United Nations Secretary-General António Guterres has warned that artificial intelligence is being deployed faster than societies and governments can control it, and that humanity may be the last generation able to decide the terms of coexistence with machines. Speaking at the first Global Dialogue on AI Governance in Geneva, Guterres called for urgent international action to prevent a digital divide from turning into an AI exclusion gap that threatens development, security and sovereignty.
During the opening session on 6 July 2026, Guterres described the current situation as an experiment being conducted on societies without their consent. He stressed that both the risks and the potential of AI are real, and that the technology is advancing far beyond the ability of anyone, even its creators, to keep pace. Guterres urged world leaders to either harness AI or let it drift, warning that the window for decisive action will not stay open for long.
Concentration of power and the global AI divide
According to a UN expert report discussed at the dialogue, 75 percent of the computing power from the world’s top 500 AI supercomputers is located in the United States, and 15 percent in China. The rest of the world shares the remaining 10 percent. In 2025, US-based organisations produced 59 significant AI models, China 35, and the rest of the world only a handful. This concentration of technology means most nations have no influence over AI systems used by their citizens and cannot audit them.
– Most countries have no say over this technology. They do not even have the ability to verify how the AI models used by their citizens work – said Jakub Szymik, founder of the Digital Democracy Observatory Foundation, who is attending the Geneva talks.
The report also warns that the most advanced models, such as Mythos, with documented offensive cyber capabilities, are being treated as national security assets, effectively turning AI into a tool of international domination. States that lack their own AI infrastructure risk losing control over standards, safety and sovereignty.
The threat to democracy and information integrity
Another major concern raised at the dialogue is AI’s potential for large-scale manipulation of public opinion. The UN report highlights that AI enables mass production of persuasive content, including disinformation, and can personalise propaganda in real time. Experimental studies have shown that an AI model optimised for persuasion shifted opposition voters’ preferences by as much as 25 percentage points.
– The threat of large-scale manipulation was very strongly emphasised. It was said that AI can be used for precise voter manipulation and can even tip the scales of elections – added Szymik.
The report notes that no democracy has yet built effective defence mechanisms that can operate at the speed of AI-driven information operations. It also points to the use of AI to generate child sexual abuse material (CSAM) and the mental health risks to minors who interact with AI models that tend to mirror and reinforce users’ beliefs, leading to documented deaths.
A race against time: the global fund and the need for action
As a remedy, Guterres proposed the creation of a Global Fund for AI to help member states build skills, data resources and affordable computing power. Over 20 countries have already responded to his invitation to join a UN-supported Global Network for AI Capacity Building. The report emphasises that AI is already being used by governments and security services for surveillance and in armed conflicts, drastically expanding social control.
– We cannot allow a digital divide to become an AI divide, and an AI divide to become a development, security and sovereignty gap – Guterres concluded. – We may be the last generation that can determine the terms of coexistence between humanity and machines. The door is still open, but it will not remain open for long.
Źródło: WNP.PL, Fot. UN Photo/Evan Schneider, Frame Stock Footage/Shutterstock
