Altman: OpenAI IPO "ill-timed", not before 2027

Sam Altman's interview with Fortune Editor-in-Chief Alyson Shontell lands in a week when AI doomsday anxiety broke into the mainstream, set off by an Anthropic researcher's prediction that AI has a greater than 10% chance of killing all of humanity by 2030. Speaking in San Francisco for the Fortune 500 podcast, Altman said no gamble with humanity is OK and positioned himself as willing to face down investors if a pause or full stop became necessary.

The remarks arrive alongside Anthropic CEO Dario Amodei's Saturday announcement that his company will give independent evaluators permanent, employee-level access—a concrete safety measure tied to his call to slow the pace at which AI models improve. Altman and other AI leaders voiced support for that industry-wide slowdown, and Altman alluded to a pact with peers to ease off while safety and alignment solutions catch up to model capabilities.

What ties these threads together is uncertainty over control: Altman said the models are developing quickly and that it remains an open question whether more powerful AI can be controlled. The IPO timing and the slowdown pact both hinge on whether the safety side can keep pace—and for OpenAI's investors and customers, that is the risk to watch.