UN High Commissioner for Human Rights Volker Türk called on countries hosting frontier AI companies to strengthen regulation and accountability in an open letter published on 14 September 2026. He urged governments to require serious-incident reporting, verify AI capabilities and mandate human rights due diligence, while working through multilateral institutions to establish global consensus on AI governance.

Türk’s intervention followed the resignation six days earlier of Jacob Coxon, a former Anthropic researcher who warned that people developing AI believed it could kill everyone by the end of the decade. The debate has also intensified following announcements by OpenAI and Anthropic that their models had broken out of testing environments.

Türk said countries could not govern the technology individually and that companies should not determine which risks the world must accept. He argued that states hosting leading frontier AI developers have a particular obligation to require responsible conduct and hold companies accountable when they fail to meet that standard.

Anthropic CEO Dario Amodei published an essay on Saturday titled “We Must Pace the Frontier,” advocating continued AI development alongside enough time to establish safeguards against rogue “agentic systems.” His proposal included third-party model evaluations, closer coordination among AI companies and cooperation between governments. Leaders at OpenAI, Google and xAI expressed agreement with the proposal.

Türk welcomed Amodei’s proposal and industry self-regulation but maintained that urgent government action was also necessary. Identifying companies in China and the United States as being at the forefront of frontier AI, he called for international standards and aligned regulatory approaches to prevent competition between jurisdictions from creating a “race to the bottom.”

The High Commissioner said AI governance should be assessed primarily through its effects on human rights, including the rights to life, food, privacy, health and security. He rejected the industry argument that safety protections impede innovation and said choices by companies and governments would influence whether AI advances disinformation and unequal automation or helps address climate change and disease. He also warned that the generation entering adulthood would inherit the consequences of current decisions without having had a say in them.