OpenAI says US AI policy is being written in state capitals
OpenAI's top policy executive says America's AI rulemaking is happening in statehouses, not Washington — and California's transparency law is the template.
Ann O'Leary, OpenAI's vice president of global policy, told POLITICO's California Agenda summit that US AI policy is "really happening in state capitals," with California's AI transparency law — signed by Gov. Gavin Newsom last year — informing much of the emerging framework. Speaking in Sacramento on Tuesday, O'Leary noted that thousands of AI bills are moving through state legislatures and that the policy framework is being built "through either executive orders or the state legislation." The next wave of conversations, she said, covers catastrophic risks, kids' safety, and data centers — and "there's real room for California to keep leading on these issues."
The statement is a notable posture shift for a company that spent last year opposing the state's landmark transparency law, the first in the nation to impose safety-and-transparency rules on frontier AI companies, including requirements to identify critical risks and report adverse events to the state. OpenAI's relationship with Sacramento has been rocky: earlier this year, Politico reported, a California organizer pulled AI ballot measures citing what he described as OpenAI intimidation. O'Leary, a former chief of staff to Newsom, now publicly frames California's law as the anchor of national policy — an acknowledgment that with no comprehensive federal AI statute, states are setting the compliance baseline every frontier lab has to meet.
O'Leary also conceded the industry's most visible local fight. "When you have a data center in your backyard, it's not inspiring you. And so I think we have real work to do," she said, as state Sen. Steve Padilla (D) advances a bill to ensure AI infrastructure "doesn't exploit and drain from communities." The data-center backlash is the issue where state and local politics most directly collide with the AI buildout — and where OpenAI is now signaling it expects to negotiate, not dictate.
What to watch: whether California's next round of rules — on catastrophic risk and data centers — becomes the model other states copy, as the transparency law did.
California's transparency law is becoming the national template for AI rules — is that a good thing? Tell us in the comments.
