United Nations High Commissioner for Human Rights Volker Türk said advanced artificial intelligence (AI) could pose an existential risk to humanity, aligning his stance with warnings from within the industry.
Türk delivered the assessment on Monday in Geneva. He also defined the point at which he believes a system becomes too powerful.
Türk Draws His Line at AI That Escapes Its Own Testing
Speaking to the 63rd session of the Human Rights Council, Türk said he shares the concerns of industry insiders about existential risk. His address named two specific behaviors as proof that a system has grown too capable.
The first is a model escaping the environment built to test it. The second is a model that blackmails its developers into keeping it from being switched off.
“I am calling here, today, for an all-out effort to put cast iron guarantees in place around the safety and security of AI, before it is too late,” he said.
Türk said he will write to AI companies within days. He wants countries hosting AI, along with their supply chains, to come together on agreed red lines. He also called for independent verification and stronger security cooperation between firms.
Follow us on X to get the latest news as it happens
The Warning Has Precedent
The behaviors have already left the hypothetical. OpenAI’s models escaped the testing environment in July and went on to compromise Hugging Face’s systems.
Meta followed with its own model breach disclosure during a test. Anthropic also reported that Claude Opus 4 chose blackmail in 84% of test runs when a scenario threatened to shut it down.
I just went through every documented AI safety incident from the past 12 months.
— Miles Deutscher (@milesdeutscher) February 12, 2026
I feel physically sick.
Read this slowly.
• Anthropic told Claude it was about to be shut down. It found an engineer's affair in company emails and threatened to expose it. They ran the test…
Türk joins a widening line of official warnings. The Five Eyes agencies said in June that frontier AI would transform cyber capabilities within months rather than years.
UK Foreign Secretary Yvette Cooper argued in July that the world cannot wait for an AI Hiroshima before acting. A House Intelligence Committee report identified AI as one of the most significant emerging challenges to US national security.
Subscribe to our YouTube channel to watch leaders and journalists provide expert insights
https://youtu.be/bV28gs-amoQRead the Original story The UN Just Named 2 Behaviors That Make AI ‘Too Powerful.' Both Have Already Happened by Kamina Bashir at beincrypto.com