California: Frontier AI Transparency and Oversight

California enacted SB 53 and later directed work on independent AI oversight and possible shutoff requirements. A study directive does not establish a deployed kill switch.

kindGovernment
siteca.gov

Updated

Law and technical oversight have different roles

California enacted SB 53 in September 2025, establishing a transparency framework for covered frontier AI developers. The governor’s announcement describes safety-framework disclosure, incident reporting and whistleblower protections. These requirements concern developer conduct and accountability rather than certifying that every covered model is safe.

What the later order directed

The September 2026 executive order directed work to accelerate independent oversight and convened experts to provide further guidance. The announcement described emergency shutoff requirements as proposals under consideration. It did not establish that a universal AI kill switch had been built or deployed.

California’s contribution to AI governance is recorded through these specific legislative and executive actions. Signature of a law, implementation of its oversight bodies and development of a technical safeguard remain separate milestones.

Sources

  1. [1]
  2. [2]

State of California history

2026

2025