OpenAI CEO Sam Altman cautioned Tuesday that "very subtle societal misalignments" around artificial intelligence could have dangerous consequences.

Why it matters: Altman has been a vocal proponent of government regulation of generative AI and has previously warned of the new technology's rapid evolution.

What he's saying: While speaking virtually at the World Governments Summit in Dubai on Tuesday, Altman was asked what AI-related fear is keeping him up at night.

• "I'm not that interested in the killer robots walking on the street direction of things going wrong," Altman said. "I'm much more interested in the very subtle societal misalignments, where we just have these systems out in society and through no particular ill intention, things just go horribly wrong."
• Altman reiterated his call for the creation of an international body like the International Atomic Energy Agency to monitor AI.

Altman also said generative AI has the potential to improve lives.

• "I think we can raise the standard of living so incredibly much if everybody has access to abundant amounts of really high quality intelligence and they can use ... those tools to create whatever they want to do," he said.

Our thought bubble, from Axios' Ryan Heath: Altman is a master of describing potential havoc from the most powerful AI systems, while keeping regulators from getting under OpenAI's hood.

• Eventually though, Altman may provoke enough fears that he turns off his potential customers, or gets the regulatory scrutiny he invites.

Go deeper: Welcome to the generative AI election era