Anthropic pitches AI safety bill that would let only a judge block a model's release
According to Politico, Anthropic has been circulating a draft AI safety bill to members of Congress that would allow a model's release to be halted only in scenarios of 'imminent' risk, where more than 50 people could die or over a billion dollars in property damage could result from a biological weapon, a model evading its developer's controls, or a cyberattack, thresholds similar to California's recently passed Senate Bill 53. The bill envisions independent evaluators registered with the U.S. Commerce Department, who could be paid by the AI labs themselves as long as their payment isn't tied to the outcome of their evaluation. If an evaluator identifies such a risk, they would have 72 hours to notify the Commerce Department, which would then have to ask a Washington, D.C. judge to order a halt, since the bill explicitly bars any federal agency from blocking a model's release on its own authority. The proposal stands out for explicitly pushing for legislation, while the AI industry remains broadly opposed to binding government oversight and prefers voluntary systems of embedded evaluators, unlike Nvidia, whose CEO has argued the industry needs no new laws at all. For a business tracking where AI regulation is heading, the proposal shows that even companies warning of serious risks from their own products want the final word to rest with a judge rather than a regulator, which affects how quickly a dangerous model could actually be stopped in practice.
Sources
This summary was written automatically with AI from the public sources listed above and checked by an independent AI reviewer before publication. For full details, read the original source.