Anthropic 提出的人工智能安全法案可能会阻止威胁 50+ 人生命的模型

anthropic has proposed a draft bill on ai safety that would enable the government to halt the release of an ai model if it poses a catastrophic risk, defined as a foreseeable, imminent risk of death to over 50 people or $1b in property damage. the bill suggests using independent evaluators, registered with the u.s. commerce department, to assess model risks. these evaluators could be paid by ai labs, but not contingent on evaluation results. if evaluators determine an imminent risk, they must contact the commerce department within 72 hours, which would then request a court order to halt the model's release. this proposal is similar to a recently passed california law and goes further than some industry regulatory proposals, such as nvidia's ceo jensen huang's argument that no new laws are needed.