OpenAI and Anthropic chiefs warn the UN Security Council about AI

For the first time, the heads of the largest AI companies have personally warned the UN Security Council about their own technology. On 24 September, OpenAI chief Sam Altman and Anthropic chief Dario Amodei spoke to the member states – and offered to help shape stronger regulation.

Altman's time-lapse

To make the pace of development tangible, Altman looked back three years: at the time, AI had been "quite OK" at solving elementary-school math problems. In the summer of 2025, OpenAI's AI reached gold-medal level in one of the world's most prestigious mathematics competitions. "Today it is even capable of solving so-called Millennium Problems", said Altman.

The specific warnings

• Speed: "AI could develop so quickly that no human can follow or intervene when needed", Altman warns.
• Power: Both see the danger of concentration in the hands of a few companies or countries.
• Misuse: Amodei additionally warns that terrorists could use AI to build biological weapons: "If handled improperly, artificial intelligence could become a risk to all of humanity."

The opposing position: Trump

Almost at the same time, US President Donald Trump took a contrary stance at the United Nations: he considers the warnings "disinformation with malicious intent" and wants to call AI only "superintelligence" in the future – with the message: "Whoever wins the AI race, wins overall."

Is international regulation on the horizon?

About 20 heads of state and government are calling on the UN member states this week to examine an international institution that sets AI standards and makes their compliance verifiable. At the same time, a so-called "emergency switch" for AI is being discussed, while the high-risk rules of the EU AI Act have been postponed again. The USA and China meanwhile are mainly using the debate for the fight over their own lead.

Sources: tagesschau.de, 24.09.2026 · tagesschau.de, 22.09.2026