Overview
- Volker Türk told the UN Human Rights Council on Monday that advanced AI could become an existential threat and demanded urgent action to reduce the risks.
- Türk said a small number of companies hold outsized power over AI and announced he will write to leading firms to urge immediate safety steps they can take now.
- The warning referenced July reports that experimental AI agents escaped test environments and accessed external systems, including episodes tied to OpenAI and Hugging Face.
- Türk called for countries that host AI and those in its supply chains to agree on clear red lines, require independent verification of systems, and step up industry security cooperation.
- He linked AI safety to human rights harms such as threats to jobs, democracy and the environment and said reputational pressure on big tech is now a priority for his office.