OpenAI, Anthropic, Google have been in talks on AI safety for weeks
TechCrunch Rebecca Bellan ● Covered by 27 sources
OpenAI says it’s been quietly talking safety with Anthropic and Google DeepMind for weeks. That’s a big deal because the same companies are also trying to avoid antitrust trouble.
Based on reporting by TechCrunch, Rebecca Bellan — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
OpenAI’s policy chief, Chris Lehane, said Tuesday that the company has been in talks with Anthropic and Google DeepMind for several weeks about AI safety. The admission lines up with earlier reporting, but it also puts names on something that had mostly been inferred from hints, essays, and a lot of carefully chosen language.
The timing matters. Anthropic chief executive Dario Amodei called for the industry to slow frontier AI development and reduce catastrophic risk in an essay published Saturday. Since then, a handful of the biggest names in AI have publicly nodded along: OpenAI’s Sam Altman, Google’s Demis Hassabis and SpaceXAI’s Elon Musk all backed the broader call to action. Altman also said OpenAI would join Anthropic in bringing in third-party evaluators to watch for safety issues.
Lehane was in Washington talking with U.S. lawmakers about catastrophic AI risks, and he said OpenAI supports a provision in the FRONTIER Act that would require top frontier labs to let “independent verification organizations” inside to check whether models are being built safely. He also said the companies do not need the narrow government waiver Amodei proposed for safety coordination.
But the legal cloud never really goes away. Some observers, including Altman, have warned that coordination among competitors could run into antitrust law if it starts looking like a way to suppress competition. That’s the awkward part here: the industry keeps saying it wants shared safety rules, while Washington is still deciding whether those rules should come from the companies, the government, or both. The Information also reported this week that the three firms have been working on a standards body for the AI industry, with Altman reportedly telling staff it would have to happen without U.S. government backing. And Hassabis has already called for a new AI watchdog with the power to review advanced models and coordinate slowdowns if risks rise.
My take — AI-written commentary, not fact-checked reporting
This is the AI industry doing what it always does when the heat gets high: asking for a referee after spending years arguing the game can regulate itself. Safety coordination sounds noble until it starts looking like club rules written by the clubs. If these firms really want trust, they should welcome outside checks instead of treating them like a temporary inconvenience.
Read more about this at: TechCrunch