Artificial-intelligence companies risk losing control of AI systems and shouldn't wait for governments to regulate the technology, OpenAI Chief Executive Sam Altman has said, in his most detailed comments yet on the dangers of rapid AI development.

Altman warned that the dizzying pace of progress could go "very badly" and that humans could lose "control of the future to AI" or that too much power could become concentrated in the hands of a single person or company.

"We welcome a federal framework that sets consistent safety requirements for frontier AI," Altman said in a post on X overnight. "No amount of American competitive pressure should justify recklessness," he said.

Altman said that "pacing" the advancement of AI tools didn't mean stopping. "Progress has been rapid and will continue to be," he wrote. "But it should be slower than it otherwise could be; interventions like safety cases and monitoring have significant costs." He has suggested that OpenAI may need to delay its much-anticipated initial public offering to focus on safety.

Altman's latest comments come after Anthropic CEO Dario Amodei urged the industry to slow down. His warning came days after one of his employees stoked fears that there was a greater than 10% chance that AI systems could destroy civilization.

Elon Musk, whose xAI developed the Grok chatbot, over the weekend joined Altman in backing Amodei's call for caution. Amodei also warned that China could use advanced AI to undermine free societies.

China pushed back Monday against claims that its development of AI was creating threats. "Spreading threat narratives and engaging in confrontation and malicious competition will only derail global AI governance," Foreign Ministry spokesman Guo Jiakun said at a regular news briefing.

The debate over the rapid progress of AI development was sparked following a series of hacks by AI agents in recent months.

Two months ago, an AI swarm broke out of a lab and went on a hacking spree. Amodei said a swarm with greater capabilities and "a similar level of misalignment could have caused catastrophic damage." Such a rogue swarm could take over the internet in six to 12 months, he warned.