AI Chiefs at UN: Humans Must Stay Central to AI Decisions

What happened

The chiefs of OpenAI and Anthropic told world leaders at the UN on Wednesday that people must stay at the centre of decisions about artificial intelligence. Sam Altman and Dario Amodei addressed the UN Security Council during the General Assembly in New York, according to reporting by The National.

The appearance comes as both companies have driven the AI investment boom, while public sentiment towards the technology has been gradually souring. Altman cautioned against moving too far towards regulation, but he agreed that some international standards should exist.

Altman told the council that "as AI becomes more capable, people must remain at the centre of AI decision-making." He also rejected the view of "doomers" who paint a bleak future, arguing instead that the aim should be to give people more agency in a more complex world.

Why it matters

The core question in this exchange is not whether AI will be used, but where decisions about it are made and who keeps control over them. For organisations deploying AI in products, services and public administration, "human at the centre" is a design requirement that has to be specified, not a principle to be stated at a conference. It shapes where automated recommendations end, where a person can review or override them, and how clearly customers and staff are told when an algorithm is involved.

The reporting also points to a shift in how the AI debate is being framed. Altman's emphasis on agency and the ability to learn, rather than on optimising every part of life, puts the conversation closer to experience design and to the work of service teams. That is where trust is either earned or lost, and where public sentiment is most likely to be shaped.

The Renascence take

Most of the coverage will focus on regulation versus self-governance. The more practical question for operators is less abstract: at which points in a journey does a person actually decide, and can they tell that they are the one deciding?

"Keeping people at the centre" is a slogan until it is built into escalation paths, override controls and plain-language disclosure. Most organisations have never mapped the moments where an AI output quietly becomes a final decision for a customer or employee. The operators who win on trust will be the ones who audit those moments first, redesign them so a human can intervene without friction, and measure whether people feel in control rather than assuming they are.