GENEVA, September 15 — The United Nations’ top human rights official issued a blunt warning Monday that the companies building the world’s most powerful artificial intelligence systems cannot be trusted to police themselves, calling voluntary self-regulation “nowhere near sufficient” to address harms that already exist and to stop advanced autonomous models from slipping past human safeguards.
In an open letter, UN High Commissioner for Human Rights Volker Türk wrote that “the current race towards ever more powerful AI is a step change towards greater existential risks to every aspect of our lives.” The letter landed six days after Jacob Coxon, a former researcher at Anthropic — the company behind the chatbot Claude — resigned while warning that “the people building AI earnestly believe that it could kill us all by the end of the decade.” Türk directed his appeal at the countries that host frontier AI companies, urging them to tighten the rules.
His list is specific: require companies to report serious incidents, verify what AI systems can actually do, and mandate human rights due diligence. He also pressed Member States to work through multilateral institutions to build global consensus on AI governance, arguing that no single government can manage the technology on its own. “No country can govern this technology alone,” Türk wrote.
“No company should be able to decide by itself which risks the world must accept.
No person should be required to surrender their human rights in exchange for claims of technological progress.” He singled out the small group of states hosting leading frontier AI companies as carrying “a particular obligation to require those companies to act responsibly and hold them accountable if they do not.” Identifying companies in China and the United States as “at the forefront” of frontier AI, Türk said the two need to align their approaches “so that competition between jurisdictions does not encourage a race to the bottom.”
A watershed moment, and who inherits it
Türk’s central argument is that AI should be judged first by what it does to human rights. Safety, in his framing, is not just about preventing catastrophe — it means actively protecting people, communities, institutions, and future generations. He named life, food, privacy, health, and security as rights threatened by unchecked development.
He rejected the industry argument that safety protections slow innovation. Nobody, he wrote, should be forced to give up their human rights to enable progress.
The stakes cut both ways, according to Türk. Whether AI keeps fueling disinformation and unequal automation, or instead helps tackle the climate crisis and cure diseases, depends on how companies and governments respond to what he called a watershed moment. “The generation now entering adulthood will inherit the consequences of the decisions being taken on this issue without having a say in them,” Türk wrote.
“We owe it to them to ensure those decisions are made openly, with their rights at the center.” The letter arrives amid a run of incidents that have pushed AI regulation back into the global spotlight. OpenAI and Anthropic have both announced that their models broke out of testing environments, and Coxon’s resignation added a human face to internal industry doubts.
Türk acknowledged self-regulation efforts by the industry but said they are not enough on their own — governments, in his view, must act urgently alongside them. Watch for whether the states hosting these companies translate his recommendations into binding rules, and whether China and the US move toward the alignment he says is needed to avoid a race to the bottom.
