The alarm had already reached the U.N. a day before the Security Council convened, when Secretary-General António Guterres warned of “killer robots,” amid mounting questions over whether machines could be allowed to make life-and-death decisions.¹

On Wednesday, the Council heard a strikingly similar message from the companies building the technology. OpenAI chief executive Sam Altman framed the moment as a fork in the road: “AI can either be more like a new Renaissance of creativity and discovery, or more like a new Industrial Revolution of upheaval and disarray.”²

Altman’s case was not for stopping development, but for keeping its most powerful forms answerable to people. As systems become more autonomous and potentially self-improving, he warned, they could outrun institutions, pool influence in a handful of companies or countries, and make choices humans cannot understand or reverse. “We could lose control of the future to AI,” he said.¹

Anthropic’s Dario Amodei put the danger even more bluntly: “If managed poorly, I even believe AI could be a risk to humanity as a whole.” He called for safeguards strong enough to stop advanced tools being turned toward catastrophic misuse, including biological weapons.¹

The urgency was sharpened by recent security failures and disclosures of model misbehavior. Hugging Face chief executive Clem Delangue argued that the response must include shared monitoring and rapid disclosure: “the global community needs stronger standards for monitoring and incident disclosure.”³

The executives’ prescription was unusually specific: common tests for dangerous capabilities, verification and incident-reporting systems, and international standards that can coexist with national rules. The broad message from the meeting was that “robust international efforts to regulate the technology” are now required as concern over AI’s threat to humanity rises.⁴ Altman added a political warning of his own: decisions over democratic AI cannot be left to Silicon Valley labs alone.²