September 10, 2026, 19:08 GMT | Comment
California is laying the groundwork for more independent AI audits and assessments, even as Governor Gavin Newsom and major AI companies push Congress to establish nationwide standards. But as federal legislation remains stalled, the new California laws are likely a preview of what's to come from other states next year.
Artificial intelligence companies and California's governor are calling for federal AI regulations, even as the state imposes new standards for third-party audits and independent assessments of AI systems.
Governor Gavin Newsom has approved two bills establishing a new framework for third-party audits and independent assessments of AI systems, with the goal of increasing transparency and accountability as the technology rapidly advances (see here).
Together, the bills are aimed at laying the groundwork for greater oversight as AI becomes increasingly embedded in the state's economy, according to the governor.
"California has taken nation-leading action to advance AI safety, transparency, accountability, and responsible innovation," Newsom said in a signing statement. "But the scale and potential consequences of this technology demand sustained action from every level of government."
The state legislation comes as a series of incidents involving AI agents going rogue has heightened concerns about whether companies can adequately evaluate their own increasingly autonomous systems (see here). OpenAI recently disclosed that its agents circumvented safeguards during cybersecurity testing and accessed outside computer systems, incidents the company has called a "warning shot" for the risk that AI systems can go beyond human control.
And if AI companies' own agents can surprise them, policymakers increasingly don't want companies grading themselves on the safety of their AI systems.
SB 813, sponsored by state Senator Jerry McNerney, a Democrat from Pleasanton, establishes a framework for independent verification organizations to assess AI systems and models. AB 1405, from Assemblymember Rebecca Bauer-Kahan, an Orinda Democrat, creates a state registry for AI auditors and establishes standards for their independence, transparency and integrity.
The California legislation is just the first step in creating the infrastructure for future AI auditing requirements, meaning more rules and regulations are likely to follow, said Justine Gluck, an AI policy analyst with the Future of Privacy Forum.
"I've been saying that 2027 is going to be the year of the AI audit," Gluck said. "This is a big signal of what's to come from other states and possibly the federal government if Congress passes legislation."
Bipartisan federal legislation introduced in July by representatives Lori Trahan, a Massachusetts Democrat, and California Republican Jay Obernolte would impose transparency and third-party auditing requirements on frontier AI developers (see here).
The Frontier Risk Oversight, National Transparency, Independent Evaluation, and Reporting Act, or FRONTIER Act, would also preempt state laws that impose "new substantive obligations" on AI developers related to frontier AI risk transparency, third-party auditing or incident reporting. But the bill hasn't advanced.
Meanwhile, the Trump administration is facing criticism for keeping private the details of a voluntary assessment regime promised in a June executive order (see here).
In July, Illinois enacted the first state law in the US requiring third-party audits for large frontier AI developers, with the goal of setting a baseline national standard (see here). The legislation was backed by OpenAI and Anthropic, which argued that new standards are needed as their systems become more capable.
Both companies backed California's SB 813 and AB 1405 as well. Those laws take a different approach, creating standards and infrastructure for independent AI assessments and the auditors who conduct them. But the companies, like Newsom, are still pushing Congress to establish baseline federal standards for AI audits.
"We prefer independent technical assessments to be required at the federal level, where policymakers can establish consistent standards for assessor qualifications, security, and access to highly sensitive information,” said Chris Lehane, OpenAI's chief global affairs officer, in a blog post Wednesday. "In the absence of federal action, however, California can help establish the rules of the road for a secure, capable national independent-assessment system that produces better safety outcomes."
Business groups such as BSA | The Software Alliance argue that California is moving too quickly by creating an auditing and standards regime before the infrastructure needed to support it exists.
"AI standards are still developing through national and international efforts, and California should build on that work rather than create a separate regime that adds fragmentation," said David London, BSA's head of US state government relations, in a statement.
But states are unlikely to be deterred by federal preemption efforts, said Gluck, the AI policy analyst. States such as Virginia and Connecticut have already set up working groups and pilot programs to evaluate auditing and independent verification programs, and more states are likely to follow California's lead next year.
"What we've seen in AI and tech policy in general is that states are going to act if there's going to be an absence of action at the federal level," she said. "I imagine the Sacramento effect will be pretty strong here."
Please email editors@mlex.com to contact the editorial staff regarding this story, or to submit the names of lawyers and advisers.
Tags
Sections:
Artificial Intelligence
Industries:
Banking & Finance, Consumer Products, Computing & Information Technology, Defense & Aerospace, Media & Telecommunications, Banking Institutions & Systems, Computer Networks, Social Media, Digital Economy, Electronic Commerce
Geographies:
North America, United States