California Governor Gavin Newsom signed two new AI safety bills into law on Wednesday, adding another layer of state regulation to an industry that Washington has so far declined to govern at the federal level. Both Anthropic and OpenAI backed the legislation, though OpenAI waited until just before the signing to announce its support, raising questions about what changed the company’s calculus at the eleventh hour.
The bills set new rules for how outside groups evaluate artificial intelligence systems for safety. One creates a registry and ethical standards for third-party auditors that AI developers can hire to check whether their models comply with state AI laws. The other directs the state to establish criteria for, and verify the credentials of, what the legislation calls Independent Verification Organizations, groups that assess AI risks.
Neither bill number was publicly identified in the initial reporting. But the practical effect is clear: California is building a bureaucratic infrastructure around AI oversight, one auditor registry and credentialing board at a time. Whether that infrastructure protects the public or simply creates a compliance industry remains to be seen.
Anthropic, the AI company behind the Claude chatbot, backed both bills in August. OpenAI held off until the day of the signing, announcing its support just before Newsom put pen to paper. The timing gap is worth noting. Companies that lobby hard against legislation in private and then endorse it at the last possible moment are not acting out of conviction. They are managing headlines.
An Anthropic spokesperson said Wednesday that the company has “always been transparent that AI will bring both enormous benefits and unprecedented risks,” adding that the firm continues to “build models with some of the strongest safeguards in the industry.” The spokesperson went further:
“This work is also why we believe the world would benefit from the industry adopting a lawful, verifiable way to work together to pace how we release powerful models.”
Chris Lehane, OpenAI’s chief global affairs officer, said the company intends to keep pushing for AI safety measures in state legislatures until Congress passes national regulation. Lehane also endorsed two other California bills, including SB 1119, which addresses safety for chatbots aimed at children.
OpenAI CEO Sam Altman had separately reached out to Newsom with last-minute concerns about SB 1119 before it was signed. What those concerns were, and whether Newsom addressed them, remain unclear.
Newsom, in a statement to Politico, positioned California as filling a federal vacuum. He cited “recent incidents”, left unspecified, as reinforcing the need for guardrails.
“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 Washington to “step forward with robust, national regulations that match the urgency of this moment.” It is a familiar posture for Newsom, who has repeatedly cast his state as a policy laboratory for the nation, on climate, on immigration, and now on AI. Whether the rest of the country views Sacramento’s experiments as models or cautionary tales depends largely on where you sit.
The governor’s broader record in California invites skepticism about whether more regulation translates into better outcomes. Federal authorities recently made their first election fraud arrest in the state, renewing pressure on Newsom over the integrity of California’s voter rolls, a reminder that Sacramento’s appetite for regulation does not always extend to enforcing the rules that matter most.
The signings came against a backdrop of growing public anxiety about AI safety. Days before Wednesday’s ceremony, Jacob Coxon, a former researcher at both Anthropic and OpenAI, posted about his resignation on social media. The post quickly went viral. Coxon has said he believes AI companies are racing toward superintelligent AI and “gambling” with people’s lives.
His departure and public comments pushed federal lawmakers to step up calls for legislative action, though the specific lawmakers were not named. The episode illustrates a pattern: a whistleblower or departing insider sounds the alarm, public pressure spikes, and politicians rush to respond, often with legislation shaped more by the companies being regulated than by the people supposedly being protected.
That dynamic is visible here. Both Anthropic and OpenAI supported the bills Newsom signed. When the regulated create the rules, taxpayers and consumers should ask whose interests are actually being served.
California’s broader political conflicts with the federal government continue to multiply. A Biden-appointed judge recently tossed a DOJ lawsuit over the state’s transgender athlete policy on procedural grounds, one of several ongoing clashes between Sacramento and Washington over who sets the rules on everything from schools to immigration.
Newsom’s signing adds to what is becoming a dense web of California AI regulation. AP News reported that Newsom also signed, in September 2025, a landmark AI law, Senate Bill 53, the Transparency in Frontier Artificial Intelligence Act, establishing first-in-the-nation regulations on large-scale AI models. That law requires companies to implement safety protocols to prevent catastrophic misuse, such as bioweapon development or attacks on critical infrastructure. It defines catastrophic risk as causing at least $1 billion in damage or more than 50 injuries or deaths, imposes fines of $1 million per violation, and mandates that companies report critical safety incidents to the state within 15 days.
State Senator Scott Wiener, who authored the legislation, said: “With this law, California is stepping up, once again, as a global leader on both technology innovation and safety.”
The Washington Examiner reported that Senate Bill 53 also includes whistleblower protections for those reporting AI company wrongdoing, a provision that may have been shaped, in part, by the Coxon episode and similar departures. Newsom himself framed the legislation as an effort to “build public trust” in the technology.
Notably, Newsom vetoed a broader frontier-AI bill, SB 1047, in 2024. The current version is narrower, targeting only the most advanced “frontier” AI systems. That veto-then-sign pattern suggests the governor is calibrating his approach based on industry feedback, which, again, raises the question of who is really driving the regulatory train.
Both Newsom and OpenAI’s Lehane pointed to congressional inaction as the reason California must act. There is some truth to that. Congress has held hearings, issued reports, and proposed frameworks, but has not passed comprehensive AI legislation. The result is a state-by-state patchwork that benefits large companies with the legal teams to navigate fifty different regulatory regimes and disadvantages smaller competitors who cannot.
That is not an accident. When OpenAI and Anthropic endorse state-level regulation, they are endorsing a system that raises the cost of compliance for everyone, including the startups that might one day challenge their dominance. Regulatory capture does not always look like a lobbyist in a back room. Sometimes it looks like a press release expressing enthusiastic support for “safety.”
Conditions across California under Newsom’s leadership have drawn sharp criticism from national figures. Vice President JD Vance compared a Los Angeles neighborhood to conditions outside the U.S. Embassy in Haiti, a pointed assessment of the gap between Sacramento’s ambitions and the reality on the ground for ordinary Californians.
Meanwhile, the federal government has expanded its own physical footprint in the state. DHS recently spent $1.5 billion to purchase two California detention centers, a sign that Washington’s engagement with the state increasingly runs through enforcement, not deference.
Several questions hang over the new laws. The specific bill numbers for the two auditor-and-credentialing measures Newsom signed Wednesday were not publicly identified in initial reporting. The details of Altman’s last-minute concerns about SB 1119 remain undisclosed. Whether Newsom actually signed SB 1119 or merely received Lehane’s endorsement of it is unclear from the available record.
The full text of Jacob Coxon’s viral resignation post has not been published, nor has the platform on which he posted it been confirmed. And the federal lawmakers who reportedly stepped up their calls for action after Coxon’s posts have not been named.
These gaps matter. Legislation shaped in private conversations between a governor and the CEOs of the companies being regulated deserves more transparency, not less. The public has a right to know what was traded, what was softened, and whose objections carried weight.
California is building a regulatory apparatus for artificial intelligence. The question is whether it will serve the people who live with the consequences of the technology, or the companies that profit from it. If history is any guide, the smart money is on the latter.
By signing up, you agree to receive newsletters and promotional content from American Frontline News and selected publications in the American Digest Media Network, operated with Patriot Mom Digest LLC, and you accept our Terms of Use and Privacy Policy. You may unsubscribe at any time.
By signing up, you agree to receive newsletters and promotional content from American Frontline News and selected publications in the American Digest Media Network, operated with Patriot Mom Digest LLC, and you accept our Terms of Use and Privacy Policy. You may unsubscribe at any time.