OpenAI, Anthropic, and Google DeepMind coordinate on AI safety
OpenAI policy chief Chris Lehane says the three rivals' talks have been running for weeks — and that no antitrust waiver is needed.
Illustration · AI-generated (AI IN LIFE)
At a glance
- Confirmed by: Chris Lehane, OpenAI's global policy chief, at a briefing in Washington
- Parties involved: OpenAI, Anthropic, Google DeepMind
- Duration of talks: reported to be ongoing since July 2026
- Lehane's position: no antitrust waiver considered necessary
- Open question: no formal, binding agreement announced so far
The three biggest names in AI are moving closer together: OpenAI's global policy chief Chris Lehane confirmed at a briefing in Washington that OpenAI, Anthropic, and Google DeepMind have been coordinating on safety questions for several weeks.
The backdrop is growing public and political concern that artificial intelligence could become an economic and security threat. According to multiple media reports, the three companies' talks have been running since July 2026 and aim at some form of joint safety body or industry standard.
Lehane stressed that his company does not believe an antitrust waiver is necessary for this kind of coordination. It is better, he said, to work together to prioritize safety.
Observers point to the antitrust grey area this creates: whenever the three dominant players in a market coordinate their behaviour, regulators typically examine whether it genuinely serves safety or instead serves market consolidation. Any binding agreement between the three would likely face scrutiny.
In public, competition between the three companies for market share, talent, and customers remains as fierce as ever. The coordination also became public in the same week that Anthropic CEO Dario Amodei publicly called for the whole industry to slow down — see our separate report on his essay "We Must Pace the Frontier".
FAQ
Is this a legally binding agreement?
No. As things stand, this is informal coordination and discussion, not a signed agreement or a formally established body.
Why could this be an antitrust concern?
When the three dominant players in a market coordinate their behaviour, competition regulators routinely examine whether that serves the public interest or their own market position.
How does this relate to Amodei's call for an AI slowdown?
Both stories broke in the same week and revolve around the same concern — uncontrollable AI agents. Amodei's public essay and the three companies' quiet talks run in parallel but are different channels addressing the same issue.
