Anthropic backs bill to stop AI release if risk is $1B
Only the Commerce Department can stop a release, and only after a federal judge decides.
Anthropic is lobbying for a US AI safety bill, Politico tells. A registered third party can check the lab and its models. If it sees a risk of $1 billion in damage or more than 50 deaths or injuries, it must tell the department in 72 hours. The department must then give all the data to a federal judge, who decides if the release stops.
Anthropic's proposed US bill lets the government halt an AI model's release only above a high risk bar, and only for three causes.
