United Nations human rights chief Volker Turk has urged governments and leading artificial intelligence companies to take urgent steps to control the growing risks linked to advanced AI technology.
In a letter sent to countries and major AI developers, Turk warned that the world is approaching a major technological shift that could have long-term consequences for future generations. He called for international cooperation to establish rules for advanced and “frontier” AI models in line with international human rights law.
Turk’s warning comes amid an ongoing debate over whether the rapid development of artificial intelligence should be slowed down. US President Donald Trump and some of his key allies have opposed calls for a slowdown, while several AI leaders have raised concerns about the potential dangers of increasingly powerful systems.
The debate intensified after a researcher reportedly resigned from Anthropic over concerns that AI could eventually become impossible for humans to control. Another Anthropic employee also publicly warned about the possibility of highly advanced AI posing an extreme threat to humanity.
Anthropic CEO Dario Amodei recently called on AI companies to “pace the frontier” by coordinating the development of increasingly powerful AI systems. His proposal received public support from OpenAI CEO Sam Altman and xAI founder Elon Musk.
The UN rights office said Turk’s letter was sent to 10 major AI developers, including companies such as Anthropic, OpenAI, Meta, Microsoft and Google. Turk said relying only on a small number of powerful companies to make decisions affecting humanity was not enough and that stronger international regulation was needed.
Turk also highlighted the risks posed by increasingly capable AI agents, warning that such systems could potentially disrupt essential services, accelerate warfare and weaken safeguards designed to prevent the use of weapons of mass destruction.
He stressed that both governments and companies developing advanced AI have a responsibility to reduce these risks. According to Turk, no single country can effectively govern AI on its own, while no individual company should have the power to decide which risks the world must accept.
Meanwhile, China’s state-backed Global Times criticised Amodei’s proposal, describing it as a potential “Cold War” strategy aimed at China. The newspaper argued that although Amodei’s call focused on AI safety, it could also have wider geopolitical implications as the United States and China compete for leadership in advanced AI technology.
