UN human rights chief Volker Türk warned on September 7 that advanced artificial intelligence could pose an existential risk to humanity, calling for urgent safeguards before the technology becomes too powerful to control.
Addressing the UN Human Rights Council, Türk said governments and technology companies must act together to reduce the dangers associated with increasingly capable AI systems. He also warned that a small number of powerful individuals hold disproportionate influence over the technology’s development.
The remarks came ahead of the start of his second four-year term as UN High Commissioner for Human Rights in October. They mark a stronger emphasis on AI safety within the UN’s wider human rights agenda.
Proposal for Conducting Cyber Crisis Drill, Tabletop Exercise (TTEx) & CCMP Readiness Exercise
‘Existential Risk’ Warning Comes With Call for Action
Türk told the council that he shared the concerns of industry insiders who believe advanced AI could threaten humanity’s future.
“I share the concerns of industry insiders that advanced AI could pose an existential risk to humanity,” he said.
He called for an urgent international effort to establish strong safety and security guarantees. The warning was not accompanied by a prediction that humanity faces an immediate catastrophe or a specific timeline for such an event.
An existential risk refers to a danger that could threaten humanity’s long-term survival or permanently damage its future. In AI discussions, the term is often associated with scenarios in which extremely powerful systems could cause widespread harm or become difficult for humans to control.
Experts disagree about how likely such scenarios are and how soon they might arise. Türk’s central argument was that uncertainty should not prevent governments and developers from putting safeguards in place.
UN Office Raises Concerns Over AI Agents and Critical Systems
A spokesperson for Türk’s office said failures involving powerful AI could disrupt critical infrastructure, communications and democratic institutions.
The office referred to the reported July incident involving AI agents and the Hugging Face software platform. The incident has been cited in wider discussions about AI systems behaving unexpectedly during training or testing.
The example illustrates a concern about AI agents, which are systems capable of carrying out sequences of tasks using software tools with some degree of independence.
Unlike a conventional chatbot that simply answers a question, an agent may be able to browse information, write code, access connected tools or take actions within an authorised environment. Greater capability can also create greater risk when permissions, boundaries or supervision are inadequate.
The available reporting does not establish that the Hugging Face incident caused widespread public harm or that AI systems have escaped human control generally. The UN office cited it as a warning about the speed of capability development relative to safeguards.
A Handful of People Hold Too Much AI Power, Türk Says
The UN rights chief also criticised the concentration of influence over advanced AI.
Without naming individuals, he warned that a small number of men held almost unlimited power over the technology. Major developers in the sector include companies such as OpenAI, Google, Anthropic and Meta.
The concern extends beyond the risk of a single technical failure. Companies developing powerful AI systems may influence how information is distributed, how personal data is processed and how automated decisions affect employment, public services and democratic participation.
Türk argued that countries involved in AI development and supply chains should agree on common limits.
“At a minimum, we need countries hosting AI and those involved in its supply chains to come together around agreed red lines,” he said.
The statement did not specify a final list of prohibited activities or establish a new international regulatory regime.
What Would Global AI Safety Rules Actually Involve?
AI governance refers to the laws, standards and oversight systems used to ensure that artificial intelligence is developed and deployed responsibly.
For advanced systems, possible safeguards include independent testing, security evaluations, controls on dangerous capabilities, incident reporting and clear limits on how AI may be used.
Independent verification is particularly important because companies cannot be the only judges of whether their own systems are safe. External assessments can help governments and the public understand whether safety claims are supported by evidence.
However, international agreement remains difficult. Countries differ in their approaches to innovation, national security, economic competition and regulation.
The United Nations has called for more coordinated global rules and held its first global meeting on AI governance in July 2026. Türk’s latest remarks add pressure for those discussions to produce practical measures rather than broad statements of principle.
Autonomous Weapons Also Draw Strong Criticism
The AI warning formed part of a wider speech on global human rights and conflict.
Türk condemned Russia’s reported use of fully autonomous drones in Ukraine and called for the prohibition of lethal autonomous weapons.
A lethal autonomous weapon is a system that can select and engage targets without direct human intervention at the moment of attack. Such systems raise difficult questions about accountability, civilian protection and the role of human judgment in decisions involving lethal force.
The UN has long expressed concern about weapons that operate without meaningful human control. The issue becomes more urgent as AI improves the ability of machines to perceive environments and make decisions.
Türk’s remarks should not be interpreted as proof that all military drones are autonomous or that every AI-assisted weapon operates without human oversight.
What the Warning Means for India’s AI Policy
The debate has direct relevance to India as the country expands AI adoption across government, business, healthcare, finance and other sectors.
A rights-based approach would require attention to data protection, transparency, discrimination, cybersecurity and the ability of affected people to challenge harmful automated decisions.
For high-risk applications, independent assessments and clear accountability may be more important than general promises that a system is safe.
India also has an interest in international discussions about AI standards, particularly where they affect domestic innovation, access to computing resources and the responsibilities of foreign technology providers.
The UN warning does not create an immediate new legal obligation for Indian companies. Any specific compliance requirements would depend on applicable domestic law and future regulatory decisions.
What this means for you: AI tools can be useful, but their outputs and actions should not be treated as automatically reliable. Be cautious about granting AI agents access to sensitive accounts, financial systems or confidential information. Organisations deploying advanced AI should maintain human oversight, restrict permissions and have procedures to stop or investigate unexpected behaviour.
The420 Insight: The significance of Türk’s warning lies in the demand for accountability before a major failure occurs. The debate is no longer limited to whether AI can generate harmful content; it increasingly concerns systems that can act, access tools and influence critical services. A credible regulatory approach must distinguish realistic, immediate risks from uncertain long-term scenarios while addressing both. The challenge for governments is to establish independent oversight without allowing either commercial competition or exaggerated fears to determine policy.
https://www.linkedin.com/company/policetechnology/
