Three companies that spend most of their time trying to out-compete each other have quietly been coordinating behind the scenes for weeks — a detail that only became public well after the collaboration itself began.
What Was Actually Revealed
OpenAI’s global policy chief Chris Lehane confirmed that OpenAI, Anthropic, and Google DeepMind have been in talks on AI safety for weeks, according to TechCrunch’s report on the disclosure, first reported by Bloomberg. Lehane made the comments while in Washington meeting with lawmakers specifically about catastrophic AI risk.
Why This Timing Matters
This revelation surfaced specifically in the wake of Anthropic CEO Dario Amodei’s Saturday essay calling for the industry to slow frontier AI development — meaning the private coordination between these three companies predates the public essay that triggered this week’s broader debate, suggesting the concern wasn’t purely reactive to Amodei’s letter.
The Specific Commitment OpenAI Made
Sam Altman said OpenAI would follow Anthropic’s lead in giving independent, third-party evaluators employee-level access inside the company specifically to monitor for safety issues — a genuinely significant level of external oversight for a company valued in the hundreds of billions of dollars, and a concrete action rather than just supportive rhetoric.
Where Government Fits Into This
Lehane specifically said OpenAI supports a provision in the proposed FRONTIER Act that would require top frontier labs to allow independent verification organizations into their operations. According to the same TechCrunch reporting, this puts OpenAI in the position of actively supporting a specific piece of pending legislation, not just voicing general safety concerns.
The White House’s Different View
President Trump has dismissed AI safety concerns as a “hoax,” and his AI advisor David Sacks has echoed that framing, arguing that fears of existential AI risk are overblown. This creates a genuine disconnect between the labs’ own internal coordination and the current administration’s public posture — the companies actually building the technology are coordinating more cautiously than the government overseeing them is currently signaling it wants.
Why Three Competitors Working Together Is Notable
OpenAI, Anthropic, and Google DeepMind compete directly for talent, enterprise customers, and model performance benchmarks — genuine, sustained cooperation between all three on anything is rare specifically because their commercial incentives usually point in opposite directions. This connects to the broader pattern of AI leadership positioning we’ve covered in how AI giants are dividing over safety more broadly, where Meta and Nvidia have taken a notably different public stance from this specific trio.
What This Means for Regulation Going Forward
Private, weeks-long coordination between three major labs, only surfacing publicly after a viral safety essay, suggests the companies themselves may be moving faster on internal safety measures than public statements alone would indicate. Businesses tracking [CLIENT LINK PLACEHOLDER] how frontier AI regulation is likely to evolve are watching whether this coordination translates into an actual, binding industry standard, or remains an informal, voluntary arrangement subject to change.
Frequently Asked Questions
Is this collaboration a formal, binding agreement between the three companies?
Based on current reporting, it appears to be an informal, ongoing dialogue focused on safety coordination rather than a formal, binding legal agreement between the companies.
What is the FRONTIER Act specifically?
It’s proposed legislation that would require top frontier AI labs to allow independent verification organizations access to monitor for safety compliance — OpenAI has publicly stated its support for this specific provision.
The Bottom Line
The revelation that OpenAI, Anthropic, and Google have been quietly coordinating on AI safety for weeks — before this week’s public debate even began — suggests genuine, if still informal, common ground exists between major labs on safety specifically, even as the companies continue competing aggressively on nearly everything else.