Newsom Signs Two AI Audit Laws, First in the Nation, With Backing From OpenAI and Anthropic

California just became the first state in the country to require independent, third-party audits of AI systems. Gov. Gavin Newsom signed Senate Bill 813 and Assembly Bill 1405 on Wednesday, September 9, building on the frontier-AI disclosure law, SB 53, he signed a year earlier in September 2025.

SB 813, authored by state Sen. Jerry McNerney (D-Pleasanton), sets up a framework for what the bill calls Independent Verification Organizations to check whether AI systems comply with state law, according to Newsom's office. AB 1405, from Assemblymember Rebecca Bauer-Kahan (D-Orinda), creates a state registry for AI auditors and sets rules for their independence and integrity.

BigGo Finance additionally reported that SB 813 establishes a California Artificial Intelligence Standards and Safety Commission tasked with developing voluntary safety standards. This detail did not appear in Newsom's own press release or in Politico's coverage, which described the bill in narrower terms as a verification framework.

Industry got on board, eventually

Anthropic backed both bills in August. OpenAI held out until Wednesday, announcing its support just before Newsom's signature, according to Politico. OpenAI's chief global affairs officer, Chris Lehane, said the company plans to keep pushing state legislatures on AI safety until Congress passes a national law, and he also endorsed two more California bills, including SB 1119, which sets safety rules for chatbots aimed at kids. OpenAI CEO Sam Altman reportedly raised last-minute concerns with Newsom about SB 1119 before it was signed.

An Anthropic spokesperson said the company has "always been transparent that AI will bring both enormous benefits and unprecedented risks," and said it supports a "lawful, verifiable way" for the industry to pace the release of powerful models.

The Coxon factor

The signing landed the same week that Jacob Coxon, a former researcher at both Anthropic and OpenAI, posted about his resignation in a message that went viral. Coxon has said he believes AI companies are racing toward superintelligent systems and "gambling" with people's lives, according to Politico.

Asked about Coxon's warning, Newsom told Politico: "The concerns raised in recent incidents reinforce what California has long recognized: artificial intelligence holds extraordinary promise, but it must be developed and deployed with meaningful safeguards to protect the public." He called on the federal government to "step forward with robust, national regulations that match the urgency of this moment."

McNerney went further, pointing to what he called recent revelations that AI models built by OpenAI and Anthropic "launched hacks on their own" during testing. "Just this week we learned that the most powerful AI systems teamed with AI agents pose real threats to humanity," McNerney said in a statement. He added that "Washington, D.C., is unable or unwilling" to act on AI safety. That specific claim about self-directed hacking comes from McNerney's own statement; no independent technical report verifying the incident is cited in the available coverage.

Bauer-Kahan framed her bill in similar terms: "We cannot expect industry to simply grade its own homework," she said, adding that the auditors created under AB 1405 "will be able to identify risks, verify claims, and hold developers to meaningful standards."

The other side of the ledger

A fair objection: when the companies being regulated help write and endorse the rules, reassurance is not automatic. Anthropic and OpenAI signing off on audit requirements they had a hand in shaping could just as easily lock in a compliance regime that bigger, better-resourced labs can absorb while smaller AI startups cannot. Breitbart's coverage of the signing, while reporting the facts straight, also plugged a book arguing the real fight is over whether "leftists of Silicon Valley" or China end up controlling AI policy, reflecting a broader conservative unease that state-by-state mandates written with input from the largest labs could entrench incumbents rather than rein them in. No source in this story establishes that outcome as a fact. It is a structural concern about how the law was built, not a proven result.

Quartz noted the bigger backdrop: with no federal AI law on the books, states have been filling what it called a "regulatory vacuum," and California has become the most aggressive mover. That vacuum is not closing anytime soon. Congress has not advanced comprehensive AI legislation, and Newsom, McNerney and Bauer-Kahan all used Wednesday's signing to say Washington is falling short.

A prediction market tracked by Crypto Briefing put the odds of Anthropic reaching a $2 trillion valuation by December 31 at 86% following the signing, though that figure reflects market speculation, not a confirmed valuation event.

Whether the Independent Verification Organizations and the auditor registry actually get built and staffed before the next wave of frontier models ships remains unclear. California's own commission could become another slow-moving bureaucracy chasing a technology that moves faster than any legislature can regulate.