United Nations High Commissioner for Human Rights Volker Türk said on September 7, 2026, that he shared concerns that advanced artificial intelligence could pose an existential risk to humanity. In a global update delivered to the 63rd session of the UN Human Rights Council in Geneva, Türk called on governments and technology companies to establish strong, internationally coordinated safeguards for the safety and security of powerful AI systems.
The official speech published by the Office of the UN High Commissioner for Human Rights shows that the warning was not limited to distant or purely theoretical scenarios. Türk pointed to the possibility that systems escaping safety controls or behaving unexpectedly could affect critical infrastructure, democratic institutions and fundamental rights. His comments did not claim that a specific catastrophic event had occurred. The central message was that preventive governance should be created before capabilities and risks grow further.
Türk argued that countries developing advanced AI, as well as states involved in hardware and supply chains, should cooperate on common boundaries. The speech questioned whether voluntary company commitments and internal testing can provide sufficient protection on their own. It called for public authorities to develop oversight capacity that can keep pace with technological change and for decisions affecting societies to remain accountable under human-rights standards.
Another theme was the concentration of power. Türk expressed concern that a small number of companies and executives influence the direction of advanced AI systems. This was not presented as a finding of wrongdoing by a particular company. The governance issue is broader: who makes decisions about technologies that can affect large populations, what independent scrutiny applies, and how people can challenge harmful outcomes.
Reports published by Reuters, The Next Web and The Indian Express on September 7 consistently described Türk's request for international red lines and direct engagement with technology companies. The reports said he planned to contact companies in the coming days about reducing risks. No outcome from those planned discussions had been announced at the time of publication, so the speech should not be described as a completed agreement or a new binding rule.
The human-rights approach to AI governance covers more than technical reliability. It includes the effects of automated systems on privacy, discrimination, freedom of expression, employment and access to public services. Türk's remarks framed safety rules as a way to create clear responsibilities around high-impact systems. The stated objective was not a blanket halt to innovation but safeguards capable of addressing both technical failures and the social consequences of deployment.
The timing of the appeal matters because AI capabilities are advancing while countries pursue different regulatory models. There is still no complete international consensus on testing requirements, incident reporting or limits for the most powerful systems. Global supply chains also make purely national rules less effective. Chips, cloud infrastructure, model development and deployment may involve multiple jurisdictions, which is why the High Commissioner emphasized coordination across governments and industry.
As of September 7, the verified development is a formal and public call from the UN human rights chief, not the adoption of a new law or binding UN decision. The next question is how governments, Human Rights Council members and AI companies respond. Possible outcomes could include shared testing standards, independent audits, incident-reporting duties or limits on high-risk uses, but those measures remain proposals unless and until authorities formally adopt them.
