The UN Faces a New Test in Artificial Intelligence

Guterres Warns AI Is Outpacing Global Governance

By Ahmed Fathi

UNHQ, New York: When I asked the UN Secretary-General whether governments still control technology powerful enough to shape elections and conflicts, his response was unusually direct. The United Nations, he said, does not have leverage. It has mechanisms, platforms and processes, but it does not have the power to force outcomes.

That moment mattered less because it was surprising than because it was honest. It put into words what many governments and institutions quietly acknowledge: artificial intelligence is advancing faster than the political and legal systems meant to oversee it.

What followed was not a pitch for sweeping regulation, but a clear-eyed description of limits. The Secretary-General laid out what the UN can realistically do in the AI space. It can bring experts together, set up scientific panels, produce assessments and convene global conversations. These are serious efforts, but they stop short of enforcement. They form the scaffolding of governance without the authority to impose rules.

This gap sits at the center of the global AI debate. The UN is trying to build a shared understanding and a common language around artificial intelligence, not a binding rulebook with penalties. That choice reflects political reality more than institutional weakness. States are reluctant to give up control in an area tied directly to national security, economic competition and strategic power. As a result, the UN is left shaping the discussion rather than determining the outcome.

The gap between ambition and authority has been hard to miss in recent UN efforts on artificial intelligence. Statements and resolutions have emphasized ethics, inclusivity and cooperation, while deliberately avoiding commitments that might bind powerful states or companies. What this produces is a system that can highlight risks and shape norms but has little ability to force change when the most influential players choose to resist.

When pressed on what meaningful guardrails should actually look like, the Secretary-General focused on one core idea: human agency. Humans, he argued, must remain in control of decisions, especially when lives are involved. His rejection of autonomous weapons that can decide who to kill, where and why was one of the sharpest moral lines he drew during the press conference.

His answer on the growing power of technology companies was equally revealing, if more understated. Regulation, he suggested, is ultimately the responsibility of governments. Existing laws on competition and monopolies may need to be updated for the digital age, but enforcement will not come from the UN Secretariat. If it comes at all, it will come from national capitals.

Beneath these exchanges was a broader point about how power itself is evolving. Power today is no longer measured only in territory, military strength or even economic size. It is increasingly defined by data and by who controls the systems that gather, process and deploy it. That shift is steadily pulling influence away from public institutions and into the hands of private actors, moving faster than global rules and institutions can realistically keep up.

It raises more than regulatory concerns. It raises questions of legitimacy. When choices that shape entire societies are increasingly made by algorithms designed beyond democratic scrutiny, familiar ideas of accountability start to erode. The Secretary-General did not frame this as a plot or a moral collapse, but as a structural imbalance that today’s institutions are still struggling to understand, let alone correct.

Layered onto this is an unresolved equity problem. Many countries simply do not have the resources or capacity to shape AI policy or share fully in its benefits. Without steady investment in skills, infrastructure, and institutions, global conversations might end up dominated by those who already hold the most power. Governance without resources does not close gaps; it widens them.

Taken as a whole, the exchange offered a clear snapshot of where the UN stands on artificial intelligence. It remains a place where risks are named, values are articulated and norms are debated, but not a place where outcomes are enforced. The Secretary-General did not try to disguise that reality, and his candor stood out.

The unresolved question is whether a system designed to convene, warn and advise can evolve into one that actually shapes behavior. For now, the UN’s approach to AI reflects a broader truth about the global order: power is moving faster than law, and institutions are struggling to catch up.