Back to news
policyAnthropic2026-10-29

Anthropic letter to California Governor Newsom urges courts be able to block high-risk AI models

Anthropic wrote to California Governor Gavin Newsom on October 15 arguing that government should be able to apply to courts for injunctions blocking the release of AI models posing a serious risk to the public, and opposing reliance on a full model kill switch, which it says can create a false sense of security, preferring targeted shutdowns of specific agents or suspending access for particular customers. The letter does not specify risk thresholds or evidentiary standards. Newsom convened an AI expert panel in September that is still working on safety requirements, and OpenAI submitted separate suggestions on October 20. A Senate bill discussed in September would have authorized bans on some high-risk models with a right to challenge in federal court, but stalled before the election recess. The letter's signatory was not disclosed.

Anthropic纽森加州州长AI 监管法院禁令模型发布