Peri News / United States
California proposes AI oversight and emergency shutdown powers
Recommendations announced by the governor call for independent checks, mandatory standards and limits on models capable of developing their own successors.

California Governor Gavin Newsom announced recommendations on November 18, 2026, to strengthen state oversight of AI. According to the governor’s office press release, the proposals include state-set baseline standards for all AI companies and independent safety checks throughout model training, development and deployment. These are recommendations to change California’s regulatory framework, rather than an announcement of new duties already in force.
The recommendations would also expand safety incident reporting and require companies to establish verified protocols to contain risks and shut down models when necessary. The state would gain authority to order an immediate shutdown in an emergency. Another proposal would allow limits on the development and use of recursive self-improvement, described as a scenario in which an AI model can design its successor without human input. The state Government Operations Agency developed the report in consultation with an expert group and the California Office of Emergency Services.
The press release distinguishes these proposals from two laws it reports were signed in 2026: SB 813, concerning a certification framework for independent organizations assessing AI systems for safety and risk, and AB 1405, establishing a state registry of AI auditors and standards for independence, transparency and integrity. This account of those measures is limited to the official announcement: the report and legislative texts have not been examined directly, leaving their detailed legal scope, commencement dates and implementation arrangements to be verified.
