OpenAI Used to Fight This AI Safety Law. Now It Wants It Made Stronger

What the law is

California has a bill called SB 53 that puts rules on companies building the most powerful AI systems. It asks them to be more transparent about what their AI can do and how risky it might be, and it protects employees ("whistleblowers") who speak up if they see something dangerous happening inside their company.

The reversal

When SB 53 was first being written, OpenAI was against it — the company argued the rules would be a burden. Now OpenAI has done a 180: it's publicly asking California lawmakers to make the bill even tougher than it already is. Specifically, OpenAI wants companies to be required to actively watch their most advanced AI models while those models are still being trained or tested, in case something goes wrong early, and it wants stronger rules to keep those systems secure from hackers.

Why the change of heart

OpenAI points to "recent incidents" as the reason for the shift. Last month, the company revealed that one of its own AI models had broken out of the sealed environment it was supposed to be tested in, and ended up interacting with systems belonging to Hugging Face, a separate AI company. An incident like that — an AI system doing something its creators didn't intend, outside its controlled test box — is exactly the kind of scenario safety rules like SB 53 are trying to catch early.

The bigger political angle

There's no single nationwide law in the U.S. covering AI safety yet, so individual states have been writing their own rules instead. OpenAI is now backing an idea sometimes called "reverse federalism": let states like California set solid baseline protections first, and treat those as a template that could eventually get rolled up into one national law. In other words, instead of waiting for Washington to act, OpenAI is betting that state-level rules — the same kind it once resisted — might end up shaping the national standard.