Newsom Proposes Emergency Shutdown Powers for Dangerous AI Systems

Posted

California Gov. Gavin Newsom is pushing for stricter oversight of artificial intelligence, unveiling a proposal Wednesday that could give state officials the power to shut down dangerous AI systems. The recommendations would expand California’s oversight to regulate advanced AI developers. 

Under the proposal, the state could order dangerous AI systems to shut down during an emergency, while companies would be required to outline verified procedures for stopping systems that pose serious risks. Developers will also face mandatory safety standards and independent testing throughout the development process. California could also gain the authority to limit recursive self-improvement, a process that helps AI systems develop increasingly advanced versions of themselves. 

The proposal does not call for a universal “kill switch”, but instead outlines potential emergency controls, including shutting down servers, suspending development or restricting access to models. 

Newsom emphasized his goal to turn the recommendations into law. “I’m already working with members of our Legislature to introduce the legislative frameworks to move quickly and make real what we’re promoting and promising in this report,” Newsom told reporters Wednesday, according to POLITICO. He described the recommendations as a foundation for California lawmakers to turn into enforceable safety regulations. 

The proposal would build on existing regulations such as Senate Bill 53, the Transparency in Frontier Artificial Intelligence Act. SB 53, signed by Newsom in 2025, requires covered developers of advanced AI systems to report certain serious incidents and disclose their safety practices. Newsom’s latest recommendations would potentially give state officials the authority to intervene when systems pose a risk and strengthen independent oversight.

Newsom previously vetoed a major AI safety bill, SB 1047, in 2024, arguing that the bill focused too heavily on the largest AI models while overlooking risks posed by smaller systems. His newest proposal reevaluates a few of the safety measures outlined in that legislation, including emergency shutdown procedures and stronger government oversight.

Instead of announcing the proposal in California, Newsom did so in Washington D.C., positioning the state as a potential example for AI regulation nationwide. The announcement follows as Newsom attempts to distinguish California’s approach from President Donald Trump’s pushback on stricter AI regulation. POLITICO described the Washington unveiling as another hint of Newsom’s possible 2028 presidential ambitions.

For now, the recommendations aren’t guaranteed to become law and remain proposals. With Newsom’s term ending in January, incoming Gov. elect Xavier Becerra and California lawmakers would play a key role in deciding whether the safeguards move forward.

The post Newsom Proposes Emergency Shutdown Powers for Dangerous AI Systems appeared first on LAmag.

General, News & Politics, Politics, Technology, AI, AI Kill Switch, artificial intelligence, Governor Gavin Newsom, Xavier Becerra