California lawmakers have approved legislation that would establish a state framework for independently evaluating the safety and risk of artificial intelligence systems.

The California Legislature passed SB 813, authored by Sen. Jerry McNerney, on Aug. 30. The measure now goes to Gov. Gavin Newsom for consideration.

If signed into law, SB 813 would direct the state to establish a framework for qualifying and designating Independent Verification Organizations (IVOs)—outside organizations with the expertise to conduct assessments of AI models and systems. The goal is to provide an independent means of evaluating whether AI safety measures and other safeguards perform as intended.

McNerney's office and Fathom, the bill's sponsor, describe the measure as the first state framework of its kind in the nation.

The legislation represents a shift toward third-party assessment as AI systems become more capable and increasingly integrated into business, education, government, and other areas of society. Rather than relying solely on AI developers to evaluate and report on the effectiveness of their own safeguards, the IVO model would create a mechanism for qualified outside organizations to conduct independent testing and analysis.

"The power of the independent verification model is that it lets independent, qualified experts assess the risk of AI capabilities, and it keeps pace with the technology instead of freezing a single test into law," said Andrew Freedman, Fathom's CEO.

Freedman continues, "It gives the public a way to trust AI and gives companies a credible way to show their system safeguards work as intended."

The bill builds on California's broader efforts to establish requirements around frontier artificial intelligence. In 2025, the state enacted SB 53, the Transparency in Frontier Artificial Intelligence Act, which established transparency and safety requirements for certain large frontier AI developers, including requirements related to safety frameworks and critical safety incidents.

SB 813 addresses a different part of the emerging AI governance framework: independent evaluation. Under the legislation, the state would establish qualifications and standards for organizations seeking to serve as IVOs, creating a foundation for an ecosystem of independent AI evaluators.

The approach also reflects a broader national discussion about third-party testing and verification of AI systems. Policymakers at the federal and state levels are considering mechanisms for independent assessments as companies deploy increasingly capable AI models, although approaches and proposed requirements vary.

For technology companies and organizations deploying AI, independent verification could eventually provide a way to demonstrate that security, safety, reliability, and other controls are functioning beyond a company's own internal assessments. For policymakers and the public, it could provide additional information for evaluating claims about the risks and safeguards associated with increasingly consequential AI systems.

The measure's next step is consideration by Gov. Newsom. If enacted, California policymakers and technical experts would then face the task of developing the standards, qualifications, and processes needed to put the IVO framework into practice.

Fathom said it plans to continue working with state policymakers, technical experts, and industry stakeholders as California develops the framework.

Source: Fathom AI, Inc.

Stay Connected with ISE Magazine 

Subscribe to our newsletters and magazine for the latest telecom insights, explore the current issue for in-depth features and strategies, and register for upcoming webinars to learn directly from industry leaders.