Lehane Says OpenAI Has Coordinated Safety With Rivals for Weeks

Classy aggregates publicly available material; we did not conduct a private interview.

What changed

On 15 September 2026, OpenAI global affairs chief Chris Lehane told reporters in Washington that OpenAI, Anthropic, and Google DeepMind have held safety coordination talks for several weeks. Lehane said the companies do not need an antitrust waiver to keep cooperating, and that OpenAI supports a FRONTIER Act provision requiring independent verification organizations to audit frontier lab safety practices.

The remarks follow Anthropic CEO Dario Amodei's 13 September essay calling on the industry to pace frontier development, and they turn weeks of private hints into an on the record admission that the three leading labs are aligning on safety mechanics, not just public endorsements.

Why it matters

Enterprise buyers, insurers, and regulators have been asking whether frontier lab safety pledges will survive competitive pressure and IPO timelines. Lehane's Washington briefing signals that the largest labs are willing to coordinate on verification even while racing on product, which changes how procurement teams should read vendor safety claims.

Supporting independent verification in federal legislation also moves the debate from voluntary slowdown letters toward enforceable third party access, a shift that could affect model release gates and enterprise contract language within the next legislative cycle.

Who is affected

Frontier lab policy and safety teams must now align internal roadmaps with multi company coordination that may be scrutinized under antitrust rules even without a formal waiver.

Enterprise AI buyers and risk officers should expect vendors to cite shared safety frameworks while still competing on capability tiers; verification clauses may become standard in RFP language.

Lawmakers and independent evaluators gain a clearer signal that at least one major lab will accept third party audits if the FRONTIER Act or related bipartisan proposals advance.

What to do next

If your organization procures frontier models, add a requirement for documented third party safety verification or an equivalent independent audit trail before production deployment on high consequence workflows.

What to watch

Whether OpenAI publishes specifics on the standards body Demis Hassabis proposed in July, whether the FRONTIER Act verification provision survives committee markup, and whether Anthropic and Google confirm Lehane's timeline on the record beyond secondary reporting.

Sources

1. Primary. Bloomberg, OpenAI Says It's Working With Anthropic, Google on AI Safety (15 September 2026). Lehane's confirmation of multi week talks and antitrust waiver stance.
2. Secondary. CBS News, OpenAI backs measure that would require independent audits of AI models (15 September 2026). FRONTIER Act verification support and Washington roundtable context.
3. Secondary. PYMNTS, OpenAI Defies Antitrust Concerns to Join Rivals on AI Safety (15 September 2026). Independent verification organization language and lawmaker meetings.