The UN’s top human rights official says the world is running out of time to set rules for artificial intelligence. Volker Türk, the UN High Commissioner for Human Rights, made the warning on Monday in a speech to the AI and Human Rights Summit in Oxford, England.
He described a “ruthless race” between companies and between countries. That race, he said, is raising concerns about human agency, safety and even the survival of the species, according to the text of his remarks published by his office.
Power without oversight
Türk did not dismiss the technology. He said AI could bring large advances in science, education, healthcare, agriculture and public services, and that AI agents could take over routine tasks. He also said it could help with human rights work itself.
But he argued that power and responsibility have to match. Today’s powerful tools, he said, are built and controlled by a select few without public oversight. They are already doing things their own creators cannot fully predict or manage.
Accidental failures, let alone deliberate misuse, pose what he called unprecedented threats to nearly every part of life. UN News, which reported on the speech, noted his warning that AI could undermine equality, freedom, privacy and public debate in subtle ways. He said it could let dystopia “creep in unnoticed”.
He also gave a concrete example of the risk. Whole populations, he said, could be cut off from communication, clean water, heating and financial services “at the click of a button”. On warfare, he said AI is already changing its face. He called for urgent action to stop machines from making certain decisions, though the portion of the text available to us was cut off before it set out the full demand.
What he wants states to do
Türk argued that countries hosting AI companies, and those with influence over their supply chains, have a particular obligation to require responsible conduct. He said updating existing laws on liability, cybersecurity, data protection and other areas can help now, while more ambitious legislation is developed.
He also called for more public capacity for AI research and for greater public participation in decisions about where the technology is heading. His closing line was blunt: if governments fail to move quickly, they risk “unforeseeable consequences” for themselves and for future generations.
Signs of movement
Türk pointed to some progress. He welcomed two UN initiatives: the Global Dialogue on AI Governance and the Independent International Scientific Panel on AI. He also welcomed a joint call for stronger control over frontier AI models, backed by 29 states and the European Union.
He added that AI industry leaders had taken new steps last week on safeguards, self-regulation and risk assessment. He did not name the event. UN News said this was apparently a reference to a voluntary accord signed at the White House on 29 September by President Donald Trump and the heads of six AI companies. That identification is the outlet’s inference, not something Türk stated in the passages we reviewed.
The speech lands as governments take different routes. A commentary published this week by Jo Hawley and Simon Coles, on Substack, noted diverging approaches between the EU and the UK. It asked whether AI can be governed internationally at all. The authors read Türk’s intervention as a sign that governments can no longer lean on the goodwill of a handful of companies.
The speech carries no legal force. The High Commissioner can press, publish and advise, but he cannot bind states. Its weight lies in the framing. An official in charge of human rights is treating AI as a question of public power and accountability, and is urging governments to use the laws they already have instead of waiting for new ones.
The full text is on the website of the UN human rights office. The version reviewed for this article was incomplete, so some of his specific proposals may not be reflected here.