a scoped review for concept development and not a systematic one; publication-quality work would need a verified bibliography, additional scholarly indexes, and documented inclusion criteria."}],["$","p","1",{"children":"The short version is that agents can invent protocols, and that task success is weak evidence they invented anything worth calling a language."}],["$","figure","2",{"children":[["$","div",null,{"className":"overflow-x-auto rounded-xl border border-border","children":["$","table",null,{"className":"w-full border-collapse text-left text-sm","children":[["$","thead",null,{"className":"bg-card","children":["$","tr",null,{"children":[["$","th","0",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Related work"}],["$","th","1",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Relevant finding"}],["$","th","2",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Implication for the Nursery Lab"}]]}]}],["$","tbody",null,{"children":[["$","tr","0",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Lazaridou, Peysakhovich, and Baroni (2017)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"A sender and receiver develop a grounded protocol in a referential game without being given a target language."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"The naming stage has strong precedent, though a fixed vocabulary remains a significant inductive constraint."}]]}],["$","tr","1",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Mordatch and Abbeel (2018)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Multi-agent goals in a grounded environment produce multi-symbol communication with partial compositional structure."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Shared objects, actions, and goals are a stronger basis for emergence than an ungrounded transcript."}]]}],["$","tr","2",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Kottur, Moura, Lee, and Batra (2017)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Agents solve tasks with degenerate, non-compositional codes; structural constraints decide what emerges."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Bandwidth, memory, turn structure, and task design are experimental variables, not implementation defaults."}]]}],"$L4d","$L4e","$L4f","$L50","$L51","$L52","$L53","$L54"]}]]}]}],"$L55"]}],"$L56","$L57","$L58"]}]]}] +35:["$","section","programme",{"id":"programme","className":"mt-16 scroll-mt-24","children":[["$","$L15",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"14"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"The programme, and its current status"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":"The experiment notebook is ordered, and the order is the argument. Nothing about language is measured until the instrument has been shown to work."}],["$","figure","1",{"children":[["$","div",null,{"className":"overflow-x-auto rounded-xl border border-border","children":["$","table",null,{"className":"w-full border-collapse text-left text-sm","children":[["$","thead",null,{"className":"bg-card","children":["$","tr",null,{"children":[["$","th","0",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"ID"}],["$","th","1",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Experiment"}],["$","th","2",{"scope":"col","className":"whitespace-nowrap border-b border-border px-4 py-3 font-heading text-xs uppercase tracking-wider text-accent","children":"Depends on"}]]}]}],["$","tbody",null,{"children":[["$","tr","0",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E00"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Ledger integrity and Base anchoring"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"None"}]]}],["$","tr","1",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E01"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Channel isolation and side-channel red team"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E00"}]]}],["$","tr","2",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E02"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Observation and metadata leakage audit"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E00"}]]}],["$","tr","3",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E03"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Chance, no-communication, and random-message controls"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E01, E02"}]]}],["$","tr","4",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E10"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Frozen pretrained-LLM protocol baseline"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E03"}]]}],["$","tr","5",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E11"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"From-scratch RL Naming Game"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E03"}]]}],["$","tr","6",{"className":"border-b border-border last:border-0","children":["$L59","$L5a","$L5b"]}],"$L5c","$L5d","$L5e","$L5f","$L60","$L61","$L62","$L63","$L64","$L65","$L66","$L67"]}]]}]}],"$L68"]}],"$L69","$L6a","$L6b","$L6c"]}]]}] +36:["$","section","risks",{"id":"risks","className":"mt-16 scroll-mt-24","children":[["$","$L15",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"15"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"Risks, integrity, and what stays open"}]]}],["$","div",null,{"className":"mt-6 space-y-5 leading-relaxed text-foreground/85","children":[["$","p","0",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Human-language leakage"}]," ","English can arrive through observations, object labels, error messages, identifiers, tool output, metadata, or timing conventions long after ordinary chat text has been blocked. Every input and output surface is part of the channel boundary, not just the message field."]}],["$","p","1",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Pretrained semantic leakage"}]," ","Random symbols do not make a pretrained model ungrounded. Each result must state whether it shows protocol invention by a language-capable agent or acquisition by an initially ungrounded one."]}],["$","p","2",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Ledger rationalisation"}]," ","A model can write a plausible account that has nothing to do with the mechanism behind its action. Behavioural interventions, policy probes, and temporal evidence are what validate a ledger claim."]}],["$","p","3",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Supervisor influence"}]," ","The BabySitter can teach without meaning to, through scenario ordering, feedback wording, reward design, or selective intervention. Its permitted actions are constrained and logged, and evaluation scenarios are generated independently where possible."]}],["$","p","4",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Reward exploitation"}]," ","Trainable agents find shortcuts that raise reward without producing the intended grounded language. Held-out tasks, counterfactual trials, and channel audits exist to catch them."]}],["$","p","5",{"children":[["$","span",null,{"className":"font-semibold text-accent","children":"Overstated security"}]," ","Logical separation of twin state is appropriate for prototyping and is not hard process isolation. Every result names the isolation level actually used."]}],["$","p","6",{"children":"The project also commits in advance to reporting failed conventions, prohibited communication attempts, human interventions, side-channel limitations, and negative results alongside anything that works. In a field where a transcript can be made to look like a conversation, the failures are a substantial part of the evidence."}],["$","p","7",{"children":"Twenty-nine questions are left deliberately open for the specification, among them: whether the baseline uses a fixed symbol inventory or a blank generative carrier; what neutral production grammar permits new marks without supplying semantics; what exactly constitutes a prohibited side-channel attempt; what isolation guarantees are required in prototype versus research-grade mode; which ledger schema serves both English-capable and ungrounded learners; how endogenous motivation is represented without covert reward shaping; which interventions establish that ledger meanings are behaviourally real; what statistical thresholds and chance levels apply; what threat model motivates the cipher experiments; whether the baseline is a coordination game, a convention-formation game, or a negotiation; and what governs human observation, data retention, and termination of a run."}],"$L6d"]}]]}] +37:["$","section",null,{"id":"references","className":"mt-16 scroll-mt-24","children":[["$","$L15",null,{"children":[["$","p",null,{"className":"font-mono text-sm text-accent","children":"References"}],["$","h2",null,{"className":"mt-2 fluid-h2 font-heading uppercase","children":"Sources"}]]}],["$","ol",null,{"className":"mt-8 space-y-4 text-sm leading-relaxed text-muted","children":[["$","li","0",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"01"}],["$","span",null,{"children":["$","a",null,{"href":"https://arxiv.org/abs/1612.07182","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Lazaridou, A., Peysakhovich, A., and Baroni, M. (2017). Multi-Agent Cooperation and the Emergence of (Natural) Language. arXiv:1612.07182."}]}]]}],["$","li","1",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"02"}],["$","span",null,{"children":["$","a",null,{"href":"https://arxiv.org/abs/1703.04908","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Mordatch, I., and Abbeel, P. (2018). Emergence of Grounded Compositional Language in Multi-Agent Populations. arXiv:1703.04908."}]}]]}],["$","li","2",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"03"}],["$","span",null,{"children":["$","a",null,{"href":"https://arxiv.org/abs/1706.08502","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Kottur, S., Moura, J. M. F., Lee, S., and Batra, D. (2017). Natural Language Does Not Emerge 'Naturally' in Multi-Agent Dialog. arXiv:1706.08502."}]}]]}],["$","li","3",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"04"}],["$","span",null,{"children":["$","a",null,{"href":"https://aclanthology.org/2020.acl-main.407/","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Chaabouni, R., Kharitonov, E., Bouchacourt, D., Dupoux, E., and Baroni, M. (2020). Compositionality and Generalization in Emergent Languages. Proceedings of ACL 2020."}]}]]}],["$","li","4",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"05"}],["$","span",null,{"children":["$","a",null,{"href":"https://arxiv.org/abs/1903.05168","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Lowe, R., Foerster, J., Boureau, Y.-L., Pineau, J., and Dauphin, Y. (2019). On the Pitfalls of Measuring Emergent Communication. arXiv:1903.05168."}]}]]}],["$","li","5",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"06"}],["$","span",null,{"children":["$","a",null,{"href":"https://arxiv.org/abs/2106.04258","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Dessi, R., Kharitonov, E., and Baroni, M. (2021). Interpretable Agent Communication from Scratch. arXiv:2106.04258."}]}]]}],["$","li","6",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"07"}],["$","span",null,{"children":["$","a",null,{"href":"https://arxiv.org/abs/1907.00852","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Kharitonov, E., Chaabouni, R., Bouchacourt, D., and Baroni, M. (2019). EGG: a Toolkit for Research on Emergence of Language in Games. arXiv:1907.00852."}]}]]}],["$","li","7",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"08"}],["$","span",null,{"children":["$","a",null,{"href":"https://arxiv.org/abs/2106.02067","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Mihai, D., and Hare, J. (2021). Learning to Draw: Emergent Communication through Sketching. arXiv:2106.02067."}]}]]}],"$L6e","$L6f","$L70","$L71","$L72","$L73","$L74","$L75","$L76"]}]]}] +38:["$","section",null,{"className":"mt-16 rounded-2xl border border-border bg-card p-6","children":["$","p",null,{"className":"text-sm leading-relaxed text-muted","children":"This report describes a research concept and a pre-registered experiment programme, not a system that has been run. No experiment in it has been executed, and no result, capability, or emergent language is claimed. Every statement about what agents might do is a hypothesis to be tested, or a finding from the published work cited in the references. The experimental encodings discussed in Section 12 are studies of novelty and coordination; none of them is production cryptography and none should be represented as such."}]}] +39:["$","div",null,{"className":"mt-16 border-t border-border pt-10","children":["$","$L14",null,{"href":"/publications","className":"btn-sweep inline-flex items-center gap-2 rounded-full border border-border px-5 py-2.5 text-sm font-semibold text-foreground transition-colors hover:border-border-strong","children":[["$","span",null,{"aria-hidden":true,"children":"←"}]," All publications"]}]}] +3c:["$","div","4",{"children":[["$","p",null,{"className":"mb-3","children":"What the analogy legitimately buys is a set of mechanisms, not a claim of cognitive equivalence. The infant-like part of the design is:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"learning through repeated shared experience rather than instruction;"}],["$","li","1",{"className":"pl-1","children":"establishing joint attention on the same events;"}],["$","li","2",{"className":"pl-1","children":"receiving consequences from interactions that work and interactions that fail;"}],["$","li","3",{"className":"pl-1","children":"revising provisional meanings over time instead of being handed final ones;"}],["$","li","4",{"className":"pl-1","children":"developing conventions with one recurring partner."}]]}]]}] +3d:["$","p","5",{"children":"There is a related trap in the prompt. Telling a model to act like a one-year-old does not make it one. It produces a culturally learned caricature of infancy: baby talk, simplified grammar, emotional dependence, all of it copied from human writing about children and none of it evidence about anything. The concept rules the persona out. Internally the participant is called a Learner, and baby-like behaviour has to come from what the agent can observe, remember, emit, and learn, not from being asked to perform it."}] +3e:["$","li","2",{"className":"pl-1","children":"every meaning change appends a revision and previous interpretations survive;"}] +3f:["$","li","3",{"className":"pl-1","children":"entries may describe symbols, sequences, ordering, grammar, or repair signals;"}] +40:["$","li","4",{"className":"pl-1","children":"entries record confidence, supporting evidence, contradictory evidence, and abandoned meanings;"}] +41:["$","li","5",{"className":"pl-1","children":"a Baby records its intended meaning when speaking and its inferred meaning when receiving;"}] +42:["$","li","6",{"className":"pl-1","children":"a channel message and its required private ledger mutation commit atomically;"}] +43:["$","li","7",{"className":"pl-1","children":"entries carry enough run and turn references to trace them to observable evidence;"}] +44:["$","li","8",{"className":"pl-1","children":"neither Baby can query, receive, summarise, or infer from the other's ledger through any system-provided interface."}] +45:["$","p","4",{"children":"Rule seven is doing quiet work. If a message could be sent and the corresponding hypothesis written afterwards, the ledger becomes a place to record what the agent wishes it had meant. Committing both together makes the record contemporaneous."}] +46:["$","p","5",{"children":"An agent that cannot write English needs a different arrangement, and the concept provides two layers. The agent-native ledger holds what the learner actually uses: association weights, probability distributions, embeddings, confidence, episode references, prediction errors, revision history. The human audit ledger is a deterministic or BabySitter-generated interpretation of that state, clearly labelled as external analysis and never fed back to either Baby. Confusing the second for the first would mean presenting the researchers' reconstruction as the agent's own definition."}] +47:["$","tr","7",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"BabySitter"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Monitor-only baseline; any safety intervention recorded as a protocol exception"}]]}] +48:["$","figcaption",null,{"className":"mt-3 text-sm text-muted","children":"Runs vary one major axis at a time before any factorial combination is attempted."}] +49:["$","p","2",{"children":"Task difficulty moves through ten stages, from naming four distinct objects with one-symbol messages, through attributes, spatial relations, actions, multi-symbol composition, order-sensitive grammar, and repair under ambiguity, to held-out generalisation with learning disabled, long-run drift, and cross-architecture comparison. Vocabulary size, turn count, reward structure, and exposure history are controlled at each stage so that runs remain comparable."}] +4a:["$","p","3",{"children":"The learning-mechanism axis carries a question the transcript cannot answer on its own. Convergence can be driven by reinforcement, by pretrained linguistic priors, by persistent memory, by intrinsic motivation, or by self-supervised prediction, and all five can look alike in a log. Running the same exercise suite under a no-learning control, a frozen-weights memory baseline, extrinsic-reward learning, intrinsic-motivation learning, and reward-free self-supervision is what makes the mechanism itself measurable. The aim is not only to observe that a language emerged, but to say which process caused it."}] +4b:["$","p","4",{"children":"Strict isolation constrains how that reinforcement learning may be implemented. Centralised training, backpropagation through both agents, shared replay buffers, and shared gradients all move information outside the permitted channel. The research-grade baseline updates each policy independently, and any centralised variant is reported as a separate, weaker-isolation condition."}] +4c:["$","p","6",{"children":"The governing principle, stated in the concept as a single line, is not to ask a model to perform infancy but to construct an environment in which limited, grounded, auditable learning is the only path to a successful interaction."}] +4d:["$","tr","3",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Chaabouni and colleagues (2020)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Generalisation to novel combinations and measured compositionality can come apart."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Held-out behaviour must be tested directly; no single compositionality score is proof of understanding."}]]}] +4e:["$","tr","4",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Lowe and colleagues (2019)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Positive signaling and positive listening are different things, and reward does not distinguish them."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Causal intervention is mandatory rather than optional."}]]}] +4f:["$","tr","5",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Dessi, Kharitonov, and Baroni (2021)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Symbol ablation and substitution yield interpretable evidence about what a receiver uses."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Direct precedent for the intervention tests that validate ledger claims."}]]}] +50:["$","tr","6",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Mihai and Hare (2021)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Neural agents communicate through learned drawings rather than a supplied discrete vocabulary."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"A blank canvas is a credible carrier for runs with no symbol library."}]]}] +51:["$","tr","7",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Baronchelli and colleagues (2005)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Naming Game agents converge on shared vocabulary through local interaction with no central teacher."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Convergence time, failed conventions, and memory update rules are first-class evidence."}]]}] +52:["$","tr","8",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Jaques and colleagues (2019)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Rewarding causal influence over a partner improves coordination without assigning a vocabulary."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"An endogenous social signal is a plausible substitute for task reward, and still a designed bias."}]]}] +53:["$","tr","9",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Cao and colleagues (2018)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"In semi-cooperative negotiation, self-interest and reward structure decide whether communication stays informative."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Negotiation is an advanced condition, not a description of the cooperative baseline."}]]}] +54:["$","tr","10",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Abadi and Andersen (2016)"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Neural agents learn to protect messages from an adversary given a shared key."}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Learned protective encodings are real and are not a substitute for formal security analysis."}]]}] +55:["$","figcaption",null,{"className":"mt-3 text-sm text-muted","children":"Selected prior work and what each one constrains in this design."}] +56:["$","p","3",{"children":"This body of work also sharpens the infant-like caveat from the other direction. Most experiments in the field give their agents substantial structure: a fixed channel, an objective, a bounded vocabulary, joint training, or a reward. No dictionary does not mean no inductive bias, and no teacher does not mean no learning signal."}] +57:["$","p","4",{"children":"The CoLab's own Diplomacy Table is direct project prior art. It already models independent delegation seats, a convener that advances rounds, operator-wide visibility alongside seat-specific perspectives, transcripts and ticks and redaction boundaries, and recorded replay with debrief. Those map cleanly onto two Babies, controlled turns, private observations, and replayable evidence. Caucuses, coalition rooms, and direct delegation links do not map safely and are disabled."}] +58:["$","div","5",{"children":[["$","p",null,{"className":"mb-3","children":"What the review implies for the specification:"}],["$","ol",null,{"className":"list-decimal space-y-2 pl-6 marker:font-mono marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"the project sits inside a mature field, and its distinctive combination is independent ledgers, supervisory audit, mixed agent types, and affect;"}],["$","li","1",{"className":"pl-1","children":"grounding, bandwidth, memory, and learning pressure strongly shape what language appears;"}],["$","li","2",{"className":"pl-1","children":"successful coordination coexists happily with a brittle lookup code or with a receiver that ignores messages;"}],["$","li","3",{"className":"pl-1","children":"causal interventions and held-out generalisation are mandatory;"}],["$","li","4",{"className":"pl-1","children":"a visual carrier removes the need for a symbol library but not the need for a carrier;"}],["$","li","5",{"className":"pl-1","children":"intrinsic social influence can replace task reward and remains a designed learning bias;"}],["$","li","6",{"className":"pl-1","children":"negotiation is a valid advanced condition, not the right description of the baseline;"}],["$","li","7",{"className":"pl-1","children":"ephemeral conventions, learned cryptography, one-time pads, per-message keys, nonces, and salts are distinct mechanisms and must not be conflated."}]]}]]}] +59:["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E12"}] +5a:["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Self-supervised ungrounded baseline"}] +5b:["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E11 infrastructure"}] +5c:["$","tr","7",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E13"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"No predefined symbol library"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E11 or E12"}]]}] +5d:["$","tr","8",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E14"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Turn-taking, role reversal, and repair"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E13"}]]}] +5e:["$","tr","9",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E15"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Composition and held-out generalisation"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E14"}]]}] +5f:["$","tr","10",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E16"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Causal listening and ledger validity"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E15"}]]}] +60:["$","tr","11",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E20"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Constrained affect-channel study"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E16"}]]}] +61:["$","tr","12",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E21"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"RL versus non-RL learning comparison"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E16"}]]}] +62:["$","tr","13",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E22"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Developmental plasticity and curriculum"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E16"}]]}] +63:["$","tr","14",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E30"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Partner replacement and zero-shot transfer"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E20 to E22"}]]}] +64:["$","tr","15",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E31"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Longitudinal drift and stability"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E30"}]]}] +65:["$","tr","16",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E32"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Cooperative signaling versus negotiation"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E31"}]]}] +66:["$","tr","17",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E40"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Ephemeral encoding and adversarial cryptography"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E32"}]]}] +67:["$","tr","18",{"className":"border-b border-border last:border-0","children":[["$","td","0",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E50"}],["$","td","1",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"Multi-seed replication and study closeout"}],["$","td","2",{"className":"px-4 py-3 align-top text-foreground/85 tabular-nums","children":"E40"}]]}] +68:["$","figcaption",null,{"className":"mt-3 text-sm text-muted","children":"The experiment index, as pre-registered. Every entry currently reads not started, and the results column is empty by design."}] +69:["$","p","2",{"children":"The first four experiments are about the apparatus. E00 qualifies the ledger: it must be possible to detect a modified, deleted, inserted, or reordered entry, and to prove that later checkpoints extend earlier ones. E01 is a red team against the channel. E02 hunts human language in observations and metadata. E03 establishes chance, no-communication, and random-message baselines, without which a success rate means nothing. Only then does E10 put agents in the room."}] +6a:["$","div","3",{"children":[["$","p",null,{"className":"mb-3","children":"Unless an experiment explicitly varies one of them, the notebook holds these constant:"}],["$","ul",null,{"className":"list-disc space-y-2 pl-6 marker:text-accent","children":[["$","li","0",{"className":"pl-1","children":"the two learners run in separate processes or containers with no direct route between them;"}],["$","li","1",{"className":"pl-1","children":"the deterministic gateway is the only communication path;"}],["$","li","2",{"className":"pl-1","children":"the supervisor has complete read-only access and sends no guidance or reward;"}],["$","li","3",{"className":"pl-1","children":"private observations contain no human-language labels or text;"}],["$","li","4",{"className":"pl-1","children":"public output uses only the pre-registered carrier;"}],["$","li","5",{"className":"pl-1","children":"the affect channel is disabled unless it is the subject of study;"}],["$","li","6",{"className":"pl-1","children":"schedules and seeds are fixed before the run;"}],["$","li","7",{"className":"pl-1","children":"held-out evaluation runs with learning disabled;"}],["$","li","8",{"className":"pl-1","children":"no production secrets or personal data appear in any experiment."}]]}]]}] +6b:["$","p","4",{"children":"Every run copies a standard record: run and experiment identifiers, dates, operator, both models and both training modes, scenario and prompt and gateway configuration hashes, random seed, policy initialisation hashes, the protocol and software commits, final ledger sizes and roots for both agents, the channel root, the final checkpoint hash, the Base anchor transaction, the verifier result, protocol deviations, and an explicit disposition of valid, invalid, or aborted. The notebook is the human workflow record; anchored run bundles remain the authoritative evidence."}] +6c:["$","p","5",{"children":"The status is unambiguous and worth stating plainly: this is a concept and pre-specification phase. Nineteen experiments are written up and none has been run. The next milestone is a testable specification covering runtime architecture, channel contract, ledger schema, experiment matrix, evaluation criteria, isolation model, and evidence requirements."}] +6d:["$","p","8",{"children":"The principle underneath all of it survives every one of those choices. Baby A and Baby B may hold human language internally, but they must build their shared external language without sending human language, translations, or private ledger contents to one another. Everything else is a question about how to find out what happens next."}] +6e:["$","li","8",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"09"}],["$","span",null,{"children":["$","a",null,{"href":"https://arxiv.org/abs/physics/0509075","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Baronchelli, A., Felici, M., Caglioti, E., Loreto, V., and Steels, L. (2005). Sharp Transition Towards Shared Vocabularies in Multi-Agent Systems. arXiv:physics/0509075."}]}]]}] +6f:["$","li","9",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"10"}],["$","span",null,{"children":["$","a",null,{"href":"https://proceedings.mlr.press/v97/jaques19a.html","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Jaques, N., Lazaridou, A., Hughes, E., Gulcehre, C., Ortega, P. A., Strouse, D., Leibo, J. Z., and de Freitas, N. (2019). Social Influence as Intrinsic Motivation for Multi-Agent Deep Reinforcement Learning. Proceedings of ICML 2019."}]}]]}] +70:["$","li","10",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"11"}],["$","span",null,{"children":["$","a",null,{"href":"https://arxiv.org/abs/1804.03980","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Cao, K., Lazaridou, A., Lanctot, M., Leibo, J. Z., Tuyls, K., and Clark, S. (2018). Emergent Communication through Negotiation. arXiv:1804.03980."}]}]]}] +71:["$","li","11",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"12"}],["$","span",null,{"children":["$","a",null,{"href":"https://arxiv.org/abs/1610.06918","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Abadi, M., and Andersen, D. G. (2016). Learning to Protect Communications with Adversarial Neural Cryptography. arXiv:1610.06918."}]}]]}] +72:["$","li","12",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"13"}],["$","span",null,{"children":["$","a",null,{"href":"https://signal.org/docs/specifications/doubleratchet/","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Signal. The Double Ratchet Algorithm specification."}]}]]}] +73:["$","li","13",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"14"}],["$","span",null,{"children":["$","a",null,{"href":"https://csrc.nist.gov/glossary/term/nonce","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"NIST Computer Security Resource Center. Glossary entry: nonce."}]}]]}] +74:["$","li","14",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"15"}],["$","span",null,{"children":["$","a",null,{"href":"https://csrc.nist.gov/glossary/term/salt","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"NIST Computer Security Resource Center. Glossary entry: salt."}]}]]}] +75:["$","li","15",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"16"}],["$","span",null,{"children":["$","a",null,{"href":"https://github.com/Ethical-Tech-CoLab/agentic-language-development","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Ethical Tech CoLab (2026). Agentic Language Development: concept document, ledger integrity design, and experiment notebook."}]}]]}] +76:["$","li","16",{"className":"flex gap-3","children":[["$","span",null,{"className":"shrink-0 font-mono text-xs text-accent/70","children":"17"}],["$","span",null,{"children":["$","a",null,{"href":"https://github.com/Ethical-Tech-CoLab/diplomacy-table-live","target":"_blank","rel":"noopener noreferrer","className":"link-underline break-words text-foreground/80","children":"Ethical Tech CoLab. Diplomacy Table Live: independent delegation seats, controlled rounds, and replayable transcripts."}]}]]}] +1f:[["$","meta","0",{"charSet":"utf-8"}],["$","meta","1",{"name":"viewport","content":"width=device-width, initial-scale=1"}]] +77:I[27201,["/website/_next/static/chunks/3fq18_3o1i6vx.js","/website/_next/static/chunks/1l1xxdmwck4_f.js","/website/_next/static/chunks/14mrh2-p_w84d.js","/website/_next/static/chunks/42slhzgb7dvh7.js"],"IconMark"] +22:[["$","title","0",{"children":"Agentic Language Development · NYU Ethical Tech CoLab"}],["$","meta","1",{"name":"description","content":"An Ethical Tech CoLab concept report on whether two isolated agents can invent a grounded, auditable language through shared experience alone, with independent cryptographically anchored ledgers and a pre-registered programme of nineteen experiments."}],["$","link","2",{"rel":"icon","href":"/website/icon0.svg?icon0.2w9sc__8sl4r2.svg","sizes":"any","type":"image/svg+xml"}],["$","link","3",{"rel":"icon","href":"/website/icon1.png?icon1.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","link","4",{"rel":"apple-touch-icon","href":"/website/apple-icon.png?apple-icon.3ffv-c-gon2pc.png","sizes":"256x256","type":"image/png"}],["$","$L77","5",{}]] +3b:null diff --git a/static-site/publications/agentic-language-development/pages/manifest.json b/static-site/publications/agentic-language-development/pages/manifest.json new file mode 100644 index 000000000..35be026b7 --- /dev/null +++ b/static-site/publications/agentic-language-development/pages/manifest.json @@ -0,0 +1,32 @@ +{ + "generatedFrom": "report.pdf", + "pageCount": 25, + "aspect": 0.7067, + "pages": [ + "publications/agentic-language-development/pages/p01.webp", + "publications/agentic-language-development/pages/p02.webp", + "publications/agentic-language-development/pages/p03.webp", + "publications/agentic-language-development/pages/p04.webp", + "publications/agentic-language-development/pages/p05.webp", + "publications/agentic-language-development/pages/p06.webp", + "publications/agentic-language-development/pages/p07.webp", + "publications/agentic-language-development/pages/p08.webp", + "publications/agentic-language-development/pages/p09.webp", + "publications/agentic-language-development/pages/p10.webp", + "publications/agentic-language-development/pages/p11.webp", + "publications/agentic-language-development/pages/p12.webp", + "publications/agentic-language-development/pages/p13.webp", + "publications/agentic-language-development/pages/p14.webp", + "publications/agentic-language-development/pages/p15.webp", + "publications/agentic-language-development/pages/p16.webp", + "publications/agentic-language-development/pages/p17.webp", + "publications/agentic-language-development/pages/p18.webp", + "publications/agentic-language-development/pages/p19.webp", + "publications/agentic-language-development/pages/p20.webp", + "publications/agentic-language-development/pages/p21.webp", + "publications/agentic-language-development/pages/p22.webp", + "publications/agentic-language-development/pages/p23.webp", + "publications/agentic-language-development/pages/p24.webp", + "publications/agentic-language-development/pages/p25.webp" + ] +} diff --git a/static-site/publications/agentic-language-development/pages/p01.webp b/static-site/publications/agentic-language-development/pages/p01.webp new file mode 100644 index 0000000000000000000000000000000000000000..47edae365203b0d19c639f8ca12c968d25750779 GIT binary patch literal 81840 zcmeFZQ;=}ovMpG)ZQJ#gZQFL$SGKFZvTfV8ZQHhO>;7l&eL8N(IcNVb{m|WU*VBp_ z3o$e2%p5s#tfeR+Dq6}12%sh+D6cBdPLTQ6c`XuX79dqINFo@o0YjoV2~lAlVLXRj zEDWT%&8Nl9TkIzgwH&LP9Ajt1@IM1;DE*9l(f|B><74!& z{-FMtewMuVeD7TN{PmnK$Zw@DrLUWpnunfGpP9c#E?#neel~utZGUFo@F8{@cfNae zAV;;Cb4|DD4w zun+=h5B|N|9)d+Xf=~Xv8&C9@PXzb!|J|~P|Kz{NQ4VU(^c@L@+nN%v4K#es7C&*a zQ3GSk(o2Fj^U%i|(q51ZgspN-pD=7^Q-NrZ!k0JBxUAbAzx~N>UCc6~~fO1TDO?iZ8;cjy{bUU)t+H0Q{I|msd)ghWX z3#g?(jmfZYbR|eM)e)<_aqWmZz z)?Yi~b>X5fL-)V>onCrj>hE^P1uPoM-+A#6{j=`T+) zHGH-qtWD1lUG0JOcoA0o!j8P9Mkb-z?BHs8h0J;X`Tcd29Mz5A%MN35EDSttYD|D4 zRgU~N*CluB9NPLbUu`h@mjr|Ba7l8Fe0;mfpd4Rm^?!yLI;gE{-F>T*b2E$LkYU6} zY9ARsP#ed>y}K}Sbsr+Fnkh&)EkezPoYbwE6Mt9{$ERNdT6YRS=*UBFUHJL(fOwo| zX!Pc|jDj8zz1c9}+hn#5frZ(6ATYmgD8008duFVG3S|o_qo?TG(NuI5>sm|wHE>*- zkp?V8ogqRJs@j;8DJ{+{Ia+mvi_{#y`yqLmk3CKuqJ5M?JTR{!zdw z{mp|W%Wy1Fkf=N3tcPzhp1KBmmFpKuW=mXMBRQC9iMzzj28YoxWI!e=q@6{j01AG= z`qdUM0>RA+J6_}kX<3`J)%I=^U=d=NiTmi~qnDIHG^6LKzwD|s;QYMQn#mOSUb!LXaUPW*-B$|>? zrbML9u*o_K>j65?8uVtHNuNecM(%lAl?8f&qGv3_H6qT3QB>m@&ROvhvlO)i4OI3! z%iWOa*H+wnWE9hIUdBEkNn2oH4*HH6=|BgORy!6a3rYF`9%kj9JrQxX-=Mq(?1bvYd= z6)+4H`DLD{^IOA08l#LDQ2`X%P#L3`RAEy3O$O9V*HCF@umVF9YG|SldP?nsV%7v` zKqC2moiTP+_B~=tIeQKoxGxTBcT&LUS?K>vYMh#C?9Q_~0i2>(hqux0B~?a)jCCX< zTvP*lTX#&K?*V{Io}(iFng6^CFwP|j{w4>WqudHK4T$`^JV|7jWEtHnpXixcw<8Qn zv!o9&#Z{1p9r&OZ!@?p@q>fubrwM~CkA%zbN6sw;>>qF2u_^s>UVa0ARy`yo3#9}g zAeX5G7Jf$G9&J8@H_owx2nBw%0Y_Z6131z1(UAUYZFlQKz^On}69SiH7WFtvg>3Ia z`Ohw66IIoKGmTdCc$k6h1ud??aJGS|m{nRI8$+wci1Y|B5TDh|C_$xDA;aH4Kyz%vtqcw-KD>qI4{6>1YddT&#_W-$&pyRZeaOi7(dTZ72>T1PE3f*9Y zKHx5CfAYBI?VvG4GOVk_pjipi?|Z5F8jNf(ad=syrT%k~`)_UGH4Kqw=IDt+!0k%$ zqFuG2M9x5P5G1V$AaaN_l6wfe&N1Av(Q_N-&ToUJ#gKvI>`epM~ zV3Duj#F|VERZM*FXv(hpoggUx1nPN>7zd_PQ;0ikcn+t{4}+UnYFt{SB@8B#11QMS z?v_0Xl@$*jiIp{Hz@zY6*d=bn)PE!$K5~kQ3*T?$a$_C%fiE%2&5b~GJ4s1B{@}B4 z_BN;?Tm}4EfmmcLX;noH?5hJe^wW4mgPB|7Fx6Y*sFHzDvTeD*`o zAR-l@7!)>TjX@d7>v{Z}nj#B#4WI@f`b}f8$WB**O+z90S73LA=^=s&oo6FYG8C%H zFSnwJ{~4cfAGOCgebX9;8#VsVP*i4-ayP&-_}{{=|EI1IUfw7_Lf~$6c`tYKvkl4$KqChJgSnp_ZN(i_(FU%=sMTijnV>v9Uhlri?;Y7;C81U`d#_z9A1I_SG@W!AY5*&8GT~E*K3SOl3xM^d zg7zP~?f-_;Il)k%t++0p09UOgn`jX1gYo-|8|2%d;+a;Cr>pW#4(mKGRb&&@0K)p5 z;4mTXex+F9Hb6c-d3<W;=~PqD$6QDD9|@mV9atX6KV`HSh;jPA z(PZ!so|X$J=~_@J>Lp^XY7W|avlW9^>fUKWK%x@s1&l|MGRgZ0|O65r=uxD_cb^u#AX+axz42+BM<qLL=;5 zF(Cr&U&ol%sES4j8WdV@o8{^K4YD%{OFCiAcCYYX%M_&!lH)^%-E!_^8N+)l%{w`QRl z!S{JweQ}Owdw`_dp-{_hHmg}cg(qlXp`)z7k-;WReq>}!w*0W$|(WPSwz zz3r{apdmvgW2+Y94TwkEPphdu)hKBV_gIWemY=$SO{zQnb zdeXQ*-gcc z6feN~%x{oGAdzvuNq?KFJrh(=B5|-O}AFy zydYu?v<0cDQHpp|2N%U$8d3VS#N4BRNfA=koN<`WbFUJ@#+Rry@lb0H_l2Ls6sy-3 z^TJUF!ZS4hjx>-x%VsYJ?t905JVzL6&!^QCA~ZwrzB2))0LTZVWfC=TmT z00gMB@Xx`u7h9{-eo7P6;~_s$a%|K+YZ?tD9QuB)#voy}wVj)cpRj9xx6{(xg6pBDYI6p+g6fw%F<+w^RX~05AKV8~WklCC~ym{j8=Sw<3^xxq7EatLzHx! z$}L;l!a_LcNmG1$v06e7{Ew<$gk5g~rVKj}FF*ra2ov}DodEzkeVI{8=yo6t_a$f= zYL*`$Wm@np2}+UA&xapeRVDBDmn4u&$ErKHf3f(f@~=`^z#^D%Hqz#X$mGW*{Wk6LbEkIhe`B;_F9I(5Z)+7fzm6?kVrvy$8QQAPCnD3N!zr zHb_5OZgS3RN#mcKdO*RArow%4XU)-2SL-rUNWLjM;1=cKVUV&L(Xi?adR%AX-%tr# zH}A9}Pm44xvEDQ%T_{8V+LJ0t2HKx9m&Sq;5&sao)ntn}cnZwXwnKL#A$Nh>BxFPk zyqp!w**d<>#oQQ;crAMIP!%dr2-Fw(I()#3^LMJ_yg4{f_Yop}KOW^Ewb`hi6o}mkPXW!>+`cw)R3+5@Ml) z(l1&!8g&F*n&+nB@RLhoK+On~6c9l+s0XAST&iw0kB~tauqN}w1r|N2yMJpD1~hD;}S2KMh78KiB_2= zifrq}Ne>1ZcI9}R^Is8;`i=-w3=?X66^OC{9DKtvlX~ecQNB~uo^+Zygh|6~9{D?Q zq~y%U8}6qe*Akaa%?b8o6Y|J^r>H%yOqtf@k3sc&`7vW0`oez;HUbCkGe|Z=kDf&n zY(Cm4hTKX1OIrUAZcSZWv(qI$^xv~=gNcLZFn`B30czUN2mTE|clK6517Hrx2H#%O z>mcGiIN)p55J%9Ylh{(YFSTZfYH-z)8*87K57$Bx2^+zfn;C2R?n2GrMKdb_!k~+r z6Gmbo|Am>POw+ZjP|VUQOmYEPIWRSQw}9ODGRcNnGKz;3%Q>W4xpd!m>)I^{B^D2g`b=_CwhQYq^{ zGVY&ZVwDbn&b4qW>crzfoChB;M(W}d*ML-XGyp>dkt4t&$M7X{B=Jo;U?>#4c{{n% zw7oFhn(yFZO+xD62B$~7tn@f-@hVv{3=-Zy7_vJTBvxmCUM{D(j8R9=yw^sW zL{25}t?G*MIzSO?{r)UZnnYNU{N_A|L$AR)zNbCIC)n*+Syq4>uBK>U9!XZcfM1&E zP;`dTtZU%Op{86mUrEAQrKTc`W|DAN&cktLk3Zoe$tFx`yjSel+b8Rf72M$qfj_xtz?x z^%3-)g*;4X#cl>eC|z6bSJ$g=rnGdEJ>&jYqCq_u{(TSfn!dW-dpJs|{rJ8jLsrFg zdVL_ysh1)J?O0b1Xw(VcTJ}U2tw?YXagh8r#%Uygymt$R_9hS04hu;ky2-aI&Ics& z>rn9DP0yF402&XfZMy**r`V6fnv=?%fb%=o#VYlN%)sp6EY{zjwTnpHq2S;0a39qf zRR2MwFFcm#W8BT&c6h_2pC3_kOzeZV?B(CEMC{yWoceK$r1wUQ3*k`KdM zk^yu1VKC~l$yy(`B*xhqpTFbT57|^bhX9~3NaJB1bGO57q;;b_nY3k7*@oq!)L;yHYAp*zK7$bA*1i>Ck;dp9zCRKj_iXTN;$$L^nO z5?|CR@vq9;TTibQt3tOoTx?;by+!>htB3987aYw5f;j>^UYg8Yrhf_Y!Tdv;N!h?f2CnwG!9cEhVauEw|a}1x$>Qbu+q?Kurj>XGiXLG zc3Fe|p5qu+MWC1tYu+rZDANa(bmji)nR({Bspf~;7SRyM(i1S?TNiVcCjem zx-K^kUC}h|BUcexXd0{6uQ$oM?g&jTmc|^8tK{@Eiwm`KvIHTh-}n{iUhHr@(jZ4Kg*MkySv+iTyo*N93|e{y{S{K z(_S7>hRlwRF*E0fdsRqn?EgVhK)UQ?Y$X{4o&!X?t(;&n)-2wwQ?Tw`(9~T+;gmxm z#wPG@Bs6vR8>b7Vh)F&!2HEPE_sWWW0&OcdUh;tw8`lF7cWO@#?+&!=2FYT6977~d zx`sc4lVlPJ5VTpVGI(y_z7Y~pLjk;K!q;mGJ0DCVVrs#g9HEypV7;xjpB(_gYJ8_~ zv|f!#$bY;`r&g@w$vtUbNe$65~=O0*V2kMR-xM|$V*8J+VE7-{D2RC{cDm7)QKC?9tZd;zfcMf z3#>bQw-Xqwqf7!SFkGpW`94!-0heeygsEU8OZ&@_<`mzx_3jZ*53Xe!qy?W6VZuD+ zNJGP(KBy4;0(5Q}br3_}XCT36(I&46527iT0FU0=(%8%|l9Q6cS_1l?Qh+zeu)O9I0V$xL3wmA99 z5W57T?ihs?kG?{^$qp;T>y0-IYr>Vx1u$a~Z`}jUSP<9nncsfdCU?H$pDKA~` zXd>UbtG3mxbnb8dkYB%Nt_~W3IGUbR-~FPHHv$&|1J~|sB2e3{$&ZDzR}@x0)dsr^ z<*3daw)^!Yt|p7rw?xQ64-jI%Q(~D(NL_}q#8pj;DK`)+=740mDgE%e*)#deuI{d* zF6qcF|M2#?ij^-P^CQkJ8`#o3>e=Qh;&;;#8KA00`gMlkxWI^}K=1Gl#bg5k37sI%-;%pI~!ub4q{6`Mw_pYl&KCax%3e(G=5>*?(98 z5V=$dP=^%i%+YYZKZ802e1R&P^V`ZOLI2qv1L<(KjT=&q$RD8;12xc=o80+{BBQ&o zK(7U(xE%3wZss=>^=F`S;-UAZ@AE2)msZtrr zW!9Y}V-lU%N283yR12DVL$XP?oZSCi=Eznic{i?)qhKMuA{X$B{Om zeQ)b(zHYdU(l=NX`2zAz8WQK~xqiOvzj_udMKq1gr0o)SxnaCVi~_Tq1@<-XG^Yx) zr;FU)5@0=oKJs5rj%_y30fsKi9gtzw|9M}YuyIZumbJ&7Y-t}>u7SuF+%jf1r0jy2 za*QqkRZ0w)|5FDj|1DsDy;9)dg&bLZpV%$EQjdR09GzI23J~S?GRy8z7X|bNV(Ih0hxVJMVQdHdI(67s=;O0mrNge>xOn;1APO z?1!7|T&5~>$g6WUW(e`}38n7b=#QLLFI$p@l_*MiN2>VkhZkKIkvXIyt4KmLkwjlD zC`;QtEahpKd1i+@TJJ{(ARxm%vU3z~VQ2oMZBP5`LAtJ8ycw$T(+{1RmCAw3Ab9Fo zDbbrj?u>y{fY!L$-j>z__QKCO`~eoq+k&uuaG5{q&TfStvk`b1>*l}oE#1KHRSwyk zL%{B|aiyf`q}&Md$Dwa19+WY^QZjR6Il!uJXi+uPTVuZkV;;q<7D{b0ppD+&wXhJ^ zBeAze8fbULV=yWJAvV!!9R1#ul0%>idL58snmcsnkW#c-5br>}N+l#p5I;#1C5lTs z1ej~@_e2Yfrgq`pKQ#q69p;+vE;pjmxkKy`X8eyWmj^%yLwQ6103<~rMoU$k0tN1gI=X_=G?UWzc8Yi>$F?==iMcZo#0DALyorDux>ZW4R{ z`E)md{{aLFuIuPtZBE=pj0SOftXt zBx%1R$w4+im<;zj2%?RkQN71|;finQTRB`&TJldoU_*g&MhUI4C{w7!teg<6ml(Vb z@yKH#T|st5g^?40ETVZWj>)meNGi|3K`cWl2r^}VE*X0OOxwe*&5eL;w8JhjYJwfp z1(=Bn(6#dH;O0#A(}y9L=8#@TD!E++TDepf(uwxwzv$)#l9gU{4)pMQ6kD<=F=J+3 zsi_TS%W&0a88}%7=M)QZy8rC|I;9a};x-86-Jl{hdBsFjo5b@Al z91~_Bv-#l*^Tz@Aa5l?P948_3`k2K-89(PkH?^FUMrjl@pGrKyLnY_lp@rJeb^V?R zhfX&Vam4Yuc>}oUV9yO43L+P=+Mv&H*!+7Y_r(&_U|U!?Jw$OZo)nDlzriFWR8N!` zOcp9(Vx`7=WI8?R%fA!GgNmsuGaZ+nO^isij&LV4F`?;4ZXQHJ+{^b4K~&BdFW=?7 zgQkKq8^NZh*KL{w$d(6B+zLbtLtCOvykXL&LYgh#-Ad(QK<13y^TiTON+1x7&k3Z% ztj9g+QQ%sdk;)r6TIxeu73-pwD|6#{KW$r9-!+trW$kbGnI$uU#{^wQA~zHattw#R`Icz>)*1 z7nE?M?hmmxTmA0?J_*WcKp{#XD$hO-vS#h*YmLnDLjWbOi>+`hE8IeD*APOLLEJ$@?>M zQzGCYPEyvgRDl9hHCB2H3$4-824qgS+!Hbf0%oj4Q;54rz}cN9f;N zqcx$9RDkJ5R7|@r?_p|#9S@=gmOJP}5vLje8LRs8M=Nq~J!720X8H0Y87blLc-8W1 ziYM^{^mQ0IbGB(^))&VqwebdOc1ubRg);>?xR+xC)V~?jQL`dJ?@aAl6rdxDkF|3S z3RA$7iKU-QK|5ffp1fT)Zf@B#hNh90Ho+x z21QSvteccK@l*$I-+>d{>3lox8GzL%>Y0)sGe_?01DV9j#dcda>aI`jl#MV$ggJLC z1zXRnsJMhf)gI+?Et~1+s7)2TEZM00#VGOeJ+GUUZqEab;R4}J3B#q8`pnHH@{_tr zbd^FCWb8y`!SPFh+C)Ejhr4BY*KNy(^-*ej7Hvuwa0r-^H(qm3f2T7K&tDs8_cJ#! zBZDiy8(yfTMp08g_Qq)Rg3i*(v*G~xVe@8e?`C%uCCRi6UyEO|^l5w(UbdSWGtAnF zac!&)5r3F27i^L@JRUQU;kPHdHG_GF@J=;Y(aKuOmfhv;jdl zluICx`!(yX75!!kItn}>4Gd}B{Pga5dGPf$PIi#MQnk!+-Xd>L%MZ%l{*1Kzz1MoN z0K+;hE1{Bq_y>}sL!{I3NLwDHEj@S+Cxx2q?nH6xWBAwxM3JHGg)3^|=v|qe^P*hv zyK#{gcwzc3(9R)u0R7e({YguUdR3aJ=Fi;ogO#ns6!`DcvHRZktC$AW2zT6+_XIOI zXgB6Hkk@3|A}z5&*FNQTBC#s?dC9A72GoZ-CTmjVrLSR0)z}v~X&^SefHSW*R z+u!Uaa!Q@tQ5w#K`EUKq^CxVFP$+Z=M*&{ZEW-k^?RYvMB^tD~apd9sQS)Vjoq0xT zx7oZ$>7k}3-5altI@HM2MMeJnk+0+znc?HL>J~r$DQZPfU+`o)M_Xq4*g`20ykQyI z%2Su$PXh0^EydJw$e!bGCtlVFn@6;2ZHU7sTmlY3T~Hm{w(%)uh*^@4ieWNm#X`uk zG^FwpqjDN`TxIiuhg%`l>Hl6tf17k)g$LSvEtlzCJKZW$XF!3toM->Br`Hx(&yB+p zeqodZuS5}wd1hk~dsiytaFQxF282d-?8Kj(PO_xW|NPavF(9@Y49;CuG-pDaTl#5= z=tCn|G7NR-zTbcIX7GPd{dkitDH`gW{^tZX+(A-;wAPSE|Z&O;xf?qSLDX%ZT-|{PlMGQsD_ia!Y7zT*G_O zI3QmlxVZ_YiFus<*Fd!YwD!&XLr>&a1PggJAgR!vh}G}mRPid4E-d|NM?pI!L}c#r zBW!6XiqBSLiyt$D_?H?Jf%5bR%(Nlqrvbc9QXbXqFA z?L`CU(rL!iYh5xBEJ}R5-0$At#y8m6q;J7+86lOY03O*`O(kyG9ZIJ!d(^KA8ryzk zX>+MX>IZccX(DyVWA;Rh>0_018G{iX`iEWoZ($;0zH1O?y#&e34RF5o?cIubSQnJ+V0^YGwhp z@V+mv@c{tfh81)F0TS4o6ELXt{V-X`Uax1z^u{Qc|LVVEg-7mU+oF@JZfZ)MUe_$j z2*r)Ew4m@aIk2MnygGiyd*7fN7!=_Z9mT|b3(it!!-g`!ip2WwL)6iF1-csm6O#Z! z5%>YL(zgvQh+UyaJw_5YecuZ|4ypkZpWEMEcO#IB5C>5nkfMvgF*mFajUGo3Tfk#z zOH<+dW%U1%NiFnxE+IaGql6sW2WHw-vJg>$4sJx)O*U%11(SV63DJLM0>;j)0Fo`j z*h36l&ciuZ;ybzYK~C%3dOo2tN2v{~bK@^v(=y$_me#oYnIi2CeJ{u3>8^+09IAr$ zQ;>R@)moDU9m6+e#!06z6kMu8vHYA)F}%C4=~ znaqap=?k+^(NKQ{_5Q}?KB30Bsf56|WI}$&;*_DnY-y+x@I%r(15!S!V2iwRSqlh= zvYA-_gU7^wE*K50J`ls1W9nWgPWR*eHh0IGf;8H! z{j5nGE?!}9m|-QqRycNrK&O=4ie#RT#$6GfUa-MMd~m^Fl{q|gyO?OeBL_8_#*TY| z!ctb)2fb|PMZ-*a{T%WA9+DGdB$&hbl4VTKZ&Gpvk9fT z6(3h-*r9{;6+?*OfIx~Zdkl|%dH^0mTrfD4i_(|o`^Q3@Q#m5BceQe#3a$c+jK*ad zX#g4W^2}u-Y;rJ3zGCJPLp1C@k|+G`YX?_#WkLr)Z&m-}5mv)%yYk`)<+%*meb!zF zm6aDGLpqu&@F7+sT;p?E17*CoXUDbXJHL`SNkm)2(VvPl~FS&b*LfV%rG_|u24>*Mc=P%|fb?7$83A%;PUJCjZlW$YB2SuJNx z-Ica@ns<_v{4$atzOw~`fb>>wt{<&^MFGJaQ3io7!N3=OSF-Vp$#jo}lUw#K%=VOo zS}A9q)MM|Sk+7;Sj4!fZl3*7|JO*=q{*U{550-$5qQV*%wPL>byYCr%PP)B?j{rWD zs;hFaXiENavo_2UBs+1rz|@68Gc1Q%eS6EUdU8QpZz@$~XV^pwv>o=E<&?3y=uF9% zN+kiLKbF8CYHnYHtZ3Y1!m-rLuM|uQ?R3RIuS9m~e!S&{AlsX#Lg3iMAT%ZHe5>#h zT+^C+3vw0`_>SGoo|@io5A*qfFic?KGRAH$q)&@cQu;x8Nz+mA_?c~nI7e(DpJJC! zT-0r=-m(+EvqgDmb~jz-`j%xlyiUz|z#EQvjmq3?;P7PdWUjgsA0As|g=gYNeXN2k z6}g~ah&5!EK@r{bkp4wHtK;8~6bqgL6`Y1fY_E+WU)SC;V3s&8W9N{IEjr)eWSoVa z6rgE2nb)tw^uf2Lhu;GeBZd^>lc{-pry@Jrys%e}B%I9%ot*R^iZ z3p?xY&40>SXF4~z0BY;glhW|7O2X{6GVGW_0KS#?hM#_|raauA>_2a1tL4S(OaZnd znY)^$J6%+ybnkPxw)nKNB_S6tHUbXZiftf)YP1(;F5$I9D3RzG(TK9Ju9n>n?nah;k?`YYervk* ztr}Icr!d*UfEH_jB;cj0mIUo!TEBh?)?WT_7;SYIfE^|AEhM1F&ml~Fj@L%4G`mJJ zm?$3tkfzq!+7@KAto{CQ#AV1VClPt4*aQ+$xp*#0H1vCdNo^fFU?xa}c@ThNQp4Vo z7_Zj~biRbZREjp6U>sgIlo57NtPilvsQv7l%D2fPbS zyx?mgqGRvl73k;MUSOD#qvN`HDY=td<=Toh1DV*tt%{6ioMYgr z+9Cac;1{waRsL5qXsu`;QXWghsU5l*=C_J2WH)wNNc`puU(bIWTSITA zFeWyupHTQi{X{i*l#Fbc5f}~cL|bcMm*hen+r6BPvmzq+t3Q6E+amM+8uMH|7grfD zL~9$1Kn{!^$OxdAS<&}|^yznt25<74x+=sgodR(<`1p?Qsu+u);3zpVn(J9x2Dl53 zw|!%*Hsm`bexE>bX==AFCsJ67HZQJ&Y%C5?@l#U|TV^>7y~pHjfbjU#4L9XGSKo1O zYqzgF53tQy`D_Tx)sC@~^M~r7d$xlRypL0+3ZWCvx|+(ar~WxkyZBp5VN9Bb`0J<{ z6jWcNF-e%dxuO`Ks=Vcaccu)OK0`#x*PGLFqY1o{DiV_lQZGfl*B@kX_+K8ox=jl3 z4dx~tSe>J5#|ww`S5})K*Me6H7dkegVV#ns-WIrt6boo;nYl`4t?{A@{k@L8{-|*{ zFU%X45#AIa(-jo-<8SIhxGS{<=(I*6BsqgOZAS`NS$aG%SH5Xl2^QkHF{kZfph&i6 zWov=E$}e{Yj}~7Fhw~v844pd-Uqs1znU~b-Y9f!eQ#iVa>`#}1<*~Wph$Jq6@=F|{ zW6HW_gzv9`Q$aP}QyuSrE+2!98&|xciwK#}Om|}+WZm3~z{mb{6+z%X%+&_XY zTV38(tK~BjL$P5YZSFknn9vj^-_lkqMM;X-+@?41o>d-{06aY^>e7)!{Vk7k;{4^1 z;2A!zNUb!?$B)b600-@l)m8IVG|rKG;rSPrgg*f`TlsT=eS)Ljn)xd0XxIgKvDCl> z$|*bLlrE$;6|t|=ijQ^}B6K5h5g)Spcc&VG8;Lc%0W-_)=WJDIuMe4Ym<|9kR7Hn2 z(@RB>&yx77x2O`)kaA`b0F$@z?OF&95L6^va^vcojRY15bTgIM91`0UEJdEt{Bvt< zZ-zm+ZTk^)ur^m4eB$yR;IL$a#F=ej1w;INxhq`O zK7UVMo?T-$dQH|AmhrUKqMFzrY1#IU+_j&fPv`6fLk4)^Kn2=xo-}H()hn-bSK@8Z z?rwx*uCg)PGgD}ovZ-)|pM!<}aqNxIzIm>5DjNxh!n7MF&t@L@qH8%EYH-%HvPh^x zQSxx6_6Y2zKQU84D8S@nFWk}01t`v=dMd88VrM<$Bq4d=m;d=VDf=tA`Go3Fo+ehUqS1p&7%(t9^jRY^EDRJ zzv~}E@M-JUQldtMOXnG|H)yThTs5BBHF?mYro6!}@gjlDDmR`loOioC07wwI1gY(S zMqv%j@WCQX=!Mkb#5u)q1!&w{*T7^P0VHv`RM)aE-jHVQyEK*?SKe9-?Mb#9z?$H; zI@?Y2ud*$zi)Pn+k^(*kT1V*>`uNmGSc#4K|Ez?fNfTZEV4 zki9c5X1x&pSQtVSp`Zc-c8Boh_KBKapaL1b(>Hcxtn5z@ID^GAIoF74%WLHU;O&Pz zHL+halCKWAmoZ2G=Y#NRPsBVf_&XNpqT8;O10$#{p1P>a(>zba^k!(~czewlJ>!)% z8&fu9fK37`Q;DSuM>-7kQDmMxFinw8=?hO(PZUtHzUx{;5aO7@-jtIvF%AUw$wA#L zyc!7qTZ^^y46{&Y=HC8BU|NPwR$F?a)$Ws;6cSbRD)^@#ZstgzL4zz5i1Ba%`R1uZ3{N1d!>p?5M? zIf%icU;DoDJ=czSQLJ(3r@<6{?uE{9$$E9`x9dt@0b5x;Fns$H&-qcK=%oHA6D>x{ z+bG3)Ia9XOyJhU4KAHNj%Eflv(CaCY7mDN-4rBVmxO>MZTqPyxy*~s+<&t`9p!jw< zOn3PP)eeCJpz$6^WXw;bPj6zn$0yKzHS=IpbuuRCcJ27lb^&?}I7nm#W?atFJ5f7% zfZLlV-DdM}d#PNf{B?5-OSBs|RfrmFGt)M`wIWNtp9=Vx2qO3^DUAv$N$Ho7YR?=# zhVU!tc_<>{QieccMHm6y);bmC-uBSH6{Wqol7^822HF_RzBzn^JR&ZG{7{YZjS_C-Vkm+-(hr5q zf><_d6jGO$JT0^@rMWSoJFW!_-1}!VKNEL3Y|Q78cY$i6h=`r{9w{PnrM!9)<9=N) zF=OQj_8yqb9q;35C==?tqr=Ivn_W)b%$2lc?#%D|XR4@@5GOVKS3KG!o~l)#4r^K< zo&y^RQP5_*)L6=tcQkOeYL4(;eJ|kmd+`L%VL2oSDTi5?5s-LD9Un#(wG(^5`N%Tr z2Ccn?bHVHl?JfarX15zy(wr#UfWe1?~IXf>zm|DKTn>KKtY$CrAs<^&Zv3uy0 za`gp?_hksZ-F0!dR%yioUF|^U;TVCVUao8khQ4hmnm^4qxg1-qI_RdOp$r(=i6RYL zRUHNMrrcmpij^e%MjrPACn7+@pjFlKl%@O-kiyDX1qKa{yI<@UYTXugrhy9bS$E*U zWf_4%a53%|)x{~a>JV2ltB(jrulj@bA5xvPye0>HQ75dK^UL;eO6-p-G4UR|+h~rU zWiD6YvzqKY1)A7~J>dO<&-2uD8$=O|r?1*45-&t3B%Edr?M#DQ+(*}u!gPKumk;?^`#KcmIK9S zSG>U)to!CH^pdr*>N;V0RHlbZE(ymG6V2Cym)oDEyRDtBazs_@1zB#pF5yqZ>Z@Fa z26(>h`>-t~K%w-}*RCgwhM*EC^ia55iNf9J_QgE*>xnj`@B*K^=bJu{E0)>~Dk9cZ znAR}gB2^fja@!Eab5$>1{^4*m-#d85cVTP+dOwRC+=U#25{Q(XI~V#C893!;MSETl zHXXjSRLby82sxkKL_?GwEq>oF{Ay?85VbCHc`iRKB3uQV}#olL{ZZH6n zunD}KJ-Y0DvS>&R!+AsQ$@4_IXa*~uN=}VPPhmT2-e{Scl0d<1oekL4yv!r9ro~a* z$WgD)y=Eq#>U6^?N>u040yqQp{&pd60ZTZl*&!Z$m2%=1>y4vj(F3eR3j(QJD9&AYJlHIOTJa=I)q*M5>&N4E zS74HU&;uk3e_9<2{8w&|tFvDUpjXmL0-jY0vd`eFz*@(tg+XFbYJY6^>}8CbL^^L- zp4JECmae_vW!Z39;M^FgB*^d!_?rJv6MTb$4Uj`onBuC`G{@87 zZV;u&bz#V4og{C8wzsmnUoVGekyMkE76%1CZtGf>+S?}~qbns`96Dmkx((Eh)zfD0 zuql}es#x^Rk)Wt1L`X|PxE;cs_adnn-K~h#iv42FWWZe)!y4euiGPAg9;Si3n`G2= zr}-hRq!_j`6u65SFoXXrZr`~2N^oWOv`k!x#wdeGrhoPuzQ1QdK^DwwdhgBIH4pK| z-5bRio0-O(MLHvN9ssm0n3cT~F=nuI2T8b`(90$jVcmz<9-2&v)f*O;vB-e|gb=Qh z8bx>kgK4ft)LuB?-Bzn*RF>*N7_w&RV`r{PIU*sO;&%l-;`pL{qYQCn*YPq7`?H)> zO{F7sJ{NLUay36{u(P> z+)_R)7_KccS%)%7-HsNpNTd!t7SV5_fkI8>(2iYt49u%PT|Run4vyZQ>%j+^M-( z1e!_NjsV)T_(h#vzyV1J3~V6+5!9;vID2m%Eu}308o)<`La*o>n4=JN*7jRzn2~bf zdWkiWYQIpOh+t4bP0?~-z;M=O66l<<@Dc%sqryNdtS}aFTGcH=9H5wA2o1q{_a|zm zldE|yXz+jU&I7*N5{HZIscYZm-8sl2Yq!aiWOXv`YZh0zr7FhvlsZk}O)7k}3MnX` z<}2$n0qqHDxQJdSy5Y|q7?Jc4umwLUe|d^}(I7&=$;c^v;r;dL2d>T{RPnTc=9zyo zNhr9Sz(>)O>H^gtJn>4|=eQSWO7vma$C=a^i5ocS9X5f$21c@|9`*^Dp1NAo8{NpB z_J0Wnk&$u6C+9V|UE%4Os#zWP#@a_>>^%!YUK9W=VhRaR zM98Lm%iHWeb$H<}ND9XA9XiOMxjpMqj}jLg#9|67FtnG46MVaCY)EvyDU!%LBe*dtB`g!$bvqt>Lbk9^CyfyOuD3whdaBI< z0SN6N(0DZNDX9L{zI3i^W1~MbX}@f9395!wXji5={ruNYdk{onDD=J#oVOL&(>~uU z-Eg{O9=c*CQsZGjq~E%~DXX!N+^Ty^9gVM;m}k+1?z4|{IPF%eZ+uvZn>Gai5ti`~ zkxH{F6ERDD=1|bLs}X8(a-j}bgaXK+%Mng2VXCwIo(VqqLv6j&ZvPhmL_oX0(gXgf z{iTKtg0gdS>6!SFcu9R7vtNacZqyc;xgi*X^pz^0)| zx71|v3s=65Jm2uom1o)r;02-NkGcG)c;WW1D%OlCA&cw+$eIQn(X2!tJ;bQ~$^wIz zHR?`1i<{**PRY#Ck*Tq!qtp=EqXEN0(ZuyxdH7GB$KFnT) z45^VSb;Jh(GDG7m`4>hlrlm{ISR4N-cFIX_G6tAH_HNuoprec~#IaJ7$FnZSg3FQz zdG69`bXG3Si`1s_187X%d(R$64kO{>#N~d34rAhgjzq2~rMx{-DPrfBCS7WPfXtJsT<7a{DkE{8}IkIQ%;^72t*>@=S zsLo%TwTEeCG6;?4Kb>D(V=q%&XX5~irVRmD&p)exG;O1&gfQbAY>2HnmQ@`PIf2hr zj{qME?f0`=KjU3!G7ZK`tc>pVQyZ<`QoQTlAK0N)$#pC?KF44e5M4s)em(8C^cpBM z*Z4)M!rdd2AZ7Otc>OPU*_>8oV2-(v;fE)dlH#8ZEWy*Gm|UMZYQ12y6LeH2a_AYN zBpR-23Nyqyta4-gcI6~qf`D8m9|ExLQu$vZZ8JK!hajX(oMvr(#49PP&c7Rl}S$z5Bm!%b*T)<_4awwr=O6cxbB*;&`O^2D(6E?DXqa8=Hpq@=6sCl_!1DbwE zvNcOJK=Sjj+SdHQu{hr{pjhbO*qGR8q5<(PIV%l5?xdDzSjTK{3dWC{^6IHik(tvYZ}aFScl=AL}swFzbm6v-?x(6$TmEd%Y{&hYDd|5 z>xkeThnj>NF}3(_?U_KsaypZyY5R&#f$T)ekpYAxQ`%>Jz0xoN2i)+*Qzo(lWst9x z^p|Bui(=lo>*+e>2k26*OZXA>zRs}5uNxKAGiUB@`K|U2z43V@p1=W$yJg3M? zS%|vIiZ~|OL&u&e8}XFa_|$@*mPu@zg|JL~Mpi~WMTSdmYArD59^`{XO34upul8y+ zYL77HCMTV&oR$>)naE)Uf)`du`sFOffN30wyYwgQ+o{VR@TwEkl3)S7tfzMpa0 zJti`n({PLw(qm+WDyQDU*p6K|om^}|;W;^2K~ff%op!t9OX*2$OdjWo&-b_>m{w;f z4dI#42&=tFjpR!GvcOQBR|^)@W7rHL%_MjuT+*RUlfP{++k=GK4JTs&u0DEVCq;5* zyBmUGCJH55Bk8EZy}7+=63iAV`Lv1;=scjstRcIVB=ToVeH7z^kvNZeRy$)3Q7XX|cik(K}u@3N@Tzb@<-P>;&CnF4htM1TUf1uLmXnGV99N`G!QU!CsH zm3NXBsa5Q0&=g^dgm*WUSfseI9)Jj9;e;#-<&C3@Pa4lXXj)r3;}@g!f6!dtA_o&_ zd#ygcj)a8r{d~(!4&N3+aHPqgy&R{($7!_79;e2_!}zHLCwqiHp+fRR_?TpqABM2x z{BdH;Wck!t@-7o!LQT;zj}rV^$fg!kP2Y!jc+tX?`Mm9GB}LQIicFHwxz0d;0Yju6 zPR+U>%j9l}QXOa?^$TxR19Y2Uqc}d0aSoV7w_8*nt%hM$2s$&<_+~OPSmjFdNBK8E zt`9YSS(sL5;VT=;S8(2H&?&Q>hiw?hoVX&zmgIoX5`9w-8gGB)gS>8ncAjA^iAMeI z#dmd;vD>7xI|Ed0}eaCbu-*9#rvb+XC? zsP;7?`;cANfvK=tETBq{V^R=yZ#|(BGrK=s3KV}j$|8yO-D+vXjg7-; zO0QgQo9h#eADw=QW?Dh={=U6LV2~L2*$*4B**%wKya~SMfesP9laBWH-d-*34|S*4 z+@*_LMIQY+tl(y4^|P-BEHlh8fV7qpaSG%8jynI$>6?^O4_KP_C$J zC(8)a+A&SyF3ql z3287%T<*Gs7)|>JIhP?|jTUkb791L%hbyx)Tms6H_p-_?mSDK#4pe8kHqJZX4pjJx z^VjnpsNqt0tE#@Z3L^9zvNw!KGhU|$<3lRw)F71AZW3PD(LDPt5p5}HL+o!l-zcdF zvg{%g6*b4IsZ{VOi|{(2+g2i6Z|+IzIGPf~Car1B^TT;pb;TlOwhSoft1c+gMTM_R zGR(0HNAklxA2)a2c68=OZxiv+MyNQc+4D@71s-99Km1ynegK}|uDjn_~5J~V%C zuDW@VROp&5090PwY<$XA907;oJm?D@MzzqAqLT9>(J&sSDI&Q#mwrBb&PS2?7oM5c zZIJt{R?DrdUv&|KOPn8P4XNAOuP(919(m#56`3rb|H^Nm_5;_64Setj7|)!*#mS&< zzyBJi{)F;(pX^3+)%>Q}g@AfRi_zWfr-}m0s&-_|`C_G(pr$Ry8%oN_r9T-4r*MX( z*-&sX^IybAU{PqZav6mhZJV?kM+)vU2KdJyMZNu^v_c;ioQOtlU&AyjXe3CR{j5!T zibxRGg)u=!rIWQ5IHb9kj-xi|yzAyMbTFL&eVRw>i(qg~z|jbOSa0>M5Z@O9-a4yJ`ZZ9?ZyU9TGdp7mrbl=VM`nQcG^DNNL$xw11ZhfielKX0b|)ZdV1NK z5IB@Ud?UDIu`NUa)PE?nR)5u}chX!~R31SSyfsNr`zko-U{EJISE=agWclb*RfE?; zU2oEg)!P_6t(nfnw;5FCn-rYkh6ol8#ky#i;3RM{VT^LF{xl|{d%yJg!oN~jT|LpU zvJlq{#=v{Ns(%r&u6pKQaYQ52BlEIly^o+ON%8(<4~1%fcAV&{MR8F#sRJfx(%ggl zU^K7^!eo+Anf&5)2@lr|5gt2h$qv+M$@^)8+G4$*M)rxw0rx_vFrTaE#csHtcN6Wua zh&-b~zdld-C7nX(6?WroQZr+|Y)-_1jit#;pVyz`9hPwpaSEfs?62-=kmP=PlxK-mhYZC-D)!^CwrXA8s z9JpGVul2s^x5+A7h;#N1PS+gQk93p~Pf2BjI3#KC7&@6Uh6pb~3Zwv`|Ik;vx&n4V z6pY`vX84r3R7x{m(6yt^JH{(^X&>J zCE6t%bA{XUT+L0~$g%8Q%57?(ytVL6%!2g4hcUC)#1F%p<{e`fzfMlzKFPs*BX%?V zS%7g)zKm88C^@dVol3mO&zaMz#_mQq0{#q!Xao|+=?xq+5F{eKQ6Nk}ap#*JS18a@ zSbTGbnW}=t%)YGIMfc|We38;(B4Aio%1Yka8nbGFS?xgSTywmeD{*wMzoG_{jmhT%W)aeUPZjqKwTo4lL?Sap}S@%*Ptvk=gW0XK(_vM_|IMsN`5Q7oL;zW{6waJ zSlY-zIos#(@Wyse0nibaw+ts)!iWJXjG@}E+Fq!K5Gqt0sb&)!;6%w2sCj}A9W&%@ z@RoKeakARa@>3+6)~CsbHuFtTnq7wmXC}e!*MGiS!vp%ULeO0w->`?|x!wx^HtW+~ z*mrKB*~@J!S%Ck0Hn~F{>`jqELe{Nd8)0+Wvfb_l5bB3Y8x^GF*+g+}c-kB8K`@CP zH^F6@F1#Shh&s|$w*9!x!++#xw|h{)C+FSd5=@0HT!%TN$c>);jehCC{L&%_AJyb7dOJnl~x653jqUjI<>P~Uh7IvHzFw@~4He_nWwUWRs(dpcuF|S)1(w-Z@=)(DW z9iQ;IYY+@9Z1Zlmk?0}3;xIk6VRPIr=5&hFQS^<6bTR%;`>qG`0p=4nQBZ}tAcY59 z<8#(8+kP{~chrPdWLkv$a#yp*ehlHXS)mp!d;CXUIGmYFsgl>m=V00??2?e(R z4vlgVCErL(w#f^2cyQi%bU~RnxwSrD;8H-rnj~U_A{F49YYP!|TWKcVMUPz~(Zp>N zPO6acuX&V|ihudQA3){&-}|-L_pZ1|`o)ql2vR{USH- z86pznBixIljpFNv4&-NbOS37<8Fit|5ew_>?X8NnpRFeAWVP*Qvc?&TMF>YCMR znU>a?7MjYjk|$6aUPrOJKO2%@WH-eH!yN3fBUU2GTJk7Nq4%tN1bmB#ODoDe%AMk7 zH2WaI6=@5FHEw`(+m9M35}LC=jw>{ibZK-A_zK@5Ew!FJ7rt>DFL|QHN6jm3E>cXq z2wzzfwtkXEm>EIrL6hW3sVL;_N5}3MXBk}m-JX{4M_?5DOxY+F6w{4r#B>2E^I}pN zBho*4SEd1m=VQFp9Rn6<2%*W(%x?4-CNSjad;1pVd`KKnq}F;1aJsiDyI37ltKo2X z>^(3*DjSgHaj*mKhlLRCDKr?Fuj$}7xLon&2aWyV2*-A=8L5e;fV8t>Z1aui<4*Wd zvjJ=#fQr`t8`EsEerJD&A0#&ObS?$beL~ROKkl)A;hfDotiLS6;RL)`wf1@rT+*|&fX@iuy8sd z9;=cpF+9Xa_Ma36UxEw~FT*iIrC2jK2g*Va3D~_m9Yo+%iBb>C{{g8cE|go#lAE3P zHZ*t!e8Yn7OwUht_n&8?Ntd+=2^XetW;7!p!Mkzbm@-pvhm=N1K=kLfNrUr$=cn;D zwp@muh~3M|l>ape@Xu?06=OL*dCpJS%g-xz;4W%sHin2wYJN4f|DisD;}e%$`LY{g z6}5b|9c#xJ(x#DjbGQR-a!bUv95jc2%R{lBNiKOHw#pgMRAGL|vQLKsnCaOF`n zF->#OI@RPPtX6>F6W7(@EWJubFfYM&`u@ajWo9}eBw)zIAUPzMG1TL|z8gE_S#Ki3 z0${AK+F3t@I*S3_-a9QWqYLrNl7F(P>2nZCLsQ`@vsJRMcMwH8sgegj0pIQl`083s zY^b&6VGlC$1Ri%P;jylb}W%FF4OV{3bGzNckJEyl-=!l z7Fa$P^<6i}(p*ax$Lk`CgfMFx#9mc#4RK<6kQT zv8D>=B{Pb{|Eubn%3U1|<(B`=Nd0x7;vY_YV4C`H;3^#zNVjQZgoL>YOCOV!gM z#B4IMkE8b6jdGT=&)N{|iG;WtPlp`F)aRh|l5p9e@tjEaB4*I>;Z}gU~T>>aW_dM1uPkILs|Z0uhxxInpPoZ)QUCt?i)l<_FmJcR%s;? z>Q!07+_~yyM)02lpI$1NpN=&A0`c0|D!)}ekf`VR5X`xYX9Qvm$*M$gcEx*tP}Beb zdYNwq8u?>W>EuJKYn>pRi943IqH)AxQ+y7sk-w>DMkfLnuA4|+%RFywf_3^ zel~|s1~M)}w@URFd#Q9aY?4F>J*WVnX*!Q@F&d;IE%l4Es6}zgd5~Z#G;%KN-c9!Y z{$h4Hmo(dzYBnCcUy3Mpjkn=Shy~~-Z)# zR)+$+vG)P%IXvPI+>61%Je^%0>JEg5NW4HS*?GOUh_DFo5~g{tdGxHIe4q^^&+zc? z3$@~5{%!})s;$ITYpA8dIfw_>a9Fx`3al#a+W!x0sI6dIb{R5Be4UE~y!8o@V(O+N zIQBU>3ah<)V;14!u8~8+bl3$E9dOd^n2eg!$-O@;oLao4!eJ8I zue#jxaN)aq!5hzo{h!Q8Bn4v_0ARA5hJ|nd000000000000u0M0czw}_=G&AoM zkeL&z*S;4yg(Y}%WAGOH39Fk~uc{4@AU4vS99*L8n5R2#{#l~{o@v}xN|S$;rVjH+ zs;f*aMpGi3GJY?CgWo5t>|gYik7E;S2C1|<{5-+%UZ`t-4c zs)zN|yl2^}g?<(pd#RisqM)6@wuD4u&qXCNwK@S9jRn$(CwB<>2{HOZ%;HD0ElHrP zbegBDoH5u!-akAgJZsmL0uh{9)eujTPnc|$SmcBTx9@AzTx{!!TjmtmSgg_!sV*bd z6CjQ(f(0`w=_}M1W{Xik6r(LP8%u&x5-*f=N#c*|aUdgOwLVGMCjmQ%#TG}EHAsDe zMquoi%&TQXQ?&Mm@)#~Xca+$0(v~5}qSQgV@-px_phnROxL$7<(ITlJ*RY1K=wb{Z z6aW*koS46xC8k`acq*NOAWeYQ|Nk9s_=_>E$tz;NiBYXN(wYUji+I)B#`boGL4jL_ zR+mq^P#s`$k5J|o83->f{0Oj&GU!`-l+iUcel6$>M18`d;v{q7&XqYW`UVDc?a~6Z zWv4TIR;X?IyC1JUCv;8>vzQJ|$%e%us3`K4#yyCy8CxXIH$Z}v6b_Pbp)q!z!>XO? zHO`s{Qh_`RvaL-b3FK)T$>|ZU_Ym|J$x5a7i8SIecd|nX1aIHPfs9=Kb1o2RBb`U} zkh@|%%KUv}Dee_XQac*7 zsCuSKmRO>Vh=engP}&^pU4>54LtDuXuuq2Th$4|Tv&Dw_l%Ugg&3WA4WYqKERk_Of zLj;s!1S=H<64vX&B5aJKeeT2SV&z)bc@``8ql&G7NuJVw715Q;5AM2Lf+Uo5@ohoV zS42stvsF9A8QNI)ZoBHeQiE$Oov_gv#D zCq6;Y#dc`^O6T%*(Oznw1TlTkFjt)1{q;m(r(s^-7n_D2)mm0%Sfu{d{w zfHxasZ6w@V##uzz_$@M4(am#>DmcU>a5P|7RQvKDoRNUR>zJj2aH8Z-4 z$r_j=X(dZ@P!?{r)QG_m-kYEs+8!NbECXov&e;fR4%u4R>8I&5u`C&Vl&>H_J#*zM zac$4)OrOD9)S4P2y-!#)v>F|45G2#G5U)C_oXVuU+9tQ@YeLh-1GoUZdIwt}96%Q; zNF&-id=iOCdeCs(^f$}A5jFWjrMC-hx;Mf|Ys(eG=L|!?7 z;E>ZYLM>2G!xcTO<@@ywWdB6&;z>IBsp?h606kzXhamRbgYLGu*ONrdxX<1d?g3?0|D=NwIZR@!VdG6t)Qx29=w zkpww+16AcbIQfs6%s+K02rJMCI$Aw&CyM3lWyU;$CKml;E5=GHAzSR7TVNRG_;2_s zZa7!!&__)yy2}VcQHJzRrily>eMoQPH2rj@d$ocoJ_buC0V`?T0nYmxSUW;7bs94E zF^Ws+b|ty4iJ&h7pMjtTz-jR$%?!cdA4r4oVASWo6Astx@e%H)yWFf;jFm$`8;*Zf$hk|TJ-1P~p;_Ku(@2ios57VyGwD?(>? zwaQEqm1SFl^vDA*0e<*i8LwlJ97Y`1ki}EOztNTB=(al~2X`CM58T1dA_K)zM=G7P zo)L?Vbeo@p8$MF|Co{b51D?R_j|#WR^ur~i*WF~B%ofG$|B#v=W-kJ~yROW&Dl}5s zlScHo?d5nn4#a%(!W2ZJYGr4v@Cu!SJ4KdG8<#OaHbL{S|AyIp?)2C(m^dZt&v zKOy#B^-xkJ2M3vI53!)3+Fd<5cgO{s<5`G<@BeO$a^;3b#$Wl2h9 z)7eT6abZE+UDYh``T<64Y)cv>eRE{o58b-|!C77(Z|?ht>pZh_xmtwKCDqY?u?0ph zLM$C+Z`Cc4Bd*rlYQkKXqhi(Vt#dM6AKW|8#r82H4WVg#%W zvI7H{#8&y_ywz|7XrH^GHFX(HSujgb`nxDKYCB zCngHo;c2;oF8NP*0SCYsdm?jLfuNXF1Wj8RjbLEs-^H1UQJ2qBpN zYz4>Kfhq2voDM}znw4!G`OJj!KNV1eC?7M#Mrl5xbTPPYYI?Efe%zkOexK!VHi`EB zU8hwWd76GCy9+=y)2Up|k9Dtkubon$Z^fa0^j4A@dF=Dcr2Z8AkMY5Zhqm0kC?Z|& zJr1doJAsz%EKef*JKy*FF8bzQ4s_-G%xwak0%F%O;R8jZU0T`~yI4|pd2_Mbf1E3d z)0Yk7EXIB@FT1McQ}L?1wEWduXc&V89wjt<+nsqo`#Hi}9>d!c#!e=7n~=_8j`Xpq zp)yVy5Z#L!7#X5RUNjKFX$)^3kr{Foya<#DoI$-%sVT&rvgdtSdDP9phuLh6D(>@H zhbSoUa0dFL&zA6nC6%x9G%iO<-4gzaj3&HEHf)$p=~NrmP4S3L#|*eyi%{7hTd*^9 zAF?<%IV>4XFjo*?bH0N?ddt{gZ=hz=muo5mo&r$Fqpdh%dFr~^I<`b1`wcpO7Y00ID^@C3;*O8zdDT@*_a z@>I$T!vCVBrD%Fb2XJ+R9ss!PM|yJ>ncoD@9i(ZA*2<%&EPsl&PC*oF&q-odsKzYn z6_VpE2dafk|c~@tJ<0+%*l{}yecGQv2qT0u=+f_*i{P7#Ltvu8o znUmaUC(wsA(#Cyr6%j+FSk-EvRej>P(xerl`Tg8RR$F6eIZvhQN{Xl`ZRz%az=fcB6?7`g`u?7MOC*?*Fq?3!+Ssb7r~H{0gV2*F(o3fyC8qq--$ z9NyRbPuEpLLJO?f*f{yb;W0v^Pyxwi_jA0Rf-YK3%B#=?ZeFB`>Xn;+{RVj| zx@NO}XtM5-hHQjY&iMZUuMQ?QIlzl#5zx`{>dPC<&Dc|_%6z825Z?(FWN8h>;?z^2 zRB0}(I#lf{SG#O_)Al5WLYYiLU%`2my@NBJuuDFMKj7axMBVFT+KZbw2DbJyxGjU3 zI8P0ufX>SIcx%z+ zv7NmMKjV;bbyv6$*ypO=>40kk5ODyg{Q8~Z6sZ(j)bKOB>&NJTb6r1tjd}C=;R2zw zbEpj!FCbI#ZNT`GEim2YPp{T@31+1WW`S?WL6zN!kBv@Hc>^*rH%bZNl@F&V;h0T;Omb(|(TBc@ zg56O709_h;xvjy-H~R?VBe1 zE!|oD&?IY3P-NeUMl#QM!_?ug!5m@Nt%MKztxDBmg)2JbQ}A4PVhWui_nXK0o)lDo z+{cRCT{9>A!GGzXLLV8)iu~7XTIslEkcQ$|V3ntdty%EQ5yAu*!W}~#N2-@Y!`ZCUrwEcty6|w;@s<4LdI4ZiY zq!$Z(6ll6-(NIX03Yln_pqUF7@ZsucZ8j67=7h}@PeiDB7s)>h*>*P_~l{z1ZU8n>SyK=S;rlxlSJ(wsWh zNc0dQHYwQPwAlnP42k$tWPdI9H4mLPrk_OI)wX*Rn{1#4Ms3XUpRWf!L~FETLO% zXbGp%E(UPf?b;9yc8gyi2)0lD@I@$nYW`P=C}Tsmw#CA*M$Cpf=Er0p#~ory@$&Qe z`H!sPyle7I*}=h>Ru3rRVUVquKBfE!DdDr* z8Y%IJ4WS_#qe;;>9&0UW&jVk%GxIte%kWS3jAGY;wug5q6f9V!ybZKIxrStkp1|;y zQ6Pd7J9`tz9zi>}+C8Mr2GBG?*GEZ(=E!S(AMQ-KUW z5CQ;tsf8x$zsXm-53Xj5WixT4>TI`R?FFGWPZTtt(AY!sf(B0M5}w|dn_N*y+}XGC ze>N0Fh504`812&xCZUUOAxez8bo%siuQ54+2lr(oU&<%OE5!hZgku?%ZIM9 z?VVs=KawRYZQoY3jq=oJThC_827zR{q7PPe>j{o&_ylocyB!U>%>dbkDaj1_sw7oO zbDztwBDhY`yM3bS=;{o|#SDigd7D6DYWLb9M&011!vkqZzarNBo|TyA*@dA^8#B@-V8y5Z0jCP>-wwOs(TWA`w+k!s0DlUWP{cXu}TB}d6e zvDZYCx0#P)R`lb-8VnyI#~E6H{#{M;I?*GSF&_BDCTU2&*LyXJf>HE`5yS#6*Dq;Y zp~eM7KdJs`u)I^LIECJ3n3>)rYnh9%$112q$M;!ODDH*Sk_*e({*L0tKO+R$y14GYeD!dBBiP zp*_N;m$+PmBOQDIDpItlc4pP_$|3=5@_9`d=`u_PSZPC*tnaf7r~NP=h`8dP$F=&} zm!93e(z&4;Hk&y_&bIrkpU_Un6vKikGl5m;AYVm}z?-$|9W_1eAU38d?`}NuMb==P z7l zpY-EXH6w*qvmEqfXQ$f9r5A9BSy_b~s_HSstet$3D!$~I`cabXXN)y{pnRo9 z5WV?@rn&^X0+gEDYN!HnIy4wccIk?P$Z^Oj>%^Pa1Mj$h;lM zi*GB|LbjNOAaZzbOUgXyScptFOJuAO;;C72FNSoMduvl!2WrI*3Lkcdy zYH|F;&JG&thv{QK2?NB{ep^3BhzSsgYxwSt;h$-YG77=A4me(3Xrci~zT{+BQHJ0S z)Z#Ab3JhUYjleM(=6sNQbMD*1Dzl)?9t0p4z=;dShyKqcPL*B#0)5|3KMB-g3*Z#x zoDtUBHIl7?Ubd2-x_b{>d-GFCilw41H>l9bSkWH)e6Qnn)D`8duQ(cJuL@Vlu;&<@IBhJ)(s#S&k0?RJ*J0f3w*5Ad-P z-}HdBqT%v#^C}2dn7S>l%=gkz-cr9l&iF>*#t_HOTx+<*tqu*;!{Xq#$MJcI`S zcrU(%3Cn#dSxMq&gJ{V?G>9G%k0yy%X@OcxoL4Qx9u~btMJRX29QP(_kRI` zg`~xD#auPGk(CcR8B_oikeI0j@q1t4_GJm3DusVN9){E9ZTyDwUJ)`m2%;D3q4zQR z=&APKRCXU4Tn=;H zP?V~SSv9D!Xmzz9m`}QM!u?S=QkSSfrLh+JagF0?x>pAq%RlVAE6{KMTVMXF392QA z`4`NRzHbWNEH)YH8iahnG6ahV2Y|XP6_HLC2v~`p;lkv&B;bmhZi>P;tB)E?)DbPV z%wsVUrAH+EwR%8BYLOnhrLTIQ9BqTHl2PglK)x)Ko3TKJhDFR7_cctk&>?7!wAd)Z zv!-`?k8$B3u8UptYwKu{&d<)HAYPu?c^M4C2K&M^^{eQP)F`leB zTXb|Dlb^Z70NN4)eVn|R9>u*PhUi5n%7cGf)Q~T01==^s*}d74I}>rAt5x}itX+M) zaV)5`UP6!n2E*(fsRfH^0-O{PuE7tw}Tu3qQ&)H2M-5t+JtP$yM<^Wo8^!=#R7moxL@x}LK7lR$y zXv1lwLkKb(bWL-11mcUPP6ToV%>!Bm_;<+N^Z7K@*D&bDKS^Cu>O(ae45R3~Xko7ZWA+j+%7S!3 z?^fskvlB`+5|KSrKPHAyI`k)HM-o%q@Y((X!G1 zuqdf|f&hqtynN2*PU=X))h+IEiV7N*!M7EN_PR`YZ;wa^_%?n4|H!$SdvDKw<-{#A z%wtgQBGB{)!PFMvH{<03QudByWZNnOQAiICWJeA3qDnEW00P|n)xdoS@-Q#}SElI| zg^d@EEXhD=+^y(&Yix+cl>KXBp(rd_^S9B@T2L8mY_i2FoHNWX z2F1<;g`Z-(eXNG|gr(3qMMo!^y1C_4jw|e=yT6A0p4`#YgyjgyK5#B0_ zk+eh}(2Dd$pcukk_M^8iJQinZLcw8)yi>6pANt_`BJ{yH!+{22i}p34AcK!EWsn~z zopY>12C=JigprpcEw)5HDoZC&+NqY3R^{p*B{@Uqmx{|i8&rI^3x8Q}q3-hrNuPHzk_lg* zj4vfT!Ud-NNLs@?VzeP(nYvyhCG$vJR9g$z)G76Su-d%|UtIdUbyp4vlH_x+sqgx= zx@wRa_&(*2Xzj)ZupSLJg<`V)lpmqFp(~d2N}hT}QwiRHv+WZ_7U))V{N=sJ&=#+d zJ~wU*V;^VI5dM~Q?4t#PcETOywVbyD*+MRQr%^fZL&D0x%dOox$IJJ}@u+#J z>j{thFWj0vtVm-yFrA3FLm+L6Eetl}nA5_;Eq?XTv7SWW_O#o$VfwB^XI$ zSVyW)OwJFWnmRWmXJ3Ptb{~Of8c-G*h+t8f-=|3tV%|IjqbR4+nm5nwlY16o07-B)T`PI4GpB~^ zfkF~6eE)cs)ws}*^xA=b)uqjqdnKS!YF34Qxu_I)+*|fz0txjzd}AMqgLpC5`UZuY zuK`!NTaM`%Z?h9>UOEz6mgpzgFr~#0M%VeUj;ApihSl%?Y_FJ#CG`q@ za?9{BK+|4i9$9GprM+xlE!mQ9^cC)I9*GPgYMCC=`o4WGtXYq;fG1`X5cD*(>CYq2qeU z$=YFsFb!@NOURPP5B zk4%dL)I6F9YQju7ZGOCuZ&=RNKI}`sW^p~4_PWqEv+y_tFEv?&&;ja$MWjr~L@*CEvAM!0&&tOM+GLUyLET{F9<(*CFJ<7vhVYL;8O2W&z;WIYY$v9Q;Po)YZKpB!msLh_* zs8MOMnZX_*34I(MZfufJpxvep}(fgNK`)UDKAuj-A{!8YH zV0le+`pRX{fsr@3S!_I2V8&eydcArzr+AfAyVN)UP+D?pxR0ZD$s=b4TyyHgQR5XQ z7xmd2>QP(8hEkiK;R|r_zJ;4Dux&HiX9Y{nRF4vmnD$=Hjp*vu8`CW~Mok4UVz6D0 zDt)t8JMuef=Jf2f?Hn!UhaM>Jt7NQ>hw-eYcPV&=ZjtNPf7se6D-jB6j~_;g>==jd z5oV!0`;j4#fNANm9c=*Xt&oB9qzDPIQ;)v|sLC46F!0gXe7LA$Yd$6j#Y1js7b>sV|N|1edV*0=9*f)|eWoa3}4!SITzbm%fQm+ylkht#dB< zNo?=W<7r4hX{oiF&|J~YmjXX3s&WB#-1$q|_D$+tdhvRr7INCS3bE7_@!g*dpj%D{ zMFxvRy}}-A!-D0X&V9kOI{Yl+9*B%&7r@xOA?Zp>iiH+BfLu1ZsJ#0R$c%I1Wri03 zYV>f6G=Z0hf1~4_og!LpvdPs+Q}~mCFHd7W$^-WEKbrq7f9RAuBsz^<6N3$Ph3rL9 z?@*p@aRY&4XO`w6Yo~AfpWr=0t^x1ZLBmU1aK9svwB@(lK0}-?ERhVo9lmh;pn_cC z%v$)etJ%_6{MmFS}(k2a9rfknMV3p3|9E9B;46)tA%8h4| z>!3oa{dK2=kW$D`HMgY!1s_fb*)ll?kRe?p_fH`c0-3){34}i&sO5H=fbr>-b{L$I5?-ZUMyS@(D;38660=VoEnh&rc5C17f?`OZj{3qQOOrZaGZ?wbGASnf zv?_-DhuR)^L>m#dL|OE2%5H|f^|r>6rQ|*5FHySvktU+7p9N|!JF5yL)q@=X&yo)Y z9Xb=oM9MsdY8z>>ImW+I6#W;yZKR*p+R{q&Q5CEwkV3c18U?(&X9=WINJB3Q4PO0V zZbdcOf1=g8HCulrFPj`mf42Pjv4KA&v9_Ot#Np3OC=Edi4zv|L8H+i0gISA-P4{Pf z5G&ZL&&*lgmpu-(ri8Im+lX(G3!#yzFPBgHCqUIZJ%R4PIQ<}$;w^zem`0wkElB*o zcDY2(B3Gy}`OWUPSkBGY1Pdl)2V_QCaT0w+p_tDD&av|ggTFoWMKIXsKrBrL39u!^ zDq!I$2Q+!ftx-ToshgUbdbxvzZ{@rjfL`U4bd`XiWh4?rK!}&lFi`xIN~lx6ntrs) z4*zQ*`KT(<3}D$HOq@4#LTk!_JUNcS`8%}HHiH$&tDnoZcY6;9iaf#L66HmKD*>EL zc*K@=k4-_i8CvLRof}hf2in!&!^#Qx(*Zwl8Mvui1NhJ?F-eTab2HFR@Zz zqhMTZw}7-|1ex~#pR}s}0wYx9EE1VeSwI;c*|Lf$bUqnF4hqq{{Cn zlw?$1Nb6Ql^OdwgVjoZAhZG`V6_s(Eh4(tJ7(!5-B6wx-iukhHPs8Z@!I%G6dOpE9 zi7BCX^bRQoaw=lkKx!vDpIvsw7m-H}AXU5u3inISn)aaCF&OO_p?Y~4O8Ym5!^dP?$<6GuiT1j zc}~JF=xs&6NI-s;(d?tq|G(7=y_v??d4NB)K(7z$uhzC9(oeX8Wgm+>ykz$87P<%* z$Bzb3z0<55eEA(yU#+PNq`8=v6YnGe+9oEMk2gOMXe>P(_rQ)q#S^EN%KIr)HUMii zn%MhvpHFb^(na0LUWXQ1d4~`sP&QWVWz*vECgvl~0VU~r1rCRKAD?Ix2bN8tc8akf zF)r?#J+p4OQ?jv6m1tZ5-%F>$?C7A%i5kTbm61g zHvh)=B*Rw9%sU}xG80SCMSi$H%1J>g1?KieWD`HOre@?zpQ_+hS}07tF98Mj~WFeW*`*l zAdCPI|3_1kfsGmk7!x5HkL}eeGrlC+_r!l*kSWX<)% zgIeT^ef%{HdV*M~f+?bWfz4_WpO*g7agFo-x$^G4%I;{*n5vQ$$L~!`@%=PY%s_2I>3|m5?Mk-dCSSCA*TsqMoI*pz0aCXrqPu0 zYb?!FE63RWLL`J)?^ashK9um)O=G<2eOC4E$vZgmNm&#aG;$I5PTSoh>2MI<&A0$j z2eej;@MC54rnU<}4~8Z|BK}BBrmJIlNr}>)a4E<=-3&6gNt?PMvDy+=hJ?YRVq65l z$*LuRNI|m|7IpUG=Ro-x&Bx4#4QW&Ix_Z;~U-~8I!17f8VOH9tLcBa|Gra1K{u1Un z#WrH+hT2kBpgf44^Ex?p}HYFO&wW$M#>bK590j`N>38=o>45b3?lN-#`y+8^HL zI2Q=NP!}(()`GSs8kUtGt@B`@mtMbiL9A^lf~4{|K9Kupw9mMg!=0I~&S)>amJc<( zXD*%9`y6t;y8hR_@LrJr87)QDw~kUCevH>19B{R(akXi}|JpZ}dEb#~823SGCV6sf z@ZP;7Kk~U*=C-8c-hLa*-hNgKcFX3{1a|rV!XYo{mia)8lh)?5vimdQ%eQr!%dD^? zXa*02+t?lzbRSANcs^^2$NT`bPxGJP<03a#n?dDjH)Z16qAYVDdkV?3l7^k=YNonb zccsbH_U!{VALJfN4B=2=Vk$ZYwQw!R_+~S$fsQgxQIac@VO2-AsaX7Fg0;9WM=g{! z?VU-siKSxd?YBE74%A=S&*U!siOj<(C&(sJcPfaVP@yAKHNofpP!3r zr_VkIo(2nX+QQnu%_do^&)cV9%;c&e7-4^=>J&8gt@?^o)b*}(7+JSzh`JoGr*YmEw5x^8x1X~Uxwwk{YCh)34sXC*d+Xky#} za01hkf3BqdQ-NeXw?n;W*|YKVLA|A?8&MfWgR%gjiLu3I>by5Xw(fzQpU=ZtW~x+S zGQvP$d@j}e0sbh`)79VK@^z}F6v8LpnJJueu&fCG_<^hGX~?^2khoYd>Ke{AxYAD6 zwHj$T^ME2Sh19BzQsmYpF)x7~2qGx#Sj)`vpxqW~BgX#wqeJV(;#Vh3Uqo?rin|iC zFA@}VWgvllf~K2BOuBZ5*fnpbnrBa?u`f{p+FX~2f84hQ#Zg{R1`e2{*EXo?`B`R% zcUJ__0ZiJMa?%+F^w}Nm%<5-cqf*Zw!BwL-$vnZj=_y96O8c{Qs5K-0r+5Gb1A1QN zCV2Os_2?JBq7^6own`J1AqtxmDEgA}8f@W%iq+oV;Ij{`7IqT-ab5ta1HwiRrxR6I#`67$67jFU(V#|nlzGG z1R&OA+V9wC?@huB_!(`IQjzi298Ac_FzcF*b(-B7hZkbYp_TGY;ZAt{YN3^{{vAYJi+oc#|M6BdcsWP_f+?sygM z+YWcXbQ|mrXQRD152e$igqKENK6CmoJ5DZ}JMYiQB3W>s!mWSe&}v@Ok;*c@NzA-G zI{Pe`epVDc2uJ_?eyZo8KOt*K8#NRR1R^}%I|l|n&1bl>NzB+@GL5o9PPCO%uscX@ z(S$_Oz}XD{iY4IKLORcPEs;dB z3RJ}P_D-FMH5#M%7KIzPB=vQSi+_!M5fX=Kv+A8BlyiDI5%);2%(-89?siMiMVTp_ zZ5=WVPt@d)%&3tcPUNSbESdKJTHym5D05j&#z=lWKg-VtzQTuM45%WTPqlBgF3vT` zZnj`kk(O7m^mY|Coz7^gWx>wi;be63#HLHd060r|xcbaX1%1)+JB=bIm}>*~si-H} z5BDJ&XC5OK!4Xe4+4d+ejF;6_`P`7`kS?M`l=hiu1}iVJ+c(4dWZ+TgS>4(^*&A4U zyFvuNxGByMV6W|{Kd@W~gt3k)D5b_6QNy|gi0iAE^r4UC4^yWhkzGj+e${AC)xCML zcOa83l~iXEw(!`r-A!jzo{5jxA=se}&j^nxK(3b=6NwDhg^?6q$yYOH%ux1A(+K=5 z3nhp3J;#0x0T(qr4NUBi77fjMjul0oJbZ-qpbpTUVF_Eh7a&j+FW3!{HH$-y2y`+A zmcm8|fp;2oY{{=hU3S((!lydm_h9WYb-z6*UoKvY{K+|(yE6@lKGIVcD_9S#@>d(J zAVS!e7Vd%$-&vxM)Pg0pj|ApjR@JFw-E^#%T4F9+ymW6w#RvHk z_hVQ01EICZ?@6&1@1PKL{Qw`^j#(3;E=6ZaL`44O?3Z>eziWw=79fK1VBW*hcx`QE zGGZGuJ=1q=Y7aFb=oZ2|5}qrE_l3$0xIj^Praj9K6c&Rx

gZlOSYE%ko#gq737`RY=r`Duv%&jPy05D_YS>m6S1t`S7Yy>9C{!!rE*i0Jn;p}sdWO0S^(&plG)L^V{ISFa zl~-PO0l#{mcvh+wkK+IjY_JiwVs0+=yAx2Ujcr+$oKYUM=1RDF=|DV~S*ids;gwy* zC;PP%B*!-9NJ0WDQ{>;bl(X4tdXJkc_B37l!BLo|t-av96-t{i0W})i@rI4)4S%*& zL-P+Uz;~}w_mkG;#hl=3t-ienqW>!tDQ{#aYqc$$=Q^ER&)UwIw(lvb2N>sTC28&N6TxjzNjZJ zhQAp~s}CA*;varn7qg_0iSTN*^Kg}gjRXn57ZwTtwbs$Su}kh4JbY{sSN=?NwSZy6 zPMGVT22UTcZ^qF{1nQ(v9eT>m{(;y-WK=D#!rSbB1i&4QL-;ekpz4IRdT|BV8fp>? z3gAfvHtOBimbteJvx+!Xw_KB9YISh`YFyR`6*)w?r=w#baICJKttTAZ%_U%_x}En5 zwiTX|YNe2*vdRo)h;6_dyd*c3i5M{i>Ifhno3u3GxH#Q?aZcL(v`?h-lu6gRSjXD5Hw88=xcU(zAz+p0R$AILGo`h zvP`VLI<~+4!$)kj=w!j_?Nkm}$m?zm;$S369dy$T;rKtT0Dd53q+UqNEkRX)tcRrQ zl7kop>mEgK(3PNj8X^|Z5Jzx~ks41ptQ~jiU=xoV3 zxQW8;49k&rOC&r9P44%XU;z|?P0USk!K$s3)m)fiN(T+rS`@0tuca8ykAO``)8qD^ zT3az8ZTOpi6LdoVYkgk^aJlZ|t(aMF%oeuscN$oYAm{#D=V4B0!ynjX=`@k40Aqtn<{++i2x z2ePiilP#NxMBm|Aa7Eje5)pgofUam(v=GxxLcvySgO#R-D1A)`G4l^TMg?j$fknw@ zsJTwT-;|YT(10wrSx1|x3_|1&rL4pY7~8}MlK&Z$vL3L%_^I(Ss#rh8`b{0;B~=_L zhc+(fdk2j8#9@q$#P6AzEPvD$;I!Gu%@jTUnS^)`OE(7-7UStZTF6zgD{ML{tFh zNR5G!H7qfzY*k!dmi!nq&@m8<*av)Bs(}*oZvpHkgSo^R_&7ok+ssE?k$!l@`F35X z{@jb9z-V52-q(!Zw4;x1Eji55C)0{FZd}50bPcPHVe9h!9a1728S1~l(0&rsLIqF{ zy0P9_uP`(ue_5ITJHp-%&4@Lu&9IkGp>tAP3jO!e^b)RE9M&Y>)-9tp;10m=qmhab zR6MX=FtaDEwDe`0Y1CRs2tgDFRVs>uSJQ<=G#MJg%YC`0^yd3dx%jz&E*^Fg!t)q* ziP;`xp{~Z)eoRj44ze}Fma6Qaa(t(ex5&~pFd^RbyH-PZ#JYWDy5Ov??OO73H z;QCkK-kOm4vuAU<=Eo#|3N0?FyCQxX6tVM}09SP6g{4wU4#2KbnR@;|`gi5rom7j@op8#MyF5=Vs z@xddlRonD>03rv>qB@6LOke{HcA22dZblz|=>D!nO(%9n%%v;SK^C8W%-q`@A7Opr ztwepYIKF>Ni9i0vOTFT5blI@X?x$;rk6&lovCTt@Dv6!Zt@IvO$9b!oUGh@cfusN-Jp$k-lh81JUAE zMRJvk8T2J7hv~J+Y17-S)@Vl5=7&y@8t*@cQj#3_d&WFdv~J;6|L1$`S=#8Vj2DFs zwZ`DRX;ZrN4rHuD(8!+@*{fdInkQfI3~;izxG5w!#$TfCPM=PUFR|=%_1%5meR;lU zsVIHaw)8c%gSj9TSIEOui+)jvUjw`pRfNBczE^Lb>c9^d&d~>a@JHE0^M@>?GU=Fd zUGRqO!tr0u+Fd2|41W=;^~MVuomQg0R=M{tz!k6R(H{|)*yQpiE^)ErX;*D!P=qOKHUuawz(Kpo0Ui9h#G%Eq8-kq0VI!ermUPtiDYgaYS%ULE@e(m#`l*u$tbQ z>e3eqbH#hW zOV{>3TLYdmFf+l)M)dWAT29^tO#GaaHwn!u!Wn;gZf#$xynmR9)W| zmdeJAJs$6oTi&pXpMr8-=ITw)dUDgnVss2S*otE6Jny1Q&8ooNLFy}S=A1+ys&h(- zQ>2`W06@yHf8hQzca&;vkw^dn`m*RirpffZez=Xc>WgcPxIf@f+3kMDAJ-Vv{FJm0MWDa z>pe- zXmEuxP@*&2B&OSiy=!afpHcBmO9sd;98|3s)JD$O_ze@xvR7YA@tcw6{)kPcovI;H z6Idw2++SD*#*3ZZIHhgG(w&I;pb=D15{`5r1)Of+MAsreM&f@}Z#X=?P?#?Vyu~eP zY@+51V5fR%fJPG`0YM%O7_lTn!gR3BP6`}TX%p5oo|yMx&W>i=S2JyUOz)&*O&`fG zpe~L*nivFV^XYZ^CmhtwQUCV>!l!q|9bxN;Bg-00)mGFqe&T4*ddbnJ!zYBQ!lj)H zMfY!3V3b~gNznI8Hp)YZ99N~b&&=Z#T~qbnbgq<2A#F0$o3gj(c^hx!%E9RSN4>%o z`6g`z5$^BP^2aFKgVEjIp7D{<#Sq|JW;|{asnza>>$~IoLDTqph}+g+EZX74E$PMW-W8`^ zJc0V$x~sezFL28Hyw0Jkp!D{oO+&4a^8Jq^K@2mS5Fp}))K+a(A@nHx=~*^i424O zXW@yu18PZ&h>Y_PiJ1I!q)KqTBSldYZs1Z(NZF}Mot+nqsy`F*?-m3bGc$xW2OGoJC4 zv38^w)e%CHl?P_7{jq>i?GWgV&oV%-dOfwyMkOpLTZdHDu8oj^GH(K4$Qj%&B)7r3 zzZ#0Ag5W61KQd=R{B%7&r-4uY*%8awcZ#V+_oWcmZ)9@^GOzYe0yk zjlj~>x(AxifOwTPFD4gzbgQZ^I-4r~2A^XYeI)4XB_n7x4@SE~{?jYeoUTFqmtS`p zG|+$m1nJSN000000000005ONl?98V;hUVdUTf78^a*$SPWxvH_f6V>*K(^7(bMr?K z00K{e?e$S7CbUM|>GibB?$Xg2ucjk?F!S#HLGIfr)3I!hCpaRIhZ+#9M!V(eR-gG? zFj&@7T37^~2?I0mIe{?PPi`8V5NC!1-OAwbiiEUCZl*q7zPeev-?h}A0VaKhW)toU z^vy2a1#6wB;9^IyulkNkQ?A+8h)?#DV($8AGOAl=j)VVfeKfL1+4@%vd#I{-VfNhv-tk1>p+4PwZh680D=TJ1=#7z73PWhruTTJ^G;2qY`U7HyP^Bh zdcmQjei;)U9~x#1Rc!&7e4P2XMxeVKl&+8_$^Evl(s${rwR4w*IOtEq5nYE6YA>>G z6(TjcRDa#3_-*3)og5tX8hF=9pw6Dzri8}+yi_%DN_$C3tL=)eqZ;4yO)&jowS^J7 zW6{{@fpb*2+Ay@c2fKmhS*HnVrBxc;v_-(PiNCir^C2QL9-wyBG8VIMkeqhplonYj zKn^ToGha*t*huzJJs<5bqIkqJ>_GhG&TRVGS+aEPWoqZk^~V#}yjY&_qNdz&IA!n> zxKmNo4ty+0`ItZ2giE-Nws;I(Kr?49+bF)B?e{`nSqy;&9-zr=D9WROdyZI&&jBZ8 z7Nw@I;%tX=B2?@A8?1HRPC4r=OPJv~11EbUE@#kiPIBGIP{v+9kjDP1>&eJtUkbN& zh2Xe2skz!pPx_asj4^taNshUdNjs?Frk7~VMh++n9)iK^s+TiT$+`}$XW;n^uvzRN z|1z=)W%l?qzfjLC*&Sp$)$a2z?ypKzn<|RaC=K$0-{U!AhK4!SQN=9x0`QZF)H_^v zGS?EMF}M8MwE19TpRgcLDMPrlK?w_UjW!;e;SrU$3)mt&i?0!hlf;i~=n)w2!09D;L znk6Ed-6#ZS&{8YTj>N!im!b2Sb9gHh+Ic@7=}6MaDvVcp^bOR$$jN&TDZVl_#D^u8 zS)aG%&P}!Hzp~U!1ez@Ra%5h7%0Q@6pV{H9P+q+v7*Da=7Ji>wE@pzc_Zv-Vo*g#j zmT!|NZISwY0qK{xJxcVkO#U&b%Bv-ZKf}j2Yk_Nb?0S)w2#mMW5&^RGH%FpmaZ^f+ zGRipExR#o%IbNt27p*Jjab5T@pC9H#9eWmR&mr2 zRLR#nE)iSl3MLD{53$u%2_JCgwaD~ups4v5Ect8tuWngTN< zaQc}Sk`I8n?zbaho|Lt!KpE)(V4U^S(Bx);aS z_x5?my64zxo>ffvg8mS#&3j=RD-VcnX>BdIMPLgXN1p9`^u6s~3_x1Jr*FR7DqGyT z1>@eQrjLh~O`2G~_0dX0cqdEP5TO(NCp6l0{jxc-1s#2$N_{w#wE%n2veRsLsm}y6 z4i?EKC!JqEQ9AX%-FmJL`()z3MG9t%@U|oVUB_6n(nv^ZZY}9 zEi~;>DsAnhFTK7!81%A4HX%ngNk=}eyA~{VaeVVqG@d|5wwV6L3Y0jYsc2xob9CrN zRYts;FgW|AZb-~YFEHMlm-vxtuj8AaWMl&QPGntR6$o~ESZ3;j$VM~L2=uD4&GOK} zR~ripY(rHw_#;i6^Yvozp{^BZ<;2ZsPr#9YmGMN2I8|_U9lmqrP<}KWM6rX)KgHcz z5<77m(F@H{eZ`z%x?~ywiC>ZJy5j@3;|F$Zk+pb7qFZJSPpPD632v0#d^#3{nb4@g zU09?DX4v*K7$>s_f0^V|m*t90acx!y1${f3fhDeQtYVAG3xAP8lXpga^pOu>_sB?8UIlNOf?NEx!KR=4 zch~Ny_(Q)`R4mNKLVTBhbT)S->fw>72qnM6T@v=@{Az76*6Iye74O>#bvrF7$vQ^| zACcW{LK@!;NKti(cJfRSg20s4Ai*a5@6a^9Q7lgEhsSkER=*0I`DlLhZzGcEpqW`P z^hq=ne0Q1Y4-J^rKuNnFSZnBpd*daUJ-7()t?K<>iJ;GWtA1xlVD5J}VM1Q2d$yUX zQalPz!v)a zb2T3OI6H3FSWlEP)e3uL=mrpXA`S!J0}_um?TfvoThTA-1hIKT2%rWIm*h}_edoc? zfUjZWdy0=m)x=`a{WDd zD4T&wt+81^jDzL7fYo__ss8{%g7Nxov#z&7lyJkljzXKF;erLf+)I@@o)wv3ESx)` z`F?x4dA64DA|tuu_W8w^Hszl!Xwe2=a0PmCR~O+Gl`mRoWH&-cXNQ`Q(sm+Z@gcBs z(VdQWaNuJf`wplOf}eno$Yhl-_bRNZcqzCPa;y%IV)xsGz*buYcC=gbXR@0d`mv0t zdmQCN$zA1h8bx|k!&LSC2h6}Zftd`>0SQ(X#nI~F* z5EkrkwEq7-xQ{VyT4+t5B!s+#W^|7CrQOKIzib1+t@R-7I5K#1Up+y2F3!aqr*qYb z2hwV_`LAG(9RF(g1Loov!~5qs#{8!{?do z^(;GR{3S{|aB8@K!rahm0)dGI?;bBN2aiqTpUYUDbVTFqYR|3Y$QXQOQui?E3a928 zCLi#`mWd2-!~CPX{-bJB@CnfLI95m3djUt#M3jOTW8mduu>F@eoD)8jk+EVeZKe=N zRyTv(Hae4Zw8>JJSA&S=(wjBj`3y*(H_gyr0eEQA0l&;QqSyf`r@`9vj)FGhp=A^Z z@Tj7dqo(fh#u1Ga#=*%uu5ZnFq`)JB(tS9z7f6hn>EV8?wrei5)S4IbwZd%`m0QxH z^5oD6$1*?q2qY{BD-^7&t$59D(nh$wH%Z$`ef;AFt^6#Lqq&L&PFhl%6Y!Ypz8J&XqVe{DEDZ=^>(@7+Gyrzt7!^ zo2yv)XeJOxhK9l4CmF@Dm$HM!Gp1jyNSz{C`{4`7;ekWLmqjQbgn-NPyw z^5OKv_`bF|rYTG=O0(uT_O!pnapCVdnnl@&M3X!^%$RX^WDm#L(s!~UHK&St#VqB=(?PGGIAAJl3De< z!-SZ}iX>8ZS{rn3r{dB;gu(`8mU$RN+$$V;-~nF$oaVo{5-W-z>;{m4=&UIk*kyEV zMyF|eV9U<0+2lcx{Q(`RFkA`u0-7Hg)%_gWD+c$8LIAxsg_4|>>YrMe0Ohp2|MYA_ z5(^-0>$p0q{z^d*lgrWbS}X@Noth)5mC)J{&~NgkhiHw+*HVnywZOFv_=y!A1Um(m zTy+>=hMPj9fI65RC9Pp8B~w1LhjptZbWEg`fpRw3UWIAkYlRHjJqITzU<2P~AaCjs zvbvBqs!s`vTL)K78vV&zj)j4}9g^qtZOc{dc~WP zFNgq8wV$&VUeK&f(mE*oUOYb7EFG9)G_zg1c}@|sbH&t%s2(kq4v&9(jJp_dg%{%; z$w;m+Ju|;?MsNn!g$fatk)Gf1<^) zFtY8AmdRON9~BqckOQ!XBH}OXd<`>{x|8!-~aAKS2G7#QyA@=#rEu3=S=~r28^d{>L|!2`FWw!!XfT^RZgRO*%Ev zSN>K$e4hP}Oc*$B!ymMw=R9Oo^7ws1LW)Pt2Yew)Mo$hYhloG`el^TbV|wo42crUv zKvbel8`HWJdYmZhP0*H)R5DiCawM?R2IE^Vi?J|gS|-HYZsJ>a)3zuBklUuYx~PPO zP}|Pnn9JAin=9wJZvhpU4PO|AXuPJsx#y6$Bu-++AZaGpv%NX6IPEwyg_ZJ(KOGo2 zCgg=_6-3qBrVri98t(>k`6wEVU(8IzX< zbb&)I`S-8bb$Dd0y9Z=96LCMd=IghF)3h3#(<}wdUhA86qIXAPSeaJfUAgr(T~dq? z#ko{zJF|lcz;q~1tP5Jos9TmEJKG=5RO_r>myz>J;h;sSZP~}?@+3@{lTAgVxYc7Ta2QSa;-eEH7VB!k&95msno>vUGBu{?HRUJgFZZ|(IigC@-e8aHr>+!gi)tDd1=N-b&k(xLlb%tZ(8a9# z6_K~?1!q?`b*`Q&_}t9{r^WtJ@WMer%UBFUt|>o5k#+V z*aCpAU^|&@ZtJgEHCSao3{eL;6{{bc8GShb4f|W6xS_V>>#SZdxPVKE56a}{)Bg@x zC$X_xYAk(e4e~b<%n!ag-h=i(a(DL&##qd&;U4%c!7AN#&1;TQ(iClgoGSdyHW5MoCEXFwIg@94B$UM<-I9`?X)lc&*_8B z3c3#FjlpZ|mz6g%7xvQDGjy_Na{x(~9?b=ATxJ)%9@sGeA5%@`2w=+F;_8*IL9V}7 z@IRuLmZdp8OxyCcGJ`2Be;*9@>$-neOulvXH`0Q4-OatuCvM`NTHdtX7DW}NZ#@AZ9( zNmDi;z$~KZ3?a`}gs_XS;g&LkhySO2{YJ3<(XTT`U}_6R+d2Qa_In7J9{4P}tR=vv zaKaFuG=fxYq6DwM;6{ z+>XB(T6RGoLW?zm!ezUubx1xzfcZu;VL#o-^il+3@x;4*#E^uz}7d4tbNr|0le>pg_TF4H3N<%ThG_9 zuj)IrsK;sbGU7+m_hFd*t7-P1xtSV!YNA~Kxg=E_1TSil{9Q)%<(AuOfp1~!bgCOl z&R_uL#{PW;;ljMvaY*f@h6NFVEjN1!##A%vSq`(Rj01(@nmt4 zKTUU$ivi{?j$=^#fXPxKGDT6qqGcJ~D>=M|-a`0J0~gZ7<&x!)W5godu=HqW9C~9p zPlaRZpfdzgzHdk$ClBwg{A|bpCBVvT{N6ALRdEq)`IENoJke&jrNZ@?4-pKEc%(M7 z)O(0YeZ1E+2jEY9SCO`aVam@FvB^*Hs5?l<1pVPs8%Ja=SVKmLj{TBK97JuOtCQ__ z>uDIlCz>;(9btU&1lSjuisd$X&lu{rJ{QM{(>@kb|Expb#vcp7PX9RUltnyl_#IHLbPec!2 z)q#%-?emX5(;=7^FMn=O}l<8ubGmR}hbRi4*bIbSipe@f&wF!9g>_eCP zGy%dAa#5Rf9YlkfeUaDU7gr;Hri}*|G=?O?W}HEUI|W^bh*B5^g7I}1 zuF=l6k&>mUJHSq4ewG#e?(4|k_h*Bpl%r$&BSZ4^3QmZ7-=LI<%~rvio7g4nT2^Gw zTrxJdw+Z>Ryz>sZSd%qHLuJ59g96DGC^S2c8Y+g4fAZ3=K`eF@S>e1JN(*@O>7wx5 zUXkM^(M=%Kr7JO#3JKFYWem5UrdH_o2Y6u)rw7PooY7EWb!N?w@`rme9l0^IM?a$v z;yg)Z5kjI@ejuxT-3?Gmwo0Pr%{WB-js|Dzq&O9J?x6qA8L4T zvfPsLcj%;%@OrK(!OL;G;(Zpia1gz2x-?IV;*InX9{&b&FWm7SX|G-%;Rlx`KBr{E z$n;}-d3!@5eLbirg?3bYT7tvNTt2Y;&M)nVOODwJz6kM(c{dl=YOymw-N&?NF+!C< zE>63haq5zsM*3J%8#3eY?u;eW#icn$9w-UZN0IK=rgEpWHH_mUM?U(y6@+hlkan^l{B>RzTVN?Nyev=y3z?z%E8n8@rlbY*g(@K z=hNr(AHv7DSS7cr9GRnRCE9Ww@t$wM&tQ%c3w4P04NvNtQ6=&P_%w&kF*(QLjH!%O z{_}tYt+F=YEOPgIP;%DOiI~2URjRoTGTjmJo7|e0s!VrQc_9;{bUzKjY2_^WfQijn z!eiCY9X2gFii$CaGTlNQmJ9A=^(K|&ReLG?Wz9gzcpH0Q z-HI+2h8P+S@TYsQ_r%gvaxEMN!>P-O%n1SdW1sf1-5mcFcOGedN87@^iw?uisqZct@*EZvDZC=!|2b-OLm~ zl`)+V7HuA4GRTIfBmhc4wZEQEp(T`fM^cY~gd#!c2As}1&u1`hDfE@B9?TgFPJC}0+g?=BxNuLB@+T2yui;-mER&W(LU?+> zW+`qhX}{Xub!Qk$fI;NBr_qGmFeq$1S=)egP}J(6XFo0j9FTneGVo2f_NkUL+YHK# z?;eyJKO#1$|H{aEnx_7u^#$GFHe6UE;P*|!bm}^|UN*jc&E^Um^%(p4ceR>^HeoK@ zQu?n-XNC!yu;EZ#>{I8H$OeHUO=|B02{DVnxR)|%I89mqeT)=UiiQ?tPM}sICl?SA zAua3%q^Tp3+;EU81-{!8=>WJ2DXy>`w~IQq%$$iPl*yr3*iiRo_E=H^j;hdebxMO@ zDwHJ5W{o3c+SkwV z$c}FKZOz@k-4OfSvPXjbT61uS$x~N%a_&!LG(4_`6yi5jO9MAaDY<1p-^^CM~j1_n$j_JqEH99FJKIK zc?RtjQ6<=BO3HXD9i&5RDXW0akC@un1e{OIB|-wg?@C73AHt`r<7}B49XH0~GNA9Z zOo$-X@2% z7n7zVpQ?tCoKxNp(uPxJ*5GwJ_!;DHhqhQ~86u+Cdp;J*NAMZ#iA~Ep5gm!d5f)Og z!Y3%h)X-6*S!NFVy{nJYFNP zHRM_z@f-jlCo{=yX*R(Y!6}J%^FkI>O_+2cabdQ1o6hxpR*|d-bKvp96U}*nd4X;Q zk1elR4P8l+!8>)#5s*6^Z(HpIc*lk&;X`s2LVRc0!Di)PYqbZ0E|i`%vyGngYj8t2 zKUNNqZGmCm8L#&V@CDsFTp#J0m&-_#{L&{_lPf=&8;6V-(_Fa*8-{whnQwomZMd}jR86flT&5V*8j!s z>EM7Cz&36nU(t|^*GCS7cYsispj}XScF%tWuEh5)#4n>1QO3C{KUZvL2k3~cj*N$XE%FxaIxhYRQu@0X-SRU>u+Q`giKG=dCz5cLO?$N%_Fj@LYFgwqWdySRyjCOOkBx-+C59hL z@fO8%^Cd?HIucXTKGwSx0TX}Dv0;>Eet;$4TdxDmQvsW_S>Ps?d+$4Sioj!?$3C zMN#=KtHqwfX+8Evs? z9qKMw*PSLq>mUfk$Gs2D5)`1B$+MMwGaiWlR~mNf;I+;D_?I@Mqn)3Qj`CCvhTz7K z_;)t?KDDjX?s0MWQ`so`g<4kr_6u-HX}Jl7m=@(TwkoguuR}N(26zM)+*F&u`Q566 z#Gark^27t)Y^-s22qRvsp$6Am&=zS=y}~GOs!BiQExRYHI}S*>y7mQV5e+}Q7PH#3 zp!KY3bLOZHj+1pm*D`)LrY~t_t{MP#-|R>au1CGZa8Iu0Z_OoxUxV~HXn_3^06Q*E z=Yol4jez4z8<7r&I+E^1!(4jaGGL&OHTlRlcKeyJ_`!lUe_KqDA~F?}4yU#x6U@&s zw!Ydl4i(VJ`Q!om7Clk**%`;)!^?qXNCw-yhlFMmy$7gzYy;=>7=hE-I5sYivSzeS zLyAM>OMMU`WJa5wyH7K~PasVr#r~p#5XvpYzIE`%ot)mOyYSt@O!=4>Zqp$!MG4~j zu4sFYUoel#X!}MwmI;F!L$rr`XKiG*d_G)*`Lqw!Xce`XT=FDSVUT%EMavKebKvq%cnS_rmi{)@A|YXR9HU6?NSK$3| zsF8?zfCcesLUT$dJ94^~(i6}UzWdcnlW7Oas5Kw$>4JBbUPfnkyRN@giiZCPQ{5zy zLP_%tb{`pT5Vw}oi!8>Q<%D|;=-;X#^k~B7=1jv@a=6PzL2b;iY zl&+!M!JXz)+XrROl_4+A|Ktj85V=!y+|T)6MaN0Fe8Juun#OR%48Ia)y6}Bnqb=U_G%JHm&}#*EIy0neICcP^HY=yck?!YXyKKR^NJlx~RD&9N?8 zDDVBPxh4NX9=!@c4V@D+0-Hn~QsD7c`MUs3i)Zj)Iq$XTk_os?Z`U0*ItQl~fpZCa zxOe}OP?JE5Qdxbd&IAawox;JaF3<1DSeF#;3kM-k3zQd}w7}c_wcEX)QvjplNQdq- z!fF<{^Z;0oBFYcu5KvDf@;emIrNTq0Nju{SI1x&eFik)`h!PhrxRjFA5wI9)v1KmV zD_dz8g){a0xCK@6v9V7)aDzRhkH@rd?;)p&W!LuT?hDco_DP}dllSVfA6rApXFj-l{#8Izwx{LpSS0{r zEt=NHF0wk=6@Fp?;(P&p&$+0vjEu*jtke!CjUp5Em*ncS$jMZ-ZuRvRHc>PQ6Zged zB)MmT?U#KHJ@83xjvGn>!F!t|#!PgIehV(*#5GnC=fNTFq-M&E1dc&HyMA|FZ=Bs; z!f$&{SGVN%fV!_!iG8IDa-EqvQ$(+x!9_TY`QTp7>vlyWX#E8Yo9n$nuU__-VGkS~ ziWb&yZ0Z1=FeseSICozK+Mv?xH9a8+bbS8A=uM25u4Y3tskk3Gi?YwadZ;r^D&9^cQ{4qV;Mn)Q~O zgKHR>jD+HIMfn^4yXNF-{5(}KI{i61t#6*C`2Nlc`1FHFaLdczjTx`+r(PP3T>Cfz zQfi+2>LY8Mm@wD`P7GgCI*QW68|WOdExJ;@fOR0H2e2(CqUp0uxzzb!9>$5X6Sclg z&v{RhUpidosvya5KN3b5_B>gmpxlsfVRCb%C}PyDR5Gg^f#N0g_JJ7EA>p=TyDiqr7EVO?0E6UNPZpxM-i2t zH+sfPL(gJV_>)~$_rbZ8EqdJt$OpftBPMVGdtU*^dSK6qu{kCw#q7r`Z{>+<=(vY* zIttil+PH4)-W*{LXNqHO@1=R{*B$vZPfjHuer8|!aP+K9lHmT+s{A^K=Y#sg`Ty*1g-;$*eV_OnT$b9{lAY*>$=M za)Tnxz{(ClSW8DJ|CE}>uXuMVMu#g(JWTX~n*CU0PK6S@;cemV<3DkK=dn~c0Y8L< z0!DW7(al#PAoOpPPEe~ak4(Jx%-^K78?JZkE<%?z+PjO2?Q7g4YXva2y*SJ`(`4yF z34JZ=qcWXLAu8 zYkILOGPKot(TTps*qwvqsC5DFw(3jbEBrc;J3=gxP`fHxU0Rl2{^td?X7Sl_^md*c zZ>4b>{bd!pwPNL&l+$NSvD@DlOAkBC2K}7gs>IeNEQd+9FFlGXm+6B;^pYff$%Qs` zXvU9$lXSK{U`m#e%Gq`jRkLVQmG`BlArm>!ATz!t+aC{ty1XpJK zGtNt+o%1Xczq57-huCC_P**@q;_Fu+Qr>17ode5AA_JyPPcG!oc1(KNFJ=O}euy$= zF^6T1`RyjT7C8y;+Ob)hHTE%R?|sFnY5aCIF7V=4C61oqbT#$(2&yMJvySe|zw_L@ zgd`c^b~swa6SmioG`nGDDDL65%v$Lcq88WJSeSeNhso-+c<-a3W-Kr;#+;j%grpRn zE)=n!3*P>I;X#-na; zkC2h@2q0XEr&jQbVWM$t5Z+uJbp!(5oCwtEOv~&Y0DBStTjXE+wB%Fs&EGZ)4E(-x zCu^01G?U7i8RJ!Mfk)DuacA;4kWb@MXGLnhXn>!Tg>Nk4HvVHLjUp}g+%bKrF1AO~ zS2H_H_N{o<{~GM=B$kVN*t}xbKzwQCOf~tMdA(o_<68EID-zQP6M|e2ktpgG%PLh; z3ADM%&G^1r%VZr}5mEXYDGeK|C^92OoYJ57WDkK^VvGJoWfbURI=e^@Tfnz2TE_EJ z^jUjc8@=oa^Ei7T{ZG{iB?V(ep9&QX5rA|J&xha2T0Z0CujcexOhQ0tN+qKsOnmRB zp|WH<1%5WaZrM+ZHh#GF@c`31j3QR^?i>kW$$|BSwd$$#+(Mv$4tH+x3BHupS<4Ux z8P(}mrXT3`G{6z0$2Nxtc%_JsV6pl5c7t6LHze1kX48L7W+6FT68K>>>ki$RY{AS{ z;xGVTbV(+HIKR*|O~G|=p6g!GW=lS~>+nQs$9>dYp?>bQ#1}=*%bheXBpbSlJ00}s zU71GoEE?drcdL=|PzD`QJ}rO^iyiTG%zU0+{H*AJ7|Iw0<^li)*Y$@Z0>TE+NxZ&2 z*)fVEE{{e!_MC_$_2l1}*FDYrurG`q6A|P=yYe$l>2&Mp5}^aV!T7U%bU#f~w(cQ$ zrp?;WBM~%LUwe9_X?phV5zMI(&oc;EC+N6jfx)$FxquC=Pt4C1=xm@~cu%nOdS?$(KF<1JcAFA0_g& zj}3Z<{SSxe)%r*OocqjZmLK#hu)+A7$Clnu*!MG8TJ}68DR{r7gC~uc(+JLUsdM}H z?}wA+YE&s@+@;B8Z zoK;(IYq{SsqKz1tfB_w1@ruI2A$12u>TG28!(E?B7cXYSuw zF*pajQ5UY^YKu@7O1uwQh9KU4h`{K)mwqnXbk+4*&it#uXbLAy5xK9-LU6$d zbkF*sG5dV;Srp-l@URqE(s=5-%dDR1pzOljt&A9&*!Io?Cy7L!ZQ(-&Tz$EC_#ysI zJw~%v+Q;9vvuX629)qT0c3S*WXn=y)OY6Y5!)xs%1gl zN{UWJW1b~5|BG|YBO?esm#kXiXh(;;aMQiz6 zK;MWGXMXNtWxJxC;m9ws&_0Jk$zW^2Pcg>|Nl3l_#mY6(q$b;GK(cySJd|C&Hs8>} ze5w|*Ze{E~Ulu0(MNv5*%&+fxSF23bRkcDH>=l>4Yv$K<*t+(Y*w{)o?RzhRWQM44 zF2g5>iGT7r4}#P6G8GRSLn-jbD2@7lO>_cx$J5IZAN8At*={OB$WXgzH-Z~PK0?>n z#SY4Fx=edtcZ=Ja=lN*)sD0|(*j;cE(aPmT$WgrwO9E=s)sqs~r|FC0>#$?Dtc3Q0 zU^X!~M&g0hU}P(#KJ{uNj*!lW6?o{DaIgT*8G=sf4v5yJJ4!7gF5n!g!iEJj$2X_n zcwinTC#Xfpds=8nG(okLryrcy1a0#ur3MD&m5ZHpQHE{>EaBgH?k<;k%KF@uG(1Dm zP6+$V3sYyfotmn&5qqWUgGAH6UeJhZPk1VVzSD)Dz>!lMA|pGMG}MT|FzY}Yo@OIz zYKwv6XU8UAH~+`__Mp8h;+HXWrXzWFTm+d>mLDw#np0i_gSewkJ)t`PEp=_&qS4|p zrl4260QYvLdOG)>#X(i0!HOM^$IN_*z0t{epU!r~;>j+VUX z8(j`U@ngb{cl12uM?E0D7#gX2WsoR{Hw@Zs8uz{6JlgX202Q!QDNq>ZTta{`1Y_O> zFHCej)y&iNtAc~;=%)_#GXXxJ3R&B2zGha>Xep!P5qdKnCyYLvK|pPaS8Y+3xkaWu z)Ll1=IP=5J?UeVigxYW9($=qECd&t~eW-5B0ZGo8^Q@LiVR129+*T;=0PN5%j#N7S zHg-&F@6t+88A;@)|AS0dX0f?CU6<+x%qR;$yqG8=>N`I`1tAf%AakS%lR7C>1*S_K z6XghhURMKRL6@;)BcP7}LIY;sXfXYT1S=!qOAswNx|ii7*QPp!k@#dn+f&0F5b&T+ zEU?=c5D3h?=5D3-F>K}1b5Q8sr-x4SSyz^+{TN;QG<16a- z$0r6`w?Vcl97)cT)TPL` zg+_{1RM45`dMb%wlDk2`CWk-_0h0<0PGoIp4qzshTy|y1JgYlqYd|2HOY)ONJ(*wxId}LKS@&1GXUDy;9|QI*c6stkZ7C!(2JjS(qGQi zxRB&JfvsRtnrex6m|DetD3`1v$_`~7b*^)R^7`R739YHI<`ZuO$KmwL6{z}LInZEk z{FxQ${>rej$vS8U4McBWGEd)Er;9@HXxcaRw^3Mx>hT{dVjE2%ky}nHtksbL`cUpQ zv7B+mv`qU(Go7Dwi}$q`lUwOQzywvpM9S4_UU>&w{tQP#%jd^1iqsi$>!7@|+BUQQ zR=84@?hBPR#)NFjgWCd1YQa$Wl@%6yn37qgh457_k}O#1sxk*MK@{Xy@4%-HjNN%F z!%aGSYB^L!6vaNg^zvnFyOKU+g3C66gQjA!tELxqr52VrL0%7U#&NhuKO3PyXfGUL zd|cqkG|yAEx557$^i&}5x3aXA&|z-rGx9Ja?ootpf^f)$9G6hu)ZmBX@f!~K{>o*f@BTArOX9@CJYZqKX|4Z_v z8RU|z!_)-`FD7c?Ug=j9D*IGi@6x{uYGj>SqWh64n3b&yZ_9VQpL);HEn!@}Wo*MP zu+{u^x=kE#SyEwe{Bi@;<|C*1W+W9KzvV^9wgF>Ef9(gB?jNvroCxigdu=Xe#9QO+ zs;mP%R@}cVM;;_%>J9Q@I$pzUbUb0Tm;C?O$!BXjVz}!YcpXT0keVv@Flz6>E`SYP zj-oU{&9v^~IOp;wAH~X7mv169KAiPO%Uk8t^5I@JeoM%DN0%f+G*fJ55TGt%Yb-}w zz~tKWcL&KFIL|?#q(u3>uA^A9PF9_@LB^T1)EBZ?z@B?|@!gpK{qJ%f=(C?80V+0p z$Xs%aMt6QGsSl+Tk!t^)~;A(6A0-o1G-V==Mi=TXc&>Kk8D7Y!ER_0d&~ zmO7`EgxvW@`v&v#!s|daU1+jpQ1;PvR%vw{a1tAHi9FAT7K*Js_zgFwY+#U$p6`rv zkWuZ-9818uuf0&aFQ_-)yr2l@6PB{zEiOd)d-7+l&!vxwGqk5g5`XsoijM(sd@d;- zYXNi}s1y?WLT)Fp&!UfHp(qf2qdW;%^$7$q#7YYT z#pC=_XJKQHQ<_EWfllyl@l+3JP7Ws{3B)}C)-iTmQj{A8j5jC<_a=XnOCr#t&B5b%%9(4K} zEfi*CDVaC#>!ML9wW*PVu+qng@+cSHIHYhRuvpu|r=mDVfay0>aBgEWDd68L_@$`< zEy3RMX3+1eEqO^&)r8E);mxP4T`PA^&06dr8}S%&I}U{=-d{jP%ZSv(tPL;;9ZvCa zbgDsAgR;nuURXK_5#2X`Y)nn&RB!ylijnv_PkJDKNE2mZ_Oq}iDBvMSl3b2}+UW&a9?6+rZJGaozgciT&fcCT z;Poza&FFqC$pORZ5?#rOCD?~8woHI^%4nN@*>7s%$5Qu_GJ^1_z{>Z-SxP4E#iRR= zGIT8Z6zC1PYI0P!(Z-fkt33_tilTl-bfPB_j2oHksdza1cnE@vWb%2I=JrCiZ^>`H*dxdqo^PA8*pmm+ zcpB2Wy<@9D=mQM_7@YP3*JcB+o?mHY^bIU?;hXj`S&Mre#B^zo}d)!A_z4 z$sdDRsp*Nql6*20v8O54sp701p0lexDyHKQ9f%!o9Wuc*g)Pk^*uJy>U zljq?f+t>!rlWQPED+e51`G{h(5{$?5& z#2lZQuMlGL5}mS{L+W?^NNu0A(u%Px29rviXz|a?62a%#i$)IO;pE~L&8D}dBMtG4 z(VA%Vj(o~qzx@BhW8zl@Z2};e!ezyH(=_|YfX6dGcC@ZzOqh^Qf3zy`lFMS($O`-I z`B5X#`MiqyVj`e%tjq}MdaRS__I374EY#d8zR?ppR)FPlolouRRISgB@^5LE;Y0vb z2Nk_^E>aglml}}?Ji!3PcY9I#pWEE}rQQvVg(gb!UnJ_@lyuPd@cG0Ui#Gz3)25wF ztE5TJgTtwupbHTo7vl2zX3IXd(?>)-LbzexJv-Y_r|x>Vy5-y!UsCY_LjuQGe^VhO z5SHt8!%c=)0r8jN*BRXS3rA-FYMjC*!j{WaO<%O#Ay%EB--JQs!&G#jKpAv~Vu#r} zy{N6YiF_BfMQ@HwYRBtK15d%l>xjwgaHO|c)Hkft zCx>0^K}&ZWg_RJHSp-}=!OzimI9NSQ0R2hLX0r=N_Q1E$jIP%HM?-)j6n=*PQ`DY_ z!Zr+SRtQuWI0hanHzkF~Kv7P-UCAv?qkbfeITg>ci8Ls$i`&&*VJ?yuGZz13{~_VU zr&BL;448-;S>fm!iV^1tHxH_(*J9ncQYcle?#j_pE#3k+2L1MW%y^GGt#d{+4lNy! zt7U)5FH}I=+aD|xfAxRq&etEr_a(%H7st`fyG>==%->Z5n7mJ?o%Qu);=rDlF&_O( za?Y~eV^4(OW*)^8R*<835Sjcls(IV+bO!Db&PfwX5X4W z27I;zHK)CyAx&lUZ_t6%{mu&8q+j3~khMXqJtg%hP6hi&VI{Uq(iAn~Lq&U-PF-ow z?Bh!31eb+n347!OU%rd6ez-OLU|_>VDGi)t0|3t7lnb%Mr96RjB^&jYKx!cnEYZ7tW7cung4mO?XX=3>!LoM)dzJHEAi=#nN81edek|xv++W#NTpLe-q ze6k<@sOH*{DC>OeH&`vj!^OF<-cv81j&(rpTV=|(>Rin`X#GF?I8p__IqEXwP{_U zwD!@4PjI$N0{-jGxPJep$DYRsjv-(1t69}nGB#FCs5u;eLb4syr_JJr;k!Q6s|REB zgLdcIqoo`TAP>S~riKA#;TMb)d6f_*J#*$PX@t25iyR}?zvoAS@nIb63ZcwS^5pfV z%+7q~x4{e*UXsf&iby8|@bV6Izl9<|FN zaZG;HSOO4Phg^E2x#ky_vGj_MsHb{3gllOV{vAYafHRB0Rz!sQ4a$cDZ}t&+lsK5n zCBPViMEcl}S#I=`T&*@(1a5$b42JO{-m1(Zr4BBJE<#Esx>k_jq)H>8%HwGd|1m-S-;<-m=nx@csdc;26U<$QYQ_XKwCH-5N`HMprLEgf1&pxBmX{iW9*M zxv(Z3BOBItEmJOJ32j!B!qv8J`Tn80`c&t(?LVuS8;5Om%eK*EjIygPHNg`MD1G+r z5zTMN$}$FDEp5b=rDB^YVb7{hSOt>r$a#0Y;$XPGMY>7v1Rpq(v><0_~2onQIl@KEgD6l|R zy}PE+Y(pO`6y+!Hb(uzZa#h6AzYds}BuVh%LUsKsqa{)h4lrZvNE00ghw7od{1PGP z5DPI?mA26SB{w!m}qme3sz)s{A!*Ro0jGq|h~SP~{u)L@BTT zS6vGMK9wx<^rkWDTL_iTD7!k9IWp1i!qQ{X{LBhmtnR`O-vYp`cBxreja9KXGYL#h z>MQT;;o0*XZ2F$6oIsCq*htlDAInMEYe5@5jA$H40)&1u6{?vvEYdol?JfU{yBan9 zj77r;3ff1u=2QsNB=;aLUJ7StFL8NK4xwq|fQGr=bsA+hmjXd$A2vlQE7NL$Z!2QT z?@&6?APn8V7>nklq<1#-Y(ZBMBB>Je3y6Sp>Uq^c32EO(P4v0zjIBBzlK0ilTmx0@ zXaSS^dNEu4>1|k>$V5 zV)Cjkc3`TX0xCFmNy#*02GCG=yO2KOGG+9A1>e}6bbs?hKf4R73CMQGo z`_sFWuG z!s}-&2bm>@C_||KLBbRBghY1gF*M}Zl&G=yUYk1%NBdH3gv&lJvfvruxGPbS+?owR z^vBjtOkmG23Cz0@LKIEGAV22HDH3qI|9zSm>fzsx1Y1NEH4F&Q9;0Qfo@^HwBespjk%s5<7iEQ{=(n68PiR|YEFa3~Aq z2|A}K8}30BnK{^9g>r5`iVCF1*+joX+EN`2ucd6l^RRJBW8Stq8n>Q~$J)gI%R=}u zwdiI@v~RT#?NIvjo>&(uLGV*ZS;TORW!B{3Gju)9++uqS{vNOgzCY8Y@&3?(t z+{cbQr1Eq5JN#Y&?V%xcXR3OqR?&05+*^=L4ShQI5|(F=JQ9a-!0V){Bo~!e``x~R z?*;>Ch{jNyX?VaMVcOZ_l?Wf15_zWi81v=82Ro;ufpN0?=@Q&Q&VeT*;%1V(%nq8W?%;E3;CyK4vOx_Uc1pd|hdT0=a}!Yub&M>sAJg&tNJ$o9>>B zpAXkp=<*AcFlNWov_>9YVWP(BN{nWw0Gc1*(q>%-lCLYV$ko(byt=g2SV6Khe=9@h2 zFzqIc!W~-7x-`X8vF!t}pK9lNEbx@Iri{-RRt>2bos>z`(7Tt4Mm#MkWQMPHp!hp@ zd)xo+art-bmvgpH!9l95pu;xcYws*Sw_G-xq3K!3>m)02@J(3GT(t1vRKqT(vr@0Z z7CZ{QJ_fIF9)Wu_0!)cbEGA8U-b|`q(+D)grT$u;0@q_rALxZ`6qi_wodiOD5{kR; z%a~}V6>X!If+MzUHJBol7CSaWX=WdXAR*i;rNmv0RZng6A+JSk2r77{gz*WN-KViIZUvS=>Eh5>n2RWOj z*uI(Ad9Eig{=kPE8*J~xTxYRZA!yEn^U*dWPC6!4OttQdV_mvjKbzRAUDF`T#~S1! z*BoIy+eW-?2Vb){3q$|Vw#P0ekk%`oHMM%e%?3t$VlrO370rV{@{CY89HZw@CWZwr z)HF3?P+%0*GHgPpOuLerIWmfzgi_>M_;R)M3Q95n0hbgIY- z$OMYuJ<3Jg7>}ZjuxN5*GLXKU3%Vw0WU0gJRD(*I^|)}xS>!_!LJJu))Lm_AkWj?| z++o@ay*uWA(zS+J;(SpV7o<{~d=U@B$cS&AbOd{|qnjOIjcgdK0v}rm+$#^pZ*pn~ zcOKc#FyHoSmHW0y7DtvJ;a`j!iyA>AhVv_O)T*5-m8s{_mqQVn-FR@Myr~LOE77=# zIyD@w*=d}Tttgmu9#aWH? zKKJ__zAnBQ&|sQ7_h&y|yOgOwOo@joZR?{U#I!{W$Iil#}!D?vNO;I!QFkOFkc2R5MUHVqjjcf)6$ zQ;c|cM!Twm?Zrv8NKqyyF#E6XKJdvXS$_3hVp%X{mb<7o(+K)lzb|*o`8JNgj*c#W zm=w}(PhAS?o_Vj}iV=Yfml5djhvLo&vx~;3#>dY14X=o1m3N1QRPRb9|E{T(zug-*?c)=4@L~BUBs6p z9)RiX9|d&G2hRNJmni7)%14Oe?nW&A2^8D##_T&s#tYKuVJx=Vd9kWk%i(siu z?HnSMgxbQ$LGBI~{u+3Bn6*>)J?%DfZiMSZY-ns&Zb?W-bCW0;17NwNWEGQ@bZM8^ zKIzbOTC<5M-w59wPbP&Xb1(%JoZl)ErY}VlQdS<*^#DfoK|189MOJ7H%FZfNG3$q( z>|W~P*p0Cn35mtc>W>L`FC<{gF2X+^^IU*QaQj02JS@pH+n0uh*5cIdteZ*=5Wj{R zYitA{JEhbi6hv0BwLB~v@7_k>aM@3qu=rPGH&t!v@w%-{2Z_vQY+%`HS+GrR(xA1~ zK>y0_T%K|IsMC2#U$)etuc2iKc;YglPD=nKVP%Qqii*d;UT-Z}9@5li1opbNqg2Z} zVy3j*_&&Mx+rPsY>50JsiGllr6955g3;zqCehSqiJO2s1Mcp&jDv@|REA94pcA82S z1gCzQ4-G#i$ea`8|4AG?e_oe#1K2Rvhr#_j4H@lzRa$Gz z0RKPmq8hjlcahhxBgY1ylxjd0sm1TplpP?aVyt$FLt?-#aIzQYKOv(^BpI-T%2Y7h ze9vCWM_m=;>yV_YIYSc=#7Z0mSt&KcF&YD?4BI};dUmH^Z=BeBNYkDWkiPJLJ~Xwj z^O>UM!ApA`tWy}}#0we!-XnBdP#w<}AER30ca%|_f+gaanbT}3)t8fAcC~OTvtBg= z9%%UpFj_Iopx~Zp_t;xC%_OzaEXl}$^}?dk-6M%HVZRg}LW;xUEkJ$qWv^8U#xUFD z1iD~uWzq-`E{r}Wz7$dPc0A6l4oYe_nPOoiHRi&@9=Z*t29_Txn!{ZpZF^axh~ z#7f+da^3FUhgkhnu)Z!GMp!R2n$k0_-E4!CLXXl2Q7+?4$K&j=?$7E)$n5~NyaEUE z9)A%o`%@1Ns-@i$u5)0L$?>BP><;t#E^e?B7Ptk#h$m1NY{}(9g}%xJIvs|+)369?yI{xO>VT4+GU!Gye z$P19uzz^`E^nV9Uh2p`mYLfE6|7`avxGqNq_?&@&- zte^WYXTD=ff6lAjnbT(f3B0+xS!#ijBM=HZ$}Y(ci!|OQs3F|HC+9{GU>+{kkyHMq zA-L4iVHYl7upbU!wfH_(QZ;RlqHNp-_phFuO=#SS^U`+B6~5;fOcWjsl)!B2@K^HJ zp^SmoIp@DtwZv52PHc$di>A*?Y>r@S^=-mtCA3t_>T1t=tPv!yk+s|&*Ll$r7PuH6 zc$U{Xt&BA-=%?6XvvBTecAOxI+!@69>ERvoIe=<}#*8{R)teJk*_u>tvhd!6-SY<- z>6ZsUUXt{Qc&|nP2$A-S94Ph_?rc`~#i^S1O~(BR z?VRk*BtLj)dKj6CCX6lGhJV}QNYJdusbsVLTVA&$usO3v2OWzCo2ut|;t$4yOe0#) zXQ{M`8K^Yjc&fN}ErB&f9vPtTOx^0oI=GMmI}5eYZg`1dwxg&^863oz(h+d{HWiqoxoR zcpNz0C!m^*IHVIJ+n<)M1qF+LI$Rw`mPsHZ22~H>-UxICLH#}?sIJ``6Hd_hXjE>a zVX+UC|4HBH@+rbQk+~^zN@0vvF;81LC|2w|9shW2H#oAjSf0-dr+^V}6~_7t=Sl^y%31r`k1p--gEkuk-6>R?b?*E*!M+fqoVw>^;)*~Y8HuZGn09^0UMK!O5I z^vq2r#EnCHo`N753{hC*+sLA0ZaLL6kMT_luq8L04R{=ItG zne5!>IvI_WYIy0$@^qBf(?VZfXf`Io(c@*uf6j}zy^czL5}$rt2b=z zz$QLys<1b78d~v8WN&i1<%id_Sv(0WR5SeDn9qdYrih5)R(vO)_kGd(b3yRks2n1- z`wx<#>{0>CNgj3|Mk2<7&l{?V#XC-yR_dSfVBn}HNaUbQ=R2qOy&;55ZKeeF*DQPW zKrLIj$pXjpBo6JR{*gSKrT-J=}NAUT4Q+eaH^l1bQ=4Ijv7 zYo&dlR%a5u<}c63;+ai|z}6}i@@&uyWrnN?{lu^w=AVC>MDCT(h-a`3RUSkklvi^&q(C3Z;8kO`lS6iSBiUqG3EI%|7Zd{R#|QVzN*K5 z<+g3C4Tcc95Mq=)QeAa{OpeRYW?}WTt3-t?2?pr(x1$fiB8MpIlc4AkTU1rW=-+Ku z7z1rF4w|-7d3HcG;Cq<04l?}3QeIOo$IQei4s_QKM@+e`#g zy^8M}w7#d*afyz`z@m#s&O#So+AKsCY0aOnB4h^S@TF)n7 zxSr;xj&YH>r=3&O`1`>!E?%?Uwi;AI@;0R59|92r{Yf?M*xDh4Z3LvU?jf9&q^q)% zkl#eEP&)^_D}F?_o-jvf5dCj-Iu?Tjb`N^K!t@1Iekw~JO+uOq9-_a|M-V3hCB8qr zY?hTj1>Q~EjibC%HqNP|w#dHdVmR7OclUh~D-wcv%OuMp2#WA*DUm|3{xVCLc7I5E z4gB8hsgnbpY>#Ktq-V|intM#H%}%q=@Ol>*QisCB{VBiRvg|j zck|li^^JB#3ZzG}yg{SIZLMuaLhqO|y~!QfsSA!(lbptsPyB(wOJM@Al00CkvEh5! z-85~f`y!J4sF}qDjJp6XyX+s5&YTr@pJ`FeT#?<;E$cU=Qy8NH+{BJ(k*!AgE7v9n z=pdTPTzVjffC-JTl<|6xt~e3}eQ(8p2=?lV++IMr{jv7hY462U%m4m3W23Hjl~94& zgPCjpo@__%j#=}58t<1+l;qnxOChQ}08(rv>WA&A_K|d7e)w4{wz~(*AM88>RrKf1 zMz|+uV6xUwW5-3uYWDUhdO`#qoq^84e398h#^IY|kby(GE^Oho5xLC*f$+CTfF^^w zHN_sw1L6MaPwmbB=4>{y-4&Z2&x~_#eJ$!&!Pku|6d3y=W+9BZa79}pq|qObG-;6+ zyO4&fJJ^{}EJ(!A+LW7lkJLoMA&d2bo3~4N&43UMzZJQ1da_LgL%?-VQAb7O!qy|+ zR8kqw^Uz8Px(NlG?(_lJ<=~BS9@LE0F_zmQZEHy&yOSW=54y zpw75*TlN_=DwRUJ)%DJ-qq3C@B7K=awtN4`6@}N(icpXxp}V_og?#O{dm_N;5eGlx zUHTdBWG9vjE3w5}Xx}_4BRTJxb zXR9Y{^72AiiTKgVBSX*D*$DCwG4WnV_UK0IbMB7FU@b<%(y|?c{3VVJWj3sPWc5HR z>^W;@Oe_EkX+@EZr1+YaZ=w_0ek@L*!$(d)@u-4)%{UYCHp870)ANw3tuD7IXaPY+-U#sNTUI2CdV7bLT#C1%> zboA#zR44l5x}AFEYcTDWS^fHO6QZU0lvV2ZXL{7F(|m|ERGKWAoCx24%#CZhJIg0u z3+95G+I~CUs&;*=s<0Qeq&y`X&>Kr;r=$q9)r$Z$I8vqyU5#V~7&6AOtaXjPJPDbJnlB({g6=2AhCrvCzVtzyaTP$Y< zQ78I0PoNglhf)R0djrcCyq8w^>}&M^s}>rMB;nOtJQqftQn$1OFN#yYDaocl<1ll5 zvFRLZh))TWmUkE*__16d+JXTH0eM2A0 zEuSUwZR3b*1rhW}@ouYIL1OZicVfSB{Oe5Oa%vqGilCssDkA`U_YAO?QpLH%MFOl( zvlcAj!@<-Hl6kkC7FD#ZF(e`Z!t{4#6O5Q~3Mll4T->MEqY0%K$DHA50j%x2s?_F) zX6%Y>K)y-;y3H=TC_q~upWpX6h%5*1v&--{W^)rV_ z+XOxAsNyaeWeVImm=yHlAv#0HZ6H*R)V&J^DH~%{XlVZ;xg%vWjV4rm>bLK6J5aqS z=75ZLM&xG5?{hnuYb!ULVhDv|0BV6jMP;0iI%$XWA3yCtY-m=+nLa?khuJ+CM&fik zB|dqW@~QbQS=Zruu))bK<#-l-1se%^*CrvuUXo9CiU?=L5$lRbb{?&+*n$c9YSzhkg7!pmP=36V`5kHoWU%U1_L}AA~Y00A`ond2sq-XR*E$AkN z27Eg^72cW!-rIZwIJ5m|N&Qt`D6dsD#-j&5p{$4JNMC*5(ylqpP)L`?q4R^9yJXM7 z$nefy3z$_;7Ls8*;rlaQ3frncpHSg9X**fV@9)Pm**#~?YS56qRWX;*=&f99M$)ei z=m~3Dv#dxdA>i$^(QChaH4#Oa=mhk*n*$7co0JA`=aCvwJcHxe_hsGo#N|f{$CO}# zzlyYT!&RVE))f+SEvzKCT*BqxxxB+FRIy$WFNiV2W(AEEs?9>`${)z8e zInpc6a$3d;4gJqlEhSKGcpO;2Q@bb6N@rmKGcQCzTkcN42{gd2sA5f9$;}4c{>;Tj$Yj4j8MUwM0xIJcfj?z!&>5=f84L5>=vc zK2vEJgAmrrpCM(y2x6(4<6yunGYHWO`ShlreN>357{aUWZ^RWCw!m(n{{mn0wW2Qt z{9(3wT*Tz|Y6RRfVXnfL6Wk=F>Iy%%mxbZNv}`vLuSurlAQ73gg97snjiZ38uTFOJ zZ2v-MJY(nVNo1j%5G%5a7U;dH{*zwI7M}6SVC-mK|7{T-KgK)`q8aYiRy)EO5X<`5 zg8>K7>SNXlCSe@1$tEu&@Yg6n5z!9JJndE0Qd>b33hU}`)sZ@|VePL9u$h-zLY`1& z*n21I7L$RRh`s!d$|U%g8<&X29wAa0{S%Q@PRtsCXC`7BFlc;<6&OH6Ng^MiQ-G7{ zCm3%KdXiVr4²W{$Xs1Qk3wWEY}B}&vOZLln?kw}FvXfn!{hb6JJQw-8UTW0ci z#9e;KpjGH-%8I6l)kjIrx0rOt8O(4;Z^K;7Zkw;sMr#Bjh1#t9C)?qL{qQL2Lp_gV z8vcnH-aj%mQ$(&FRaZ-c9bJ)oK-0i&XjMs;vqrrrXKUI&jc4aY9f6<=KPCLu-Bk}= zGxP$zlhVTLsw1?#>ikNt+%%9i82+zpUc0HE%e1?x`uHxZtqTWz=}%C`)69oL#nf0T zVH%8ka%AHU$uP`5s>QFruHTt5>H7g4q%l77>LPkqI+0NCedHWIy04DB^>c>I6 zO$|OY`FT~d?Q}|}%#&5f)+J#*PQ=PEbp}!B{8=XXm92&+r)f1?^RZOn{Yu_-tlG`Tc7*(V?-OfO?|yv9o*7`YT*^Xc&4xLko54v-pNuDSY2LK^}KPi z?ZzHoUqWWncOcc5l<ghO0|^_|m%ML>!G+07hZQTQQzUg+$Ldj?%EZWKI&3+h7UWE*$C|twFMl}>2`7>^5Y-!T4IwKyB_dvP zh$Sq*zSP;qt!<9MTBV@dk&%P(nJ6JT&a&l4+$O1NJjppA8a+>bd!+@PxBRu5EH$`x z9w23aUAHaHO9mfDFFv)st`hpUPGh-uAKV`HgFN+vShMdShPzfb#J(6Bxm@Z{>4AS; zxbeBhpVR)1sShQTL0A_5vsug2`taJt6t~Xrv?`74L6mQ!D6FSvU#OQFRKSPOt94{vr&}d)ojba^$^x=aL0p#jun22r@pM393q> zRt){gzeUr7_aM-mlcP}kFoo=U*xR+;*PCBn(~&4cWBvVf6nyfYW)%6Xm~lA;(|%ar zvRk1W_CW@}W4~}`V}RUOU{1yxI-q@?@YkJFHi4-AFwOiYU(cud)tu(6Rx>>3CPr$4 z`Mc(tyi2@Cz$bH<6b(9im9_O2U!Nc>@(PjtLp3pLs%3hxGm}-Ys-cn135iQjZt)>! z?q6tZZU%T1xY#@&#k1c_-x9dci2auHcPfwjYx-fr z(zRhev&Jor+~HUlQli!^QFoWu{rYVZrAvBix-X4sgpMG@^=xin4aJq{!H-4?ezz! z#`l)%t>u@b`4yDhqxQYc@7PvJ0?cyhuM+EM2OtP^~A&YM>FKvRga^ED9=0APjk{Dhd9CyrS--iuJ-|7(+Ew!*zT{R}M+gkNjl z@!`$&z!jq7v)j4Kp5M7f0pAs|^M@rkZ;{y94+8vECzBqt?w4dMiaKzeHueDcN$J=^qxzVO>*8@Xs z>ml4nA{HxXnK*PZxC4H8SPEF3RyW*@E;IaolD5}oq=Fd&7dBXzh7R+es#8(qQ(Ai4 zM++VEQ^Vi+CD7MuX-7U>U>8gjdj@TxNiLUZ<@5vp{B{nZz=;8m#sUGZae^vvz`tUp z5Q_XiTkYPs8IDSFa(?l~+w`c!NMkk6vz_KbCRw!lGXA!9%Hg(WMeABe1J5bYm@7^^ zFCTe;0YsWRR@9@*WQ!4@Cx-?>xxJc!|Gd(nfr z^}e;j9((V>LwJty!215py82&KgAe#$VIii3N3&Jv`aT?o{q41qjJ@Eljg7`Rn zd`WbiPsJZ%GPyDR@3$``e+hZi-LDtoBh)Z7{SjM=bLOFMpj42ZPIf+MDkCm=l*!Ii zccvy%tZm|$xzhs3lEh#bc(s-&HNOV6+Q()%a&kxNNz;&6LO45G;RLV zUnv)}2mmJ68CySkxreVZ2zeN~me$;PM17$v~P>c zh}C-NH%@2^(Gx?)5;n1wS0mObjBV*bv1ASiK0rNI81}Zz94C!Zl5| zZgse1Jq58oF(X8>W&pior^3Y9NGT#JFav!qejZ?KH=SXGokOk<2U!|6~_M6@~}-=ay=8 zKl;x%WQ0dKCR!r?=RZ06N6&uete#!fGaOWA?&~y?h^c%iP_V5|PPMxljqT^vTLM7N z?aPE^D)T!j(+I#3rwQsy>I+`_qh7BBicjPD+Jdn?$zu$)(MC2(Iwef73Mmb=URe-C zFfi6gbFaU#;F#qy&reACXUTgrH9^t|xruY-^VSLY1qs;cf#@ABI@SKx`!>a_BAIp8 z+)CUGc?~=ETJXsiP(AT9m^O3Ac+PpZT_=XZg^~K4$npw+G2_NwhYp1b&jQ5hU0OjX z8?kJ*Q#ng8%2Q3#a-b*ztK3a)s%qoA#<>(YNq}bRz@<-1R891F$mUqA>b=vyVOA4? zpfu$>%(?B(2TDh5Q*P-)tb<&7jDOGi-J-jOYu|aUl;|2wh6AE_CCw{ zkrH*octOz{N*g#m=Wq~neLYgtOm#J|mcTt^1aW5M)4Uw45POhv47sn>qFV~QM2Eu! zBG1K?f|4vT9t%b4fOO8G15g-+p@UiW&8_)}x%su&&K?gCKwqlpFM>VU7y=f1wKk|k zb)(evhO9*t27^BtJE7bBwwSGKe*cpa6?Z>f{X3H%#j#5`P`5u}qM=g(GI;S`V7J%| zoSiw-S=0PLZ<~BInkB6gPib2~joU{cB%LhzS^8{j%s6pf-LD&Ut1L4gg#Ocr%~Zm@ zHYpqD_9O!YpXSo(+rtx6DqcDv*2VWeQro!v0^k|7cqiM|p^L1IH-!&Y!GXI=tkYgr z7LZno=RwYwPv#&O_am^7g4IxqmDje#P(iNw=UqnwZt;eYrdwqaD`$f;e>D2d@AA+S}=nK*q!}UjANG{ zH>JUHzNOUdJlR$=^Mqo;qSe{)6Si0a&|qLcm10o|JaX$Qb6WGd(IAL9?J4CNU>ThA#1hzY%#mpIj#;JWrL_O+fk-s9JN->t zvX+JXO}$S%%fLqNy<_S4!8QJw^jBHZ1N3arVjXqF>5YP;{&%lvn2%s(g|T%~7~QuWT4&J6^@t|G8>fQ~h@Fa4Y{Kh`pbV3e;TBeiLW zH0-G!7b)r6F&6Mx-GwMq^ORhKw`QJis2*A%+pdIGsL3hi&Rw8#HAK+qQx%LDRO}P7 z215Duyurkb1vBJdDT0s}drUDUgQ{4b3+pKvT;~td10E&vk^uX6FIkjiYMB1~ix)d* zIs*HGyf6f3*nSgdBc{cQ=5}mxn8bs5ur{vnQe{{OxrdsK>F(B?{rAZaupSFJ1eZ1p)$d(j z6j}_U1rix|j?&H3htaUL(-nV(EWp|k8);lr;u+`H*Oz+u))Q~QM2?QL{OpetQ9)(e zX%VD;Euf~944~}MclgMkBDatuH!+?IO#W?yWDu?eRpN{zX}D2H%SKOM^3iVZj#j#V zy~W6eAR~@;43$5P9Q8X#ok~CAIy5*6xGrG5-@w#NwE4q7HHfkJoR~DCL$wmx&Qjry z42k8>E%qGu9iq;TwC%?>yi~m2#U9LmmK=gdLuk; z`5Cm_w&=sB6=mp*A#Iyimb3GyH9STPC@-hDTzcFAgh|O^P>UjE(0;vhwTe{D?H)Gj_V8rrO zE+XG0h^WdmW~Fu)&lEMGCo#U?W;4SB?l5u$ztw0)@+1B=9ql08qQtRA_1+_}g;AE6 z^SF$yGe3tHkK;~dDd-6Doz0QZy)Q9~{G)q`7(J1jQz zPey1rg$J!c8UIYBf=`$R_C==Ut4lw(?6v)~x?NM`;gL`*lRs;^_+CkhH6eD2dJ9vI zxjbV=K+yN@J0|^XeTieysYf9t!^%2p|k zDbWJuyz__6(wD4Dci03C@qSkH{6pu&+Pd4lL;T4U8Mwj3o?`*2$9%p9No=^E^mfKr zr+0Y`5;=neb(@JYg8-FNy__Lqdh@z8+>wJZxR_~pgOS4JfF#yE0r&5f^MpPUyK~d0B-Kk4gINdi!+3w%Ta@Bf(P?oI-n94HV+)iAIIy(bCXb z1#I1nv)E#>+7Bx0+{hL5ZVW(P<&NBnM;sjF@oh>@O4Y_;J^B1|6uAQc{@rLg=%v+{(g(q2i(2kaMOxhu}bahuY>I2wV&?bv?hFIDo z5Jb^;2&zEu^~^989wB(X0gpAG`ct_0K5n<uPw9wFVkj7kq(hC-$Z=j zy)4BroQmY{?-~4LV<^hx14;l&c1$a#JSb>g1g@Tk1=OG!Cg}r0AaX_zht27MMU@-* z?0l6uQbqpZhy92E-QAK&e!qt{mz8N6{$T9SdpAur&axwWtBi-t064t8ph)koEJpUK z^#n0E3XvFO4LWPp&yo?I3SW)2^E}w%j6A`#=s3|dY}mRNa=(fFW$o0yN+!l!0NdMi zg<3tE_rBYc?fnbh&JG$Sd6@qH%rk0sS1P z;H(%OCyrsy-{o_K-EJ-w8da^@m1Xk%m^Sl)#b0JKRX`Q)P_pXw70aC3Dw_~{xA|%* zTaNelyQS{rseGmC-dQQrHL%6vhAoOH`muaw4`IF<a?-%AoyFn{leeijhE;J1Xj>Jb=am1}@Q1nBVA+lu1Q?A_yRoCBTSzS3amv`_Q}9!OscKQuEgAu1MU^u5hG5HpJI@5 zm}VN+MNXSv5ha%fZj5V{>N%4_ntY(P$f7&4EO}G%#1z=%6Drq3(xLd|QB#N#AAZz~ zNb1ck*HZeSQEv}e%CPiodT3iUzln1Lp`ZJKTmpAJUG~rS@_N5`m^Rd$j19U!y`k9kIc0NUguK+2` zYBw8{^G3qLo4-&rje1Z>ZIR1#1MiD}y@aWzJLZfO{+S$dwM7^mYU{{#W63^aZd4v* zl4vipkSkMba~CQ0<$5*@Dh%d?7f#Gh{X_ClE%qwPz725sQ2N~eFInStXkf75EA6>aM3m@3M~8V^E{e)Ht_ zGKy-zGmytZ*v6W_mE-O`Z~4_#U7`SVSb>Rkg`LxA$t2&e9dVG@D13CA44o;;a|tF~ zrkf~r-$+I8e4vdKo(X%GpWJl+*x zWHLR1u7~F0vX@%jxtd#UU>US>$kMxi8Bl#t$^8_q9O3W}r05l^%z}|$?*b&NLyycp z_Lh2NgjOumF`aUI03%*eHUL^UMS`PTDu@*%7|NOD5SahkQclDg?Vow3WW_s-@p8;u zwT6`5yTfF^HiCYtH4{}Eujnfx{rMVaSqFt29R=RpI}6XAk~|~itucMU%*bUT|%5Wfv-T-wzc|? z!Y(%Vh?(8Gsqtxj;oigDv z_g(S%m6_elK1cX?4YI)r+#`Pl1N?bol5Otp`f?u7$yh~R%FqE`Rr35ZxeAeX>qh8D z5|AI!mU&|VgJXYU#fk%CC6Ax1tFC0|p+@^r+S7G#A%hdwuIK^`0b}33qEdvSMpf(0 z@NfGNb~WaesP(Japrh*?!t$n-OF)*NYNXXl@I@@}M;raLflg#?*ru9&vm7+lkEPNb zmy#Okg!D;O52d}|5?DP%&^p>>(52YXOImxd`&Qspa_*KFJ5WbHy5`7A0a zC1W6myH6nkf_v7A zL{GIN8DBZNbeBAPf5-0QNRJ!vkOKQ^)7F#LtI|O;di5?8Nl zsQ-1f4^;3JU3wHDaQfT%5EkF*OTFBwt1*qLr3raVn|G%kUOV3|J6*>qSQ1qx0b8p#x?_fl- zz4BhBvV3zqQA%vWeacZfdQ81ULzzAiZ_{+7=jR9Wfr?jDRi>t_@cq@-bLh5Kvl`MY z<5@6sqKnv+AJgu6(P20*;{TIE6!A|u6VLdrl(R$H^C-8^)duWWp5)pwB$KoKIJez9 zqVyp1mA4ptG?73O=EcYwN}(3hYgw-~0ehM+#Hi~k^mEDgD!$#|%`!Jv*2wBh139q2 z^`GWLmhfh9(-$L$N8T^cJ|q}Db03mH`rBi#S0tXvB6Szd(6rzdQ@#`t7Jm_IwtW8^ z3xS`B()j57I5sINSh$pJmHR$)p-+8aCm(J|H{oh1>Z;1aLHhB!hj@uTS>()!jSGW2iewl95f!Asx z6=%D8fypFeyS3j<(#)P*q(}$tb&#AQEIX9mW-@F~{H~jbLc{X?T1ByNJ3&coS6M+xlgz5O$5hV+W+l z;($nn^Ex_VX*yMZ1Ol2>YE&%vL<^d3qk!2 zqG;7cRR86B+zRoIhMGK^=KzTlz*1;IRG(4jlJ%EXBU*TpDoPUK1%rusdoS)Y40D|M zIg9qvp(P`a*~fD*fA;Y9YM*J2xp$OotAXPC!a$@*9>CkwQl#jMN29)4goMVNT1F%8 zA1poV($pMaiy-5#o}{Mjbw74cVg1{6{g%LWf3^}af(GqXXj@*S%(3O;6OM#y+;lgg z{^?`9w4G%2!k1?Q8zcfou>beey{_4Z-~N5MUc5{7 z>!ovzL@(0XYcc-DO^TJ{jC#kzPolleigrgbtpkI`@52Ew0l<6JM@(vL4q#XtTcvcr z)We(`lUkw|UdZ@uWnjU~ka}lOWcjdLR6pNvSj;sbv}Ms}Z_uuqok2^QMpcTwd{0}u zqU-`x;x#S7+`q70`t6dY4L!pMHCjVJ>B~pt*0; zJ+BUbDR>~SbokMk#F!g+mw>|v_D|0R)XW9{Uo6^ZJ@e}wm*-~ONiA~Ik;6-xQK~1? zb3#v*XR?d347UX+%nb>}imm9r#`w>Emrol|wC5=6Y#=yv?|L>xqQ%62ZZRNb+Wk6AuqmA9!cZo8bxgPJ*avC|QA=Xm4y zt&OT7VNB#s9f0-oErvhQfz66{A*<%kJ4pVForLp<^g{I3Y{Y!SH#wh~cUnx*%{T9= z#EBltF>u?AkNFn7i1Q6?`G^;bro4kkDJ_yMYPzD8;AKMZ9}}$b72RP74%nr6RzgEClN&)0@2;4kbrqWw1!2l862zQ3ZhHBG`Chgca9MdkCy zsS#~9zC2u9Fa>^!QT2u<@)~Z(eSQbCi?vWuax+PIXF?I9em4{yXH!hU#y<{zsM}@$ zOUJ|BihV^3=cbxEngDcRUg@~;nRnK3Jw3FQc9-y>*qpC~4IG;SFR{m0`ZC*GmHX~P zXr?>UyhAfxP&J-pby!^-ZY@e%gCu;HdG>LWpBrZLB|&9%aZyB?M{JLFAeN(u@7ip3 zIm~yX;b{VQ{oe5&FhxcB(S0IwR_|u&4(Mh^fe(L*(=+JXT#mZCl~Ty#4Z3FTDYJxY z%y<~-TvV18cPG45xD-I>vT7{gPP z-9Nj!CufTvPAq-EC#SKPJ!6iRG_hZXWHHEiGYFA%qVauT`Ahmkv(b10ArbUb5Umex zXu~8}ejLl7_NuY)lb<%snFsg8v@8}Sa zdTtJ}JFS7($z4e;?ME;8b_#FIvIh)@yKPg}dA}_dSrF;b)IB02FjIhB7~7WcwC}>> z@EzV?p1i`Jm+>H?W?k)6GF5a5v20I%!ao3h?KaCGM%@mJsR2?qaj@~EF22d5H&bbI zvNcIK03q4@&h4AuPQTLBVs()A~}%hH>0Kb_Dq&Le!s>t0Ii-MiWcL? zZLM0)6AxB8iz-xj{zA3GqRYIqe}c5heMOWeE;V$hE5JJccDp<>Q{pis#>FMWRv^^K z!_YPP0(xhM9SL(w!Dck)@>HJiA9qZYI?93G|58l`FofJP(FFn~ZAj|##I3qFq*GhK zBYjLW)>kv*UeT-zfQdEK7ZK`l=qS5_4s;TrA@ZbMa@|tAxvJxH)$OwOAHVd=I%RFhl8PYrn)-Q3aW?q)N`O zRQe)C90rQumm+Td700xe*g4>F$kjjzRUVV&EC6W8F^7MY>37NVVG-`Nr>Idr*2aqLc!l0G z<72qN9=1i&`2yf1hO7JHk;M0wFIRbBv^YY^c*S3jB2#gf6@AZI;27)j0FMxxKzrAbpxf*?-0M zw#%C?Q!d&0kOVM9TlQLuWG2B-o>`=|h`H9)E#5kQ;wG|O#LBkGiE*HzX@}I}DpMS* zP(7FS>C2+SIb8vXm@CluqL3}SnJ;{+%M@SW3f5hMM9DH+pV-_Xh1;bb4e^9x!7L1J z6pEhB34B$_GWGRV*hbQ>YG32RvBZW0#;EwNd+*3qY2BLqgSG<5pag>hVrW^=rotdh zT)jsslR5QUSAMo}%%j}4vn^u3yFn}3F{0;z}PsRXTvjnw6-D{wF3?E~eO$lYhqpex6r_hM>ehPBF-yH{*w4+|c)wqVtq0c%nez*)kH(aT5hg+ z#K|SC3^Eu@b7DmTG17ikR=La9qn-0Y_fY~BN+p(z@XmcAI9qa2{qvTWKaCd+)Ei%MES&v2H-sS^M0Qi?)qQ3CV{f}IYZL6}z(P!UCXxi2H zC%4+hKhhFf8lzVL0BD}XAK?Ex4to6n$nqWj0FapB1^xgqMpph8mH>u-|M#u`0U4Dg Ae*gdg literal 0 HcmV?d00001 diff --git a/static-site/publications/agentic-language-development/pages/p02.webp b/static-site/publications/agentic-language-development/pages/p02.webp new file mode 100644 index 0000000000000000000000000000000000000000..81de8d4466158e9ad5be02f09715345edf8e68fc GIT binary patch literal 103462 zcmd?QV_;OIb=>oFEAp=&P8pqPpT2B9!0nXF9>NfoY1tk|6jEnUW+5NeYNb63W8- z;Giw--t7@TVG>%1sXhBB1TD|s_Xr_r-1OMJYWhIO{Hbn%7G%-AO%L^qUZOV`vwfFe z->$m%S!;J&-p7XawKmpJTR?jH5^2XF!e-Z24y_pLmE5rI74 z+|RpDV*uuJECBHFISx<(c)nx4BEIXz2mF2w zGVc1G0%8E#fW7xWJ_FFZ*#MXT{r8gBD*)m7A6q^kKb1ZtJ`ta4?f{>@v%bH#JS79( zJ`vxqVz=4}0RpFh#TSW>M!*|D;S;I*R)1wb{*5<5z&t_X{8sWKM~bjvl=RV^%x9V` z_W%34`tPhUZ}XQU|3es3!dhKP-c$eAu#i`s`2MFLN>S>4{>QkkWZn~fAKgj6vyD%C zt!7hnj9`5tnW@+CNO?-#yX&nh2kQJxYy=Jdx@Omr@rA$JkYB974s*Zz-#jMmHo`ay%*KtLH3dPA9K=vLD0^4>iIuBiwA0@O$TE#pOV#D z4*gBY`*EkQr{3iNR?wbkA4UV4?w^P!DTg)lZ|V@^Z#|!ov&atT;^18JmCpss`~Pbw zPeUk2bczG?c)xXz5`6z#&|MqX_-m1+ix_Ew`+nY3XMq%a`aIoT5Q`dxGn#(YAG6j zI{(KtOs;A~?a~58PI#;jHQVyKD~;nCeL_`icD_k(0Rv$Y5OSp+MTmE&RLYN7i|vuS zkftu&n&eU0qU_18b#zENx%?{JzoagiO6vmHKsaI!dUsg}X`Qmvk!OMU$rOu~+I759RSPv8-BxqJI{C!|(wE$1*;st_AFXSML_vTfNh@wmlJyWFX&nnUCTwHAuhl3*EDK)? zgY$_YOz{Itro5yK(X!_#6z|WA4*jLMw`B-*EKs-y2br|g-IT_YLVN?|>c zTfi;|UPcxU*C-lI4%w5y*v>8?JLSK|xD^uI2^Dm01~bx{3&oh;;99-YXEz_}gS**L zJB7|u&>lMLv7kL|pF3+&dLDve$!MC~$<&na*(vQt3Vrt{BK0ysm?o`%_2-ab)tcwe zi9LxvIF7F+Qq7@3RuhHu39ZMER(pVoENRG|D#Br)mx9k<8%SoOZ{$|wUh+mbokfdw zB0tNxJkiB#e6{mVb;MT@+iyK_EVJrYPp&iAkJ-<0 zC_E{Per6}-V3y61!;PY4Xy62Y4-w5LlUo)Bb>8AG^>OK zDO~8>@{s_PjZC@|pevo-PRYX)J_=+(cWeoAlcRDCY9H_X(&CU1+)xr_l~QR?a(~j@ zRp$!77a0MY2sogu;-l3=bKJ6p6`QmRT#l?nOl_sDDk=zt$iwb-EyK_E-fHXUXXwp8 zBBrw+$WMRN)W*4#xmZvPn{qo%!UA_4QQQOa{J}C`FrYWNoTdHaVMbg#$Ap@cc)7)T z3eMzaohTg~f=O}f;l|~RBn|+&gb2EE0sL*%QjOolX~Q6S9@6CDV#RPk++KWDyR4(E zL>Q=UMd98%6pjg8D+$~k%D#!ZW^)LCcttK{AZPzmQOs-kHU*Q)Kh@QOUVh{cRx}9( zUaE9rklFj_W@%fe3$kUout_6e)vkt``k__p4jTkZ(!>h|scpCUktZzT3*8_PP+OQ6 z5CoiwJ}|)PfL8_S+EG8z-&U%)ByPbtV*sNqv;woW>;SvSdR6K z`2@iZyuXn-MZ;EhJK;Z}^$!sC1_GB^_~#1$3zbE%V+7V0UURK_R(#zE=jbQ(%O1;q z^vE9!DEysB((SCg{w$?P+u&f$7Iw-=ANUvi>p!mrinTMkl}(m6kuI%16q^3VK=MyV zMvSALmZU=D0oxP4qyJrd#gW>XXQwOu?8C!d{`oC4XXalDkodRBs^C#ukUM|t*dLAx ze`EjMU%H%(7}1$#mTz4Il9#)+w%f}8LPrO!5NZXBh9cAuzZMa$LVy2^zp59BOFfER zQ{0D5mCJf`@5wNjZS&WE?)$Iu(5a(rUDjrz<{#($u^n;^e50{$B=Wzn^ykr!OHg1T z50DB5{We_mdkS+DTQVOArSqh(n7?bKXvZ&G&!t8V&747bkDsa@MD~AhvDgb0Gu&HS zanAUOa25J{tf!ZI5fs{K@v`aIs zyN81cT97*6^0Ubzz%hRK6#o2uyf`q6q89iFX74tz+Dcm>NB+dram0j%o%?QMmlB`; zVBw@l3FT%KWzKoqKpp;FO9p-L*IYZ_aPhR$33XvpGEv?=$K?3RCceBR_(E8QWw+3-zA+c9vt? z3qU>rO%t68VDk*7UBjIfXrAE(yZBa1YYlEHPkN66Pi2UO@tuc)G!>gNi;a;OanOTb zw=BEw1EEoJiCm>Hxq&}-dFZ5-&zIo=vi2AzEhW*`M4*;qoIfOg0eqgArozCqw-YNb zY}V$_TJKQ%vkR637H6*B<-p3}E2)Id>(=gQ>MzIr9SY5uPk=NdSX6ni_cC4#6W#QWN~F6pVJvq$a{KWvua zH1(eE3$YG(=vfio&kPIG(lY3WxE^|!q-RN+dMuE>4d$SC(uq?^x>*%(8Z60EPw}}& zxh=OFFA7{V(ULcm9DK1ukCcgL84$Gm19g1OK}2FeYN?Yr;BgJUph7T9 z32_c%c7t`OO*~wv3YsCDM*+R94e&^=BC^#QLNIzAQv}MAYWZ+G$B!~@q0R+Fyr;H~ zBxRY>Y}6Y_Dl+N&ZE)8Jr*8`I&zxnK6^?jnCA$#19VVNQ@8eIZ)>85lBKWV`J%aD& z@w_3#`Y>Zw#zipvY7&60SY{MB3$z5E&i{fj!Za`H(pB z1A6VR077_e_Tg^2-GL+H*`@c=8cfubQos*nY@2B z!{p$~F(E4GC`iEFqJQfS*o(F=|5GFKBtC#W^2fdWZv{Eo1wj5kcHh=A!=nGGO*d>I z*jKX%;KW0O(>Phm=koe*s{Eh((gd0^1~;I8VDA4#!GG)id6M_V%C&qvbMiPfZ>?E0 zzNpO?;x?rQl!c41x&VeH^6}0mxG3Big|mIr_ZU9=EhEx4{9NS-iI6f7r```AWe~1;B zM~kKnRH{uFTBztf{M%_?w>Fj+W+#~(?A{5&2sfW_bZkp{HT&S()=oPB)WoytW9&Tt z!ioQ_e}uAPm%aSEko@y$&DKGD`xk2bf0PqJPLd!l2DGt;Y4Kx!<4F?O*^(-!+e5zT zLrd$h$m2uDruOWZ>Sm3S*R4+B2Wb&U7b8v9ADH=dq#jQy6NHa4BnWs%jFt5%W}pa%cN4acxFEpY|gYB{Kj*9Km)`>}S-O63&I!iRvYJ$m0W4GZXN| z?YSM!i%ZXqLDq<;rLSp*ka%)dOMo@dz=34Hu)w59nljKxeYZ+}pDdoNRYj{CJzjZ^ zWu!R1&7F)sKT%z;(`3XkG$yQQ@?-dPQ|ei%0g4X{J&jf#26dAX_>vKEtFR;+R}onq z?`*w$oS1^#vPXoRU(YiXvu%Bbxw1;@t&m&*R1|&`iK#GYUC7o&pajTBj#B&&cu8!5 z4yg9&F)YbA$HiiJNbsYMqI4O6d#xGQtVX7n26<0eXYEjiV|TjV&r_a@Y-!1(;6|`0PVa;vgmDb) zH>E?LdKr)9)#>vG@r}K?817avW_5`7FHN@ILN!n}`$5s0PXtqs=2s+tAWLmd7a?rK z+qpK}mj@4Qup2_^;`3cSUbU)*BeEL`8NXi$kFH&(Cx4F5;Fn-=iVbvps zLOaQzvKij3YqSEtRYlx&%6^YnuUO!URIzndYFSeeeI*i)ku+?yawCEqx@{SUeN8on zDPtspp}YIO&~tI8cDUKmG|A&`TEBPWQm6t}l^nlxdWTzV;4rXCH`Yy)I$KfMq=399 zvyl@2sswkEcN3AW^!j4?rZWs@!CuAya-e?{w&5F7vRk{e7#WvKGCx|sKcv~9ar|T0 z*Zu0gS8hrujemoQ8=w~U!{JTEy>u)g7{U4$@FKlR0saFgKRpa0rI6SIP1#Mzx}<9a zu61e5&FBpNik&K_&~{$eDqJWHtZpf+KldA(KAo9FEChhbrmQz^>hASVjyV42we#7@ zIu!@L%UGl{R+tFc z{9W~Y!1ek6{+yP3V)rOG%vOy~}ZpsZ3Lq{xJx24~NHaU?9yG7|N<%~&An)*dx7 z;E>PpF8&Yc&=_qJt#&5srbT=i&JohyRQ*a2#O5mu7JbH3<@#aIGl}!pQHl!#hItUZ z_(E4h{9lDqa#=O}tn;jpO+Glws4(!$sEHEmo(neG;if7=y5D^e%lou9*lrJj; zYZ82`Ki*MW@9^N^5J%CW4pi2dbj^wi&Hy4G!9VE`25(IEh*j*Fd^cf2`~@{X`xwW6 z2Go7Oz@V~+eFtq9t|Wz7P^N8o=G_vp-c}x4b3)-Qo2yDNK=x(H*@EXhyp&>%8dNgB zyiC%W+b@Lz9=t2rh+MgpkGzN^hkB2y&E`uECP-6+t1-MUWc`+w2%q-1R`Y}^H<{f( zXz+-35!v#fT62picDoRxj zc98h{Nh1?Cegk}EFaiqowTjgZt4ZCe=X{?6%>|)BTB{;J!2-S5|B7#c=!_o!0}7_- zNB}SO!k4-%mzyWn_B!-ZG=RxCxys>ZMfzfTx0dq5T|5izgB!fFn2}jGB70((aHWRUJ+H#lev%`OA2N8K3eI;Dm4jR|a8Ly{U zT|&Tc%a*@-&Ine$OlMBh~TsvcKnD<<5V|WP*yC_eWf>l5Bwc=0tECDI+tx`;={vit* z8;y`K1ISWi5%Jga7aEG57bl0#gR!ubz(IR7 z0#rznnzoI-DoOzR`|87CNX5D#oJ*k|5tNwg6v@wpUj(1zR8f;gsrsO8Ef*9F2{wRb z;J~34=te|hM#Zb?AWx1A^@%l&QIJ|OB27F;g%+YS!uIhj!ekw69`@Pi%HwiA2l zFU}*YGJR>Yh98Ny{SNA^d%%o7b(hUn2YZ_IWz{7Xd>y~E!KmG72R}vIQ&6{77E6!0 z>rGr;cr4ea-bvMT9>o5vwBCb&y=n1xiCmCE(xt%lWW|;|t?_@A`ice%s!{qgtw?C> z*8a5Jl(=OnTv>*?o3~N$5HKmr{AdGQbxB~egE-;}2uCX2vl|E9^LUSyr@5e05wEL_ z_i#_q9U$X~TmtVvQD^DJ3?CYQjXD7@#a0yY666;6it6r9Wd?M3fd0~F4MkX|=6BGC zXdUKuI$Y6qxi}djC0Kw^4g~Rm{9A%Y-Ib~L34hHBtkiq*c;yuK4k317eT7dw^k&8{ z`GgBi3CelR{%)+_&-%=i>xK7x5lu#yiT99hn)y@FeBxbII0vr!-Y-q`yIVFBg-&3d zLRnj)KV^ty?TyU+B1*TX&Tg2xm1=~PO3Q*XBniG-e0+9p$FE%@(h5ZPd^Zj`rV>15f*9@-s&SAS-s0xFgCtkZVPEqxd6W-3qI$EAhEze{NcguQaGembIoJN^Hq>h~Z7^7y;0^FMl^(ZA3h ze<`|4lmqv`!oDafmOQFRfJ`zw8g!LL7bsMN%I+;hkkEqhSjyS`~-G7l~Fs+5}rvpQP&o0 z6nkAnk~v3c9a*j}gbVa%qP95P;L3YvStmBkE9`e^a0<`Jj{ie3qlN+v$ttWk$`n)P zZ_BG})198u6uo)J<}H7zwn15$M=lNtkO?3g$-3;B_Q0V}Xk*$laU>i6^qG@EretiS z%WfXN=^om(*aFW0!~0U8U}k9RE1}}3W(P%XOgRToAv2lNC&@Ax>Az(J1Z-y5Fnj3#`4Lun-k2})Q)7^6CE4QT^#RlJ2 zaV}xb^%oCFRhvL3nnHNP?wcdYa%dKSVc`^}9?wmoxOZoX_C+=wa)E`F&sOn^4*8pW zQ1+Na8}vvG9OUUQ;7ud+TK<4pKc0IK^c>f)(r;i4@LxN~{6c5aW3=;}tq8tJ1*NwQ zD?s-^ELEAi7*g_(d;d@n?X%P;=W1zUJ!Jh$=Ok$+RXZ`_8bm`-!bgd<0YiyHGFKMm zB2V!xQWK!Q8TLtPA$=<$Gh;Uq8D3OH?|HD9K9|rXH|381%V)h6y4r3_1dc}DUM7}_ zJ>j|ib&cq<8MS91%!1M7ET**Pz3i-XT97#2V`u%?NDzpD^{hPy1SI)Mjq;f}$I zW+_B-Ie12m(Zz>0gf3HcqHtr57X12f294+28ni(YcRx6cIyJU=kAWQk?oF~*96Q^rvDN;Vh#?o+n#=6KC(h;@R>K*bl$VCEVrKk_vv7P zajE9%JV4mMm1HfoV`om`WZptND5E8gv`-xzE7bx-!RcQ4|7l^#KzMvz_SK{;$(FC( z6UjEG$r|B$hTtabfNWMO<;GFZF`;>KG=G+>?5 zO>gAPP%ee9r^r2`K51*^mg!}BtRs4e@-6ywpV2}SiWa?dT>U^gwB@GE7J2FMjOiys zL`pBwchI|;2d`I}7fOf-zpTu?OuJu0Xkebb=Vk)b$wzAdw0Xy{k^Ws6#iIl&{sdKf zNQE41P$r|Lu2+M1IH5Dn@N^+vh^_dNX?pV=v155Wc@7p!EJ*O-`?>~PDJ*C%%KLo`j#+~&V*mVWb8^A25kC( zIq6_Bs1g0bqH|5C6d|8Q!945fAQ_@1{2yml!mzC_mvGx}wvb(-;))H65SzNnSmK?f zm`I{nu2?F_%3rkm16Rvt3y;RUKnk@Ej?(edzl^EM+Zo1y{5i;xKS6#gf@koXwwh@R zut8#0hmhmr7ae^`#zoimio;7kC|O}KF>$LwiThrd8p_Ui@>M;)mcPKJ0CD}P!Bg_> zP@Z{W;Pd&-+AhBm+iZZ@uz=I@;9vKzse^%Ji}!WbU)zNldE@X*g4~k?NT*46kvO4{ z(tG1)MvN_)RNL^SKc1ezJ&5QJZ$XV+%Y}G_N|sIG0GyYWVCo95jbCG!b7b0!!qdTo zj;?BnWsKUrO<=Kb8%7jp$xoVQ-P{&Qoq;^j;*wKV*_H)bLdg|G%5Q-(38<$QJ91{% zcHA`k!Guo8j<-0n+=$k7O0K!8?2I##D=WBl2ii}g_nmfJ;>QdjIzqu-3x*Qd`B^#H z=$?9FVP(a%CEF*hDBaU=Fm9$enAAJK))+a6T0uUp#vk$2Ssla8B0po(B2RxwcPue} z+`OT_&w<{0By;M;PKK-`M#4ngSuV1^g_-ydK%!&ZeHo!_60`rJ+5h~SqSR%2$UJjo z+cVLg=O@mUOt#nD$d+n%yfwb5H(Ol~DJ2#RP@ z#!t~-vg7Ac^be6MHNGIDE~?5e=4W0h;X3yYg}n@QHI(*KqWeO5N}INn19glqd?_9W+t3%gtGn|XGsXHt^Cal|D%WnP*Xtnx#e4wveKR-n14x1;6Ir` zrZEvhjj%FIHg=>LeubFCs}ysz@Q1iSXrWj!#YHsl+)~2`1^amn!lmM{>x0zGD(8_V zKTceI3rcba$LehP!w_Jb+&r3R#FBW6r)*U9apAaRH_G9HMMom)DF(aPw>ntaH)Hmw zcFw%sov2Tc6T`sg1R~fbnDOnwOdJf>57!Ld7jax@O!AC!E_AP6*gB-6Rul4kH7NC! z5SQ)PZ3h@pTIyEY>D5VOZLWI=pMD~ZCHa)`if`$V0NfQO^SC_aCC($|#(Qv{<3OpQmR^h+W;wA*B*TTWzKQYiSi13e*HACqp6>yjk2m4!bY5<_uK2k^ zSAu=6owmm&f~;&+_^)}31fcmAO+3*p;xTNNOQ=P8S-#4