Staff Reports
California Governor Gavin Newsom just signed two new AI bills that set up a state‑run system for third‑party audits and “independent verification” of AI systems. The measures — known as SB 813 and AB 1405 — create a registry and a state role in picking who can test AI models. This is a big move. And it deserves a hard look from anyone who cares about free markets, innovation, or basic fairness.
What Newsom signed and why it matters
SB 813 tells the state’s Government Operations Agency to choose and certify what it calls Independent Verification Organizations to assess AI risks. AB 1405 creates a registry for AI auditors and sets rules about independence and transparency. On paper, these sound like sensible “AI safety” steps. In practice, they build a power center: the state decides who gets to be an auditor and who doesn’t. That’s not just regulation. It’s gatekeeping for an entire technology sector.
Who backed the bills — and why that should make you raise an eyebrow
The most interesting part is who cheered as Newsom signed these rules. Anthropic publicly backed the bills earlier, and OpenAI announced its support right before the governor signed them. If big AI firms like OpenAI and Anthropic help shape the rulebook for auditors, we should be wary. Companies inevitably prefer auditors who won’t bite the hand that feeds them. We already saw a public scare from a departing researcher saying companies are “racing” toward dangerous AI. Now the industry gets to help name the referees. That’s a little like letting boxers pick the judges.
Big risks: capture, politicized audits, and chilling innovation
There are real problems ahead. First, who really controls the selection and the standards? GovOps will write the rulebook, and that process will be where the real power lives. Second, a state registry gives officials the tools to favor or freeze out competing firms or ideas. Third, the new assurance layer could become a legal safe harbor that shields favored players while punishing newcomers. In short, this system could slow innovation, hand big tech more control, and give regulatory busybodies a new lever to pull on speech and business decisions — all in the name of “safety.”
What should happen next
Congress should step up with a clear, national framework that protects innovation and competition while setting real, technical safety standards — not just paperwork and approved vendors. In the meantime, Californians and free‑market defenders should demand transparency about who lobbied on the bills, push for open and competitive auditor selection, and insist on guardrails that stop the registry from becoming a political weapon. If we let the state and the largest firms carve up the AI ecosystem now, ordinary citizens and small innovators will pay the price later. That’s not safety. It’s capture — with government help.
