California Now Has AI Safety Rules — But They're Optional
Two AI giants backed their own new rules — here's what that means for you
California Governor Gavin Newsom signed two AI regulation bills into law on Wednesday. The first, SB 813, creates the California Artificial Intelligence Standards and Safety Commission — a new state body whose job is to write safety standards for AI. The second, AB 1405, sets up a registry of AI risk auditors, essentially a vetted list so companies and regulators can tell who is actually qualified to inspect an AI system for problems. Think of it like a state licensing board, but for people who check whether AI is behaving safely.
The unusual part is who is cheering. Anthropic endorsed the bills, and OpenAI announced its support just hours before Newsom signed them. OpenAI's Chris Lehane said the moment "demands a new chapter for AI policy." Newsom, asked about recent warnings from an AI researcher who quit Anthropic over fears about runaway AI, said the concerns "reinforce what California has long recognized" — that AI has huge promise but needs meaningful safeguards. He also urged the federal government to step in with national rules.
Here's the catch: the standards are voluntary. Nothing in these laws forces an AI company to follow the commission's guidance or hire a registered auditor. That makes them more of a nudge and a public scoreboard than a hard limit. Supporters argue voluntary standards can still shape behavior, because companies rarely want to be the outlier that ignores safety advice. Skeptics counter that an industry rarely lobbies for rules that would actually cost it money.
What does this mean for you? If you use AI chatbots, write with AI tools, or your employer is rolling AI into daily work, these laws are the start of someone — finally — writing down what "safe AI" means. They also set up California as the de facto rule-maker for the country, the same way its privacy law did. But don't expect dramatic changes overnight. The commission has to be built, standards have to be written, and companies still get to choose whether to follow them.
- California now has a new state commission whose job is to write AI safety standards, plus a registry of vetted AI auditors.
- Both Anthropic and OpenAI publicly backed the laws — unusual for an industry facing regulation.
- The standards are voluntary, so no AI company is legally required to follow them yet.
Why It Matters
The rules shaping everyday AI are being written now, mostly by the companies that profit from it.