Amid the intensifying debate over whether AI should slow down, OpenAI has made a landmark gesture: publicly backing federally mandated independent audits for the first time. On September 15, OpenAI's chief global affairs officer Chris Lehane said at a Washington roundtable with lawmakers that the company supports the core provision of the bipartisan FRONTIER Act requiring large AI developers to submit to independent safety assessments.
The FRONTIER Act, co-sponsored by Rep. Jay Obernolte (R-Calif.) and Rep. Lori Trahan (D-Mass.), would require frontier labs — the small set of developers with the most capable, most expensive models — to let independent "verification organizations" (IVOs) assess whether their safety protocols keep catastrophic risks within "acceptable" levels, and to publicly share model reports and safety issues. Lehane told reporters this is the first time OpenAI has publicly backed a federal mandate for third-party assessment — previously it supported only state-level regulation.
Even more notable was Lehane's other confirmation at the same briefing: OpenAI has been coordinating on AI safety with rivals Anthropic and Google DeepMind for several weeks, and the three firms believe they do not need an antitrust waiver to cooperate on safety matters. That moves Amodei's "industry coordination" from vision to actual negotiation. Lehane's framing was plainer: "It's better to try to work together to prioritize safety."
The road is far from smooth, however. Politico reports that while the AI lobbying world warms to mandatory independent audits, critics warn of "regulatory capture" — conflicts of interest between auditors and audited firms — and some lawmakers say the independent monitors would be too weak to impose meaningful change. President Trump, some venture capitalists, and academics also have reservations. In short, OpenAI's endorsement is just the first step; whether the bill passes a divided Congress remains an open question.
From a broader view, OpenAI's support for the FRONTIER Act echoes Anthropic CEO Dario Amodei's call to "embed evaluators" — despite their visible differences over IPO timing and safety narratives, frontier labs are converging, unusually, on the concrete mechanism of independent third-party oversight. The moment when AI governance moves from slogans to institutions may be arriving.