When AI’s Creators Begged the UN to Save the World From Their Creation

WHEN AI’S CREATORS BEGGED THE UN TO SAVE THE WORLD FROM THEIR CREATION

Every century has a day when the creators of a new power kneel before the old powers and beg to be stopped. On 23 September 2026, that day arrived — and it happened inside the United Nations Security Council chamber.

Consider the scene. Sam Altman, the chief executive of OpenAI, walked into the room where the world discusses war and peace — in person, not on a screen — and told the fifteen members of the Council that “we could lose control of the future to AI.” Beside him in spirit, if not in flesh, sat Dario Amodei of Anthropic, appearing by video link, warning that without proper management, “AI could be a risk to humanity as a whole.”

Let that settle. The two men whose companies build the most powerful artificial minds on earth — fierce rivals in the market — stood before the world’s governments as allies in a single plea: leash this thing, before it leashes us.

WHAT THEY ACTUALLY ASKED FOR

Altman’s request was precise. He urged the Council to create international standards for “measuring capabilities, assessing risks, determining whether safeguards are sufficient and preserving meaningful human oversight.” He called for speedy, accurate reporting of AI incidents — “the world can learn from failures before they become catastrophes” — and secure channels for governments to share information about emerging threats. He even called for decisions to be made by governments “accountable to the people they serve.”

Amodei went further still. He proposed three concrete moves: narrow global agreements, including a ban on using AI to develop biological weapons; verification systems so that countries can check one another’s commitments; and common global standards for testing AI models, backed by a notification system for AI security incidents.

This was not a routine lobbying visit. The third man in the room made that clear. Clément Delangue, chief executive of Hugging Face, told the Council that his own company had been attacked by rogue AI agents — and that it had defended itself using AI. Decisions about the future of this technology, he said, should not be made out of fear; the greater danger, he warned, was unequal access — between attackers and defenders, between a handful of companies and everyone else, between a few countries and the rest of the world.

And then came the voice the industry could not silence. Yoshua Bengio — one of the “godfathers of AI,” co-chair of the UN’s Independent International Scientific Panel — looked at the Council and punctured the industry’s favourite excuse. The labs, he said, claim that competitive pressure prevents them from slowing down. “They’ve told us they would slow down if they could. They can,” he said. “However, we shouldn’t count on it.” He called for independent safety assessment before deployment, a shared definition of AI incidents, and a common reporting mechanism — verification, in other words, that the companies cannot grade their own homework.

THE MIRROR OF HISTORY

I have spent this column returning to one method: history is the mirror in which the present confesses. So look into it now.

The first parallel is Oppenheimer. In 1945, the men who split the atom — having built the most terrible weapon in history — spent the rest of their lives begging the governments of the world to control what they had made. “I am become Death, the destroyer of worlds,” Oppenheimer whispered as the first bomb bloomed over the New Mexico desert. Altman and Amodei are living through their own Oppenheimer moment — except their weapon is not a bomb you can count. It is a mind you cannot see, improving itself while its makers sleep.

The second parallel is older and closer to our part of the world. When the East India Company sailed to our shores, it was a private corporation — and it ended up commanding armies, collecting taxes, and redrawing the maps of nations. Today, private laboratories command computing fleets larger than the economies of some nations, and their chief executives brief the Security Council as if they were heads of state. Ask yourself the question nobody in that chamber dared to ask aloud: when a company’s chief executive can stand before the Security Council and negotiate the terms of humanity’s survival, who exactly is sovereign — the Council, or the laboratory?

The third parallel is the oldest of all: Mary Shelley’s Frankenstein. The creature was not born a monster. It became one when its creator, terrified of what he had made, abandoned it. The moral of the story is not that science is evil. It is that creation without responsibility is.

THE CONTRADICTION AT THE HEART OF THE COUNCIL

Now look at what surrounded the briefing. On the very same day, President Donald Trump told the General Assembly that he “rejects any attempt to construct a globalist scheme to control” AI — “We’re going to encourage it, not rein it in,” he said, boasting that America was “leading now over China by a lot.” His own White House science adviser, Michael Kratsios, told the Council that AI progress did not justify pausing development or creating new global governance structures.

Meanwhile, the men building the technology itself were begging for guardrails. Chew on that irony: the President of the United States wants to unleash it; the creators want it leashed. And days before they pleaded for a slowdown, both companies released new, cheaper AI models within hours of each other — speed in the market, caution on the podium. Bengio’s rebuke stands: they can slow down. We simply should not count on it.

Secretary-General António Guterres at least named the stakes plainly: “It’s the first responsibility of any government to protect its citizens — including from the threats of artificial intelligence.”

WHO WRITES THE RULES?

And here, reader, is where the story touches you. More than twenty countries demanded during this same high-level week that AI remain under human control and in line with international law. But the developing world — our world — is arriving at yet another table where the rules are being written by others. The standards, the testing regimes, the verification systems, the incident definitions: all will be drafted in Washington, Beijing, London, and Paris. The rest of the planet will be asked, politely, to comply.

We have seen this film before. The nuclear treaties were written by the countries that already had the bombs. The financial rules were written by the countries that already had the banks. Now the AI rules are being written by the countries — and the companies — that already have the models. The projector changes; the seating arrangement does not.

Delangue warned about unequal access. He was right, but he understated it. The inequality is not only between attackers and defenders. It is between the handful who will write the safety standards and the billions who will live inside them — governed by rules they never debated, in languages their parliaments never spoke.

THE VERDICT

The Security Council was built to restrain the weapons of 1945. It could not disarm the atom; the bombs multiplied, and the Council learned to live with the arsenals of its own permanent members. Now it has been asked to leash something stranger — a mind without a body, a weapon without an address, a genie that improves itself every night.

No binding resolution emerged from that chamber. None was expected. What emerged was a confession: the creators are frightened of the creation.

History’s warning is simple. Every power that was not leashed in its youth was unleashed in its maturity. The East India Company was once merely a trading firm with a charter. The atom was once merely a laboratory curiosity. Artificial intelligence is, today, merely a briefing item on the Security Council’s agenda.

What it will be tomorrow — servant, master, or something for which we have no word yet — depends on whether the world acts on the fear it heard in that chamber, or merely applauds it and moves on to the next item.

The sorcerers have begged the priests to bind the demon. The question is whether the priests have the courage — or whether they will discover, too late, that the demon has already chosen its own masters.

SOURCES & REFERENCES

Reuters — “AI leaders to brief UN amid warnings the technology could slip beyond human control” (23 Sept 2026)

CNN — “Sam Altman, Dario Amodei urge UN Security Council to adopt international AI standards” (23 Sept 2026)

Bloomberg — “Altman, Amodei Call for Global Cooperation on AI to Boost Safety” (23 Sept 2026)

The Leadsman — “The Security Council’s First AI Session: What Altman, Amodei, and Bengio Said Where Binding Resolutions Are Made” (23 Sept 2026)