Anthropic drafts US AI bill with judge-approved shutdown power
What's new: The proposal would let Commerce seek a court order to halt a model tied to mass-casualty, bioweapon or $1 billion damage risks.
Anthropic is circulating a draft bill on Capitol Hill that would create a federal process to stop the release or use of the most powerful AI models when they pose extreme safety or security risks. Under the proposal, the Commerce Department could ask a federal judge in Washington to suspend a model if evaluators find an imminent risk tied to a biological weapon, a cyberattack or a model escaping its developer’s controls. The draft sets high thresholds: more than $1 billion in damages or more than 50 deaths or serious injuries. It also would require frontier developers to test top models, disclose findings publicly, submit to independent evaluations and maintain security programs. Congress has yet to pass a federal AI law.