Anthropic has privately proposed federal AI safety legislation that would limit when the U.S. government can block advanced AI models and override certain state regulations, according to POLITICO. * Intervention limits: The government could block a model's release or use only for foreseeable, imminent risks exceeding $1 billion in damages or threatening death or serious injury to more than 50 people, involving biological weapons, cyberattacks or models escaping developer controls. * Federal authority: Only the Commerce Department could seek a court-ordered suspension. Other federal agencies would be prohibited from independently halting development, deployment or internal use. * Independent oversight: Registered evaluators would assess frontier models and report qualifying imminent risks within 72 hours. Labs could pay evaluators, provided compensation isn't tied to results. * Safety requirements: Developers would need safety frameworks, transparency reports, whistleblower protections and critical-incident reporting. * State regulations: The proposal would override certain state AI disclosure, evaluation and reporting laws, despite Anthropic's previous support for state-level safety legislation. * Background: The proposal follows U.S. export restrictions that shut down two advanced Anthropic models. Congress remains divided over federal oversight powers and state-law protections. 16-OCT-2026 07:33 PM ET