Anthropic urged California Governor Gavin Newsom in an October 15, 2026 letter to allow government officials to seek court orders blocking risky AI models. Submitted following Newsom's September expert safety panel, the letter opposes mandatory full-model kill switches like those in the federal AI Kill Switch Act proposal, favoring targeted interventions and embedded auditors. This comes amid federal legislative debates and past Trump administration foreign access restrictions on Anthropic's Mythos 5 and Fable 5 models.
Anthropic's safety proposals to California
- ▪Anthropic opposes regulatory requirements for full model kill switches if control is lost, advocating instead for targeted safety interventions like disabling specific AI agents or cutting off customer access as more effective alternatives
- ▪Anthropic's October 15, 2026 letter to Governor Gavin Newsom advocates having outside safety auditors embedded alongside company employees to monitor AI developments
- ▪Anthropic stated in its October 15, 2026 letter to Governor Gavin Newsom that California's approach to AI safety could become a model for other states and the baseline for future policy debates
- ▪Anthropic sent an October 15, 2026 letter to California Governor Gavin Newsom recommending court orders blocking AI models posing serious risks, submitted after Newsom convened a September 2026 panel of artificial intelligence experts to draft regulatory proposals
Federal Kill Switch proposals
- ▪A bipartisan group of US senators debated legislation in September 2026 allowing the federal government to block unsafe AI models subject to federal court challenges, but Congress recessed without advancing that legislation
- ▪US House lawmakers proposed the AI Kill Switch Act in July 2026, which would grant US officials the power to order AI firms to shut down models threatening human life or the economy
Story comments
Loading comments…