California Becomes First State to License AI Auditors

California now has the country's first licensing regime for the people who check artificial intelligence systems for compliance. Governor Gavin Newsom signed two bills on September 9 that together build a registry of accredited AI auditors and a legal framework for the independent organisations that will employ them.

AB 1405, written by Assemblymember Rebecca Bauer-Kahan, creates the state registry. From 2029, nobody will be allowed to conduct a covered AI audit in California without registering first, and registered auditors will have to follow independence rules modelled on those that govern financial accountants, according to the governor's office. The companion measure, SB 813 from Senator Jerry McNerney, sets up the category of independent verification organisation that can assess whether a model meets the state's legal requirements.

Audits with teeth, eventually

The accounting comparison is deliberate. Independent audit only means something if the auditor cannot be quietly captured by the company paying the bill, which is the problem the new independence rules are meant to address. Quartz reported that the laws make California the first US state to require this kind of third-party review, rather than leaving safety checks entirely to the firms building the models.

The three-year lead time before registration becomes mandatory gives the state room to write accreditation standards and gives a professional audit market time to form. It also means the immediate effect on any shipping product is close to zero.

The industry signed on

What sets this round apart from earlier fights over AI regulation is who supported it. Both Anthropic and OpenAI backed the audit bills, a contrast with the lobbying that surrounded California's vetoed 2024 safety legislation. Newsom used the same signing to approve what his office called the strongest child-safety chatbot and social-media laws in the nation, part of a larger package the legislature sent him as the session closed.

The move follows months of state-level activity while federal rules stall. California's legislature passed 26 AI and social-media bills this session, and Massachusetts advanced its own safety bill that split the major labs. With Washington quiet, the rules that bind American AI companies are increasingly being written in statehouses.