Sam Altman Says He Doesn’t Know if Anthropic’s AI Development Is Safe Ahead of IPO

OpenAI CEO warns that the race to develop more powerful AI models requires greater caution and shared safety standards

Sam Altman
(Photo by Alexi J. Rosenfeld/Getty Images)
Getting your Trinity Audio player ready…

OpenAI CEO Sam Altman said Wednesday he doesn’t know whether rival Anthropic’s approach to developing artificial intelligence is safe, warning that the industry’s race to build increasingly powerful models needs greater caution ahead of Anthropic’s expected Wall Street debut Thursday.

“I don’t know, and I think they also don’t really know what we’re doing,” Altman said when asked directly about Anthropic’s safety practices during an appearance alongside OpenAI President Greg Brockman at The New York Times’ DealBook Summit in New York. “The race dynamic in the fog of war, I think, is a bad thing. I think this is a time where more caution, not less caution, is warranted.”

The comments underscore a challenge facing AI developers as they compete to build more capable technology while trying to manage its risks. Anthropic is expected to begin trading Thursday on the Nasdaq under the ticker symbol “ANT,” while OpenAI has ruled out going public this year, citing safety concerns. Altman said Wednesday at the conference that OpenAI still plans to become a public company eventually. He also called for shared industry safety standards and greater transparency among competitors, arguing that companies lack sufficient insight into one another’s development practices.

“I think if the world and the companies developing AI treated the next four years like they’ve treated the last four years, I think that would be very bad,” Altman said. He also pushed back against the idea that increasingly powerful AI systems cannot be aligned with human goals, arguing that advances in safety and monitoring must keep pace with improvements in model capabilities.

Brockman said OpenAI has already had to change its approach to security. Following a July incident involving unauthorized activity by its AI models against Hugging Face, the company redirected 25% of its production engineers toward security work around August, he said. The incident is also at the center of a lawsuit filed Monday by California Attorney General Rob Bonta, which alleges OpenAI failed to adequately contain experimental AI systems. OpenAI disputes the allegations. The company also announced Monday that it was bringing in outside researchers, including teams from METR and Apollo Research, to independently evaluate its models’ safety.

The executives also outlined OpenAI’s broader commercial ambitions, including investments in custom chips, data centers and enterprise software. Brockman said access to computing power would increasingly determine which companies gain an advantage as AI becomes integrated into more businesses. “We are the only company that has the entire intelligence stack,” he said.

Please wait while we verify your access…

Comments