Loading page…
The package adds a route for third parties to assess AI systems, while leaving California to define how voluntary safety standards work in practice.
Listen to this story
California’s September 9 signing of SB 813 and AB 1405 moves AI oversight beyond developer disclosures and incident reporting toward independent verification. SB 813 creates a framework for organizations assessing AI systems and establishes a commission for voluntary safety standards; AB 1405 adds an auditor registry, credential checks, and requirements for independence, transparency, and integrity. OpenAI and Anthropic supported the package.
AB 1405 creates a state registry for AI auditors and a process to verify their credentials.
SB 813 covers independent assessments of AI systems and models for compliance with California law.
The new framework supplements existing requirements for safety protocols and critical-incident reports.
California Governor Gavin Newsom has signed two laws that create a framework for independent assessments of AI systems and state standards for the auditors conducting them. The measures place outside review beside California’s existing developer safety-disclosure rules, making the credibility of reviewers a central part of the state’s AI oversight.
Newsom signed Senate Bill 813 and Assembly Bill 1405 on September 9. The governor’s office described the package as first-in-the-nation standards for independent AI assessments and third-party audits, intended to increase transparency and accountability as AI becomes more embedded in California’s economy and public life.
SB 813 establishes a framework for independent verification organizations to assess AI systems and models for compliance with state law. Separately, it establishes the California Artificial Intelligence Standards and Safety Commission, which will create voluntary AI safety standards.
OpenAI and Anthropic backed both bills. Before the signing, OpenAI said it would advocate for compatible international approaches to measuring AI capabilities, managing risk, preserving human control, and deciding when development should slow or stop.
The earlier California law focused on company safety protocols and critical-incident reporting. The new package creates a path for outside assessment and sets standards for its auditors. Newsom has also called for robust national AI regulation, saying the technology’s scale and potential consequences require sustained action at every level of government.
Editorial analysis
California’s earlier frontier-AI law centered on what developers must disclose and report. SB 813 and AB 1405 add a different piece of infrastructure: organizations outside the developers can be evaluated for their role in assessing AI systems. That does not make every model automatically subject to an audit, nor do voluntary standards by themselves settle how demanding reviews will be. The consequential next step is whether California turns the new framework into a credible route for verification that complements developer reporting rather than merely formalizing it.
Loading discussion...
Make your case
A view to debateState rules can build trust while shaping who performs outside AI review.
Explain which safeguard would make outside review credible.
Be the first to share a perspective or an experience.
Reader comments
Newest comments first. Replies stay oldest first.