Anthropic pushes US bill to block high-risk AI models
What's new: The draft would allow court-ordered suspensions if a model could cause over $1 billion in damage or kill or seriously injure more than 50 people.
Anthropic has circulated a draft federal bill that would create a legal path for the US government to stop the release or use of the most dangerous AI models. The proposal, first reported by Politico, sets a high threshold: officials could seek to block a model only if it posed a foreseeable, imminent risk of more than $1 billion in damage or of killing or seriously injuring more than 50 people. The covered risks include creating a biological weapon, evading developer controls or enabling a cyberattack. The draft also would require frontier AI developers to test models, disclose findings and submit to independent evaluations registered with the Commerce Department. Congress has not yet passed a federal AI law.