Anthropic circulates a draft federal AI safety bill that could block a model's release over catastrophic risk

Anthropic has been sharing a draft AI safety bill that it hopes Congress will consider, according to Politico, as reported by TechCrunch. The draft describes cases in which a model's distribution could be blocked because of a "foreseeable, imminent risk" that more than 50 people could die or that property damage could exceed $1 billion. The risk would have to come from the creation or release of a biological weapon, from a model evading its developer's controls, or from a cyberattack. The thresholds are similar to those in California's Senate Bill 53, which recently became law and uses them to define catastrophic risk. Under Anthropic's proposal, independent evaluators registered with the U.S. Commerce Department would assess a model's potential for disaster. The labs themselves could pay those evaluators, as long as payment does not depend on the results. The draft matters because a leading developer is proposing specific federal rules for when a frontier model could be stopped before release, and who would make that call. AI labs and lawmakers are still debating how far such rules should go, and Congress has not taken up the proposal.