The chief executives of OpenAI and Anthropic took to the United Nations stage Wednesday to warn about AI safety risks and call for international cooperation to address them.

"It doesn't matter whether people put the risk of catastrophe at 10% or 1% or 12% or 0.1%," said OpenAI chief executive Sam Altman. "None of these levels are remotely acceptable."

Dario Amodei, the Anthropic chief executive whose Sept. 12 "Pace the Frontier" essay rocked the tech world, urged the U.N. Security Council to establish "narrow agreements" to garner broad support from member states.

Altman and Amodei diverged from President Trump, who said attempts to regulate AI amount to a "globalist scheme" in Tuesday remarks in front of the U.N.

"We have unilaterally slowed down in the past," Altman said. "We will do so in the future."

Amid the hubbub, Anthropic's IPO plans have shifted to November. Altman confirmed in a Sept. 13 Fortune interview that OpenAI won't go public in 2026.