Four federal agencies sued to reveal secret rules used for AI safety testing
Ars Technica AI · rss · 2026-09-03
Nonpartisan nonprofit Protect Democracy has sued four federal agencies, demanding disclosure of the secret framework the Trump administration uses to review frontier AI models before release.
- The group says "almost no details" have been released to the public or Congress, apart from vague references to "trusted partners";
- It remains unclear what the review process looks like, which companies help construct the framework, and what legal authority officials rely on;
- Neither the identities of those entities nor their selection criteria have been made public.
More from Safety
- Anthropic launches browser tool to detect Claude-made files via C2PA content credentials — btibor91 · 2026-09-03
- OpenAI's CoT monitorability hit: paper authors double down on 'fragile' AI safety window — GaryMarcus · 2026-09-03
- CoT monitorability not abandoned yet, but new techniques risk a race to the bottom — DavidSKrueger · 2026-09-03
- Japan joins US Genesis Mission as first international partner with $1B joint AI-for-science bet — mkratsios47 · 2026-09-03
- Microsoft tracks malware campaign using fake software-download sites to deliver multistage payloads — yuridiogenes · 2026-09-03
- OpenAI's Report Reveals Its Model Crossed Its Own "Critical" Cyber-Risk Line — eyishazyer · 2026-09-03