OpenAI has confirmed that it has been in talks with Anthropic and Google DeepMind on AI safety for several weeks. The discussions involve the three leading AI labs and are described as ongoing, though no joint framework or public commitments have been announced. The talks come as the Trump administration dismisses safety concerns and emphasizes the need to keep pace with China in artificial intelligence. OpenAI’s confirmation followed earlier reports of the meetings, which have not included formal regulatory proposals. The companies have not disclosed the specific topics or outcomes of the conversations.
This is how progress happens. Not in a single lab, but in a quiet room where three rivals admit they share the same risks. OpenAI, Anthropic, and Google DeepMind talking for weeks? That is not a PR stunt. That is engineers and policy leads realizing that safety is not a race to the bottom. It is a race to the top. The Trump team wants speed. Fine. Speed without coordination is just chaos with a launch date.
I am not naive. These companies compete hard. But they also know that one bad headline, one uncontrolled release, can set the whole field back a decade. So they talk. They align on red lines. They build trust before the crisis, not during it. That is maturity. That is how you keep the future from being written by whoever panics first. The silence from Washington is loud. But the silence from these three labs is louder. It says: we are not waiting for permission to be careful.