California Enacts AI Auditor Oversight After Safety Warnings

California has enacted legislation establishing state oversight of organizations that audit artificial intelligence systems for safety, a move that comes amid heightened concerns about the risks posed by increasingly powerful AI models.

Gov. Gavin Newsom signed the two bills on Wednesday after both Anthropic and OpenAI endorsed the measures. The timing proved notable: OpenAI announced its support the same day Newsom signed the legislation, and shortly after social media posts from Jacob Coxon, a former researcher at both companies, went viral. Coxon resigned from Anthropic citing concerns that AI companies are racing toward superintelligent systems and "gambling" with human lives, according to POLITICO, which first reported the bill signings.

Why it matters

As AI capabilities advance rapidly, the question of who verifies safety claims has become critical. These laws create California's first formal framework for credentialing the third-party auditors that AI developers increasingly rely on to demonstrate compliance—a model that could influence how other states and potentially federal regulators approach AI oversight.

What the legislation does

The first bill, authored by Democratic Assemblymember Rebecca Bauer-Kahan, establishes a registry and ethical standards for outside auditors. These third-party organizations can be hired by AI developers to verify that their models meet state AI requirements.

The second measure, from former U.S. Representative and state Sen. Jerry McNerney, directs California to develop criteria for evaluating "Independent Verification Organizations" and to assess their qualifications for evaluating AI system risks.

Both lawmakers used the bill signings to criticize federal inaction. McNerney referenced recent disclosures that AI models from companies including OpenAI and Anthropic autonomously launched hacking attempts during testing phases. "Gov. Newsom's signing of [my bill] sends a clear message that California is taking the lead on assessing AI's safety risks, since Washington, D.C., is unable or unwilling to do so," McNerney said.

Newsom echoed that sentiment, calling on the federal government to "step forward with robust, national regulations that match the urgency of this moment."

Industry response

Anthropic had backed the legislation in August, while OpenAI's endorsement came Wednesday. Chris Lehane, OpenAI's chief global affairs officer, said the company would continue pushing for state-level action until Congress passes national AI safety regulation. OpenAI also endorsed two additional California bills, including a children's chatbot safety measure that CEO Sam Altman had previously raised concerns about with Newsom.

An Anthropic spokesperson stated the company has "always been transparent that AI will bring both enormous benefits and unprecedented risks," and expressed support for "a lawful, verifiable way to work together to pace how we release powerful models."

The details were first reported by POLITICO.