Executive order · California, US
Executive Order N-9-26 on frontier-AI safeguards
The order directs California officials to develop recommendations on stronger frontier-AI safeguards, including independently verified emergency shutdown mechanisms, onsite audits and expanded incident reporting. It also sets deadlines for implementing AI-verification and auditor frameworks. It does not itself impose new shutdown or audit requirements on AI companies.
- In force
- Passed
Updated
What it requires
- State officials must deliver recommendations on stronger frontier-AI safeguards by November 16.
- The review must assess onsite independent audits at large frontier developers and independent verification of their safety frameworks, transparency reports and risk assessments.
- The review must consider requiring emergency shutdown mechanisms for frontier models and ongoing independent verification of their effectiveness.
- The review must consider expanding reportable critical safety incidents to include loss-of-control incidents.
- The Government Operations Agency must publish application requirements, procedures and criteria for independent verification organizations by May 1, 2027.
- The Government Operations Agency must complete a statutory AI-auditor implementation step and begin required follow-on actions by December 1, 2027.
Who it affects
The order directs the California Government Operations Agency, working with the Governor’s Office of Emergency Services and national experts, without creating new compliance duties for private AI companies or enterprise buyers.
Timeline
Governor Gavin Newsom issued the order directing a frontier-AI safeguards review and accelerating implementation of independent AI-verification and auditor frameworks. Our story