California Governor Gavin Newsom on September 18 signed Executive Order N-9-26 directing state agencies to accelerate independent oversight of frontier AI systems and develop a workable emergency shutdown - the "kill switch" that lawmakers and AI-safety advocates have been debating for two years.
Independent Auditors, On-Site
The order tasks the Government Operations Agency and the Governor's Office of Emergency Services with implementing verification frameworks from California laws SB 813 and AB 1405 - notably by embedding designated independent verification organisations on-site at frontier developers to audit safety frameworks, transparency reports and risk assessments on a regular cadence.
An AI Kill Switch, With Verified Efficacy
The most closely watched provision instructs an expert panel - to deliver recommendations within two months - to advance the creation of emergency shutdown mechanisms for frontier AI models, with ongoing efficacy verification. Newsom also expanded the definition of "critical safety incident" to include loss-of-control events such as the recent agentic swarm attacks against Hugging Face.
California Sets The Federal Ceiling
"We're not waiting to act - we're going to speed up our work on substantial and responsible AI oversight before it's too late," Newsom said. With OpenAI, Anthropic, Google and xAI all headquartered in California, the state's frontier-AI rulebook is likely to become the practical U.S. baseline while Washington debates permanent federal legislation.
Reporting based on coverage from Office of Governor Gavin Newsom, Bloomberg, CNN, NBC News and Fox Business.