OpenAI’s global policy chief Chris Lehane confirmed on Tuesday that the company has been coordinating on artificial intelligence safety with rivals Anthropic and Google DeepMind for several weeks. The disclosure came during a Washington briefing and was first reported by Bloomberg.

Lehane said the three organizations do not believe they need an antitrust waiver to work together on safety matters. He pointed to long-standing industry practice in other sectors where competitors share information on safety without triggering competition concerns.

The confirmation follows Anthropic CEO Dario Amodei’s Saturday essay calling for deliberate restraint on frontier model releases so that safety systems can keep pace. OpenAI CEO Sam Altman publicly supported the broader idea and said OpenAI would embed third-party evaluators to monitor safety.

Key takeaways

What the multi-week talks actually cover

Public details remain limited. Lehane described the work as prioritizing safety and exploring industry-led standards. Earlier reporting indicated discussions around a possible standards body, an idea previously floated by Google DeepMind CEO Demis Hassabis.

Primary sources include the Bloomberg account of the Lehane briefing and Amodei’s published essay. Coverage from TechCrunch and other outlets has tracked the sequence of statements from Altman, Hassabis and Elon Musk supporting cooperative safety steps.

What to watch next

Any formal announcement of shared evaluation protocols or a voluntary standards framework would mark the next concrete step. Watch whether the three labs publish joint principles or invite additional participants beyond the current trio.