Anthropic Lobbies for an AI Safety Bill with High Intervention Thresholds

Breaking News from Beating AI: According to Politico, Anthropic is privately lobbying U.S. lawmakers on a draft bill on advanced AI safety that would establish strict thresholds for government intervention in AI models. Under the draft, the government may only seek to halt a model if it faces a foreseeable and imminent specific risk likely to cause losses exceeding $1 billion, or result in the death or serious injury of more than 50 people. The draft limits regulatory triggers to three categories of risk: AI models being used for biological weapons development or proliferation, models escaping developer control, and models triggering cyberattacks. The bill also requires AI developers to conduct transparent testing and undergo independent third-party evaluations. If an evaluation body identifies relevant risks, it must report them to the U.S. Department of Commerce within 72 hours; the Department may then petition a federal court to suspend the model’s operation, with final authority resting with the court. Additionally, the draft prohibits other U.S. government departments and officials from independently halting the development, deployment, or internal use of AI models. Anthropic states that the proposal is derived from its advanced AI framework released earlier this year, designed to require developers to undergo safety assessments while granting the government the power to block dangerous model deployments. The draft is currently in the lobbying phase and has not yet become law. If enacted, its provisions could significantly restrict the U.S. government’s authority to take emergency action against advanced AI models.