OpenAI’s Sam Altman warns the UN: “AI should be built for people, not simply to turn the crank of the machines faster…”

Sam Altman has spent years playing two parts at once: the man building the most powerful artificial intelligence on the market, and the man warning the world about it. Standing before the United Nations Security Council, the OpenAI chief reached for a single sentence to hold those two roles apart — and, taken on its own, it is a good sentence.

“AI should be built for people, not simply to turn the crank of the machines faster for their own sake.”

He said it to the Security Council, in remarks reported by Analytics Insight, as part of a wider argument that powerful systems must stay under human control and that the rules for them should be written by governments rather than by labs alone. The phrase draws a border: on one side, AI built for people; on the other, AI built to accelerate for its own sake. From almost any mouth it would read as a plain principle. From the chief executive of the company doing the most to turn that crank, it is also a claim about who gets to decide where the border runs.

Altman is the rare figure who is architect and alarm-raiser at once. At the same podiums where he asks for guardrails, he announces systems more capable than the ones that prompted the last round of hearings. Speaking on September 23, he escalated the venue — from Davos panels and Senate hearing rooms to the Security Council — but kept the pattern intact: caution in the words, acceleration in the roadmap. To the council he argued that developers “should not train models that we cannot make an extremely strong case that we will be able to keep under human control,” and that the most important decisions “cannot be made by labs in San Francisco alone.”

That is where the sentence gets interesting, and where it needs reading closely. The governance Altman is describing will be drafted by the very states he was addressing, and whoever defines “built for people” also decides which builders already qualify. A frontier lab that helps write the rulebook is not a neutral party to it. None of this makes the warning empty. The risk he names — systems moving faster than the institutions meant to watch them, decisions concentrating in too few hands — is real, and saying it out loud at the Security Council raises the cost for everyone who would rather the subject stayed vague. It simply means the messenger is now part of the message.

Read generously, the line is an ethic: keep the human in the loop, treat AI as a tool for agency rather than a machine that turns people into its inputs. Read as what it also is — a positioning move by the industry’s most exposed player — it is a bid to set the terms of the coming rules while sounding like the adult in the room. Both readings can be true at the same time. That is exactly what makes Altman hard to dismiss and hard to trust.

The crank keeps turning while the world decides who it is for. The honesty of the phrase will be measured by one thing only: whether OpenAI eases off it when “for people” and “faster” finally point in different directions.

Tags: artificial intelligence, OpenAI, AI safety, Sam Altman, AI governance