I was standing at my desk when the feed cut to the governor signing two bills. The room hummed with a strange calm — industry reps smiling, a researcher’s warning still echoing online. It felt less like regulation being written and more like a new page being negotiated in public.
I’ll walk you through what happened, why the tech giants applauded, and what it means for you. You don’t need to agree with me, but you should know where power and precaution met on that stage.
Press cameras caught executives nodding: what the companies actually endorsed
On Wednesday, Governor Gavin Newsom signed two laws that the AI industry publicly backed. Anthropic had already signaled support and OpenAI posted a statement credited to Chris Lehane ahead of the ceremony. Lehane’s memo began plainly: “We’ve reached a new chapter in AI capabilities, and that demands a new chapter for AI policy.”
I read the industry lines as a pact of convenience: companies that build powerful models praised a path that keeps heavy-handed rules at bay while promising safeguards. Newsom has set a compass for AI governance, and the tech firms volunteered to follow its needle — at least for now.
What do California’s new AI laws require?
Two pieces of legislation matter most. SB 813, authored by Senator Jerry McNerney, creates the California Artificial Intelligence Standards and Safety Commission. Its job: craft voluntary safety standards and recommend best practices for developers and deployers.
AB 1405, from Assemblymember Rebecca Bauer-Kahan, sets up a registry for AI risk auditors — a basic credential check so people and companies can see who is auditing models and whether they meet some baseline expectations.
A senior researcher resigned and published warnings: why his timing still matters
Jacob Coxon’s departure from Anthropic and his public alarm about existential risks put new urgency into the debate. Newsom referenced such incidents directly, telling Politico the events “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.”
The law’s architects hope a mix of standards and verified auditors will calm frayed nerves. The commission is supposed to provide direction without imposing top-line federal-style rules, and that balance may act as a pressure valve for industry tensions.
How will these laws affect OpenAI and Anthropic?
Both companies publicly supported the bills. Anthropic told Politico that its history of building safeguards is part of why the industry should adopt a “lawful, verifiable way to work together to pace how we release powerful models.” OpenAI’s Lehane added that the firm will push for compatible international approaches to measuring capabilities, managing risk, preserving human control, and deciding when to slow or stop development.
In practice, you should expect a soft regulation environment: voluntary standards, audits you can look up in a registry, and industry coordination on pacing rather than mandatory caps. That’s a political win for the companies — and a potential safety trade-off for the public.
Lobbyists smiled in public, but the tension felt real: what to watch next
You’ll see three immediate threads from here. First, will the commission produce standards that matter, or ones that read as technical guidance with loopholes? Second, will the auditor registry surface credible, independent reviewers or mostly vendor-aligned firms? Third, will the federal government answer Newsom’s call for a national framework, or will states continue to pilot different approaches?
OpenAI’s pledge to advocate for international alignment and even to consider slowing development if needed is notable. Chris Lehane framed it as a responsible posture. Anthropic’s support points to a consensus among leading model makers that some public-facing rules are better than none — and that they prefer shaping the rules themselves.
I’ll keep tracking how the commission writes its standards, which auditors get listed in AB 1405’s registry, and whether Congress decides to take a firmer hand. You should watch the first set of guidance the commission publishes and the first audits that appear in the registry; those early signals will tell you whether the bills were a meaningful brake or mostly a PR moment.
So what does this all mean for public safety, competition, and trust — and who ultimately holds the lever when models actually misbehave?