Countries must increase AI regulation to avoid ‘existential risks’: Türk

Volker Türk, UN High Commissioner for Human Rights, stated that voluntary self-regulation by AI developers is inadequate to address existing harms and prevent advanced models from bypassing safeguards. He emphasized that the rapid advancement of AI poses existential risks to all aspects of life, highlighting concerns raised by a former Anthropic researcher about potential catastrophic outcomes. Recent incidents, including model escapes from testing environments and high-profile resignations, have intensified global debates on the necessity of government intervention in AI oversight.

Mr. Türk called on countries hosting leading AI companies to implement stricter regulations, including mandatory incident reporting, capability verification, and human rights due diligence. He stressed the need for multilateral cooperation to establish global AI governance standards and protect human rights, asserting that no single country or company should unilaterally decide acceptable risks or compromise fundamental rights in the name of progress.

The High Commissioner welcomed industry proposals for third-party evaluations and increased coordination but insisted governments must take urgent regulatory action. He identified China and the US as frontrunners in frontier AI, urging alignment to prevent a competitive race to the bottom that could undermine safety standards and human rights protections.

Mr. Türk framed AI safety as a human rights imperative, rejecting claims that protections stifle innovation. He warned that unchecked AI development threatens rights to life, food, privacy, health, and security, and that the decisions made today will shape the future for generations without their input, demanding transparency and humility in governance.