Rivals talk safety; antitrust lawyers take notes

OpenAI confirmed Tuesday that it has been working with rivals Anthropic and Google DeepMind on AI safety for weeks, confirming Bloomberg reporting from global policy chief Chris Lehane. The disclosure follows Dario Amodei's essay over the weekend calling on the industry to slow the frontier and coordinate on catastrophic risk, an appeal Sam Altman and Google's Demis Hassabis publicly backed.

The coordination is legally delicate: rivals discussing how to slow development sit close to antitrust lines, and Amodei proposed a narrow government waiver. Lehane said the firms do not need one. The Information reported the three are working toward a standards body for the industry.

OpenAI also endorsed a provision of the FRONTIER Act that would force top labs to admit independent verification organizations. The White House, for its part, continues to dismiss the risk debate, with the president calling concerns a hoax.