Anthropic Proposes AI Safety Bill to Regulate Model Releases Based on Catastrophic Risk

Anthropic has drafted a bill aimed at regulating AI safety, proposing that the release of AI models could be blocked if they pose a foreseeable risk of causing the death of over 50 people or over a billion dollars in property damage. The bill suggests the use of independent evaluators registered with the U.S. Commerce Department to assess potential risks, with a requirement for evaluators to report findings within 72 hours. This proposal reflects Anthropic's ongoing advocacy for AI regulation since June.