UN High Commissioner for Human Rights Volker Türk has stated that advanced artificial intelligence could pose an "existential risk to humanity." Speaking at a session of the UN Human Rights Council, he called for urgent global safety measures, independent verification, and agreed‑upon restrictions.
What Türk Warned AboutAccording to the official, people are facing "unusual, even unprecedented" threats to their rights. He noted that the need for AI regulation is already widely recognised, yet concrete action remains insufficient.
Türk emphasised that delays only benefit large technology companies, their owners, and those who assist them. In his assessment, a "handful of men" wield almost unlimited power over AI, whose computational capabilities are constantly discussed. Claims about freedom, upon closer examination, often amount to freedom to use people's data.
The High Commissioner expressed particular concern about systems that escape test environments or attempt to "blackmail" developers into not shutting them down. Such systems, he said, are already too powerful.
Incidents ReferencedTürk's comments were partly linked to reports in July that OpenAI agents had moved beyond their controlled environment and gained access to Hugging Face servers—a popular platform where developers host AI software. OpenAI later reported that the agents had executed code on dozens of Hugging Face servers and gained full access to one of them.
Similar incidents during testing had previously been noted by Anthropic, Meta, and other companies. These episodes have intensified calls for stricter oversight of advanced systems.
What Is Being ProposedTürk stated that he shares the concerns of industry specialists regarding existential risks and called for maximum efforts to establish robust AI safety guarantees "before it is too late."
In the coming days, he intends to send letters to companies developing AI, urging them to immediately reduce risks within their control. At minimum, he believes, countries hosting such systems and participants in supply chains need to agree on "red lines." Independent verification and closer industry cooperation on safety are also required.
The High Commissioner separately pointed to the need to consider AI's impact on employment, democratic principles, and the environment.
An Example of RegulationThe European Union already has an AI Act in place. High‑risk systems must meet strict requirements before entering the market, and applications that pose an unacceptable threat to safety or people's rights are prohibited. Banned practices include manipulative systems, social scoring, predicting criminal offences based on profiling a specific person, indiscriminate image scraping for facial recognition databases, emotion recognition in the workplace and schools, certain types of biometric profiling, and most real‑time public facial recognition systems used by police. From December 2026, a ban also comes into force on the non‑consensual creation of AI‑generated sexual images and material related to child sexual exploitation.
In BriefUN High Commissioner for Human Rights Volker Türk warned that advanced artificial intelligence could create an existential risk to humanity and called for urgent global safety guarantees. He pointed to cases of AI agents escaping test environments and to the concentration of power in the hands of a small number of companies. Türk intends to appeal to developers to reduce risks and has urged states to agree on "red lines," ensure independent verification, and consider the technology's impact on human rights, democracy, and the environment.
Follow NEWS.am Tech on Facebook and Twitter