Anthropic asked Newsom to back court powers over risky AI

Anthropic sent California Gov. Gavin Newsom a letter earlier in October 2026 proposing that courts be able to block AI programs posing serious risks. The proposals, reported by POLITICO on October 27, also called for AI companies to develop kill switches and place outside safety auditors alongside employees.

What Anthropic proposed to Newsom

The letter described three proposed safeguards: court authority to block AI programs deemed to pose serious risks, company-developed kill switches, and outside safety auditors embedded alongside company employees. A kill switch is a mechanism for shutting down technology; the letter’s reported proposal does not specify how one would work.

Anthropic argued that California should create effective oversight because the state’s approach could influence other states and future policy debates. The letter was a policy proposal. It did not itself establish the court authority it advocated.

For context on the company’s earlier position in the debate, see NeoTeo’s previous coverage of Anthropic’s support for government oversight of AI safety.

A separate ruling addressed a federal designation

A California judge’s August ruling involved a different question: whether the federal government had adequately justified a supply-chain-risk designation against Anthropic. In a Bloomberg report republished by East Bay Times, Judge Rita F. Lin’s order on August 27, 2026, required the administration to lift the ban tied to that designation.

The August order did not require the Department of War to use Anthropic products and did not prevent a lawful transition to other providers. It concerned the designation in that case, rather than the general court authority Anthropic proposed in its letter.