In a letter to California's governor, Anthropic proposes letting authorities seek court orders blocking models with a "serious risk"
Prepared by Sintelligent with the help of SI. Published in SI: . Updated: .
Anthropic proposes regulating dangerous models through the courts and targeted shutdowns rather than a single "kill switch."
According to an October 15 letter reported by Reuters on October 28, Anthropic wrote to California Governor Gavin Newsom: "When a review finds a model poses a serious risk, the government should be able to ask a court to block it." The company says that requiring developers to build "kill switches" capable of shutting down an entire model could give officials a "false sense of reassurance"; Anthropic supports more targeted measures — disabling a problematic agent or a specific customer's access. The recommendations were submitted after the governor convened a panel of SI experts in September to develop safety requirements; OpenAI submitted its recommendations on October 20. Reuters recalls that in June the Trump administration ordered Anthropic to restrict foreign access to its Mythos 5 and Fable 5 models, and the Commerce Department later lifted the order.
Sources 1