Leaked Anthropic draft AI bill is weaker than FRONTIER Act, critics say
agstrait · x · 2026-10-20
Politico published a leaked federal AI bill draft Anthropic has been circulating to lawmakers. Transformer reports it is far weaker than the FRONTIER Act: third-party evaluators face no standards or licensing requirements and can self-certify, which LawAI's Charlie Bullock says will drive a race to the least bothersome audits. Blocking a dangerous model would require Commerce to petition a federal court with a detailed case, considering harms to the developer but not the public risk of inaction — Encode's Nathan Calvin said the bill made him 'viscerally angry.' Critics argue it undermines Anthropic's safety-first branding and could foreclose stronger future regulation.
Related event: Leaked Anthropic Federal AI Bill Draft Slammed as Weaker Than FRONTIER Act(2 posts)→
More from Companies & People
- DeepSeek's Liang Wenfeng will never sell: ex-quant founder doing it 'for the love of the game' — teortaxesTex · 2026-10-20
- OpenAI DevDay Exchange attendees get 30 days of Pro plan for free — gabrielchua · 2026-10-20
- Runway CEO says AI could 'absolutely' replace human actors: 'It's already happening' — c_valenzuelab · 2026-10-20
- Before leading Anthropic's first round, Jaan Tallinn asked Dario to get rival VC offers — JacquesThibs · 2026-10-20
- OpenAI's naming mess: ChatGPT Chat, Work, Codex and four Pro/Ultra tiers defy explanation — cedric_chee · 2026-10-20
- Nando de Freitas: using GPUs for deep learning since 2009 with Hinton, LeCun, Bengio — NandoDF · 2026-10-20