The first United Nations Global Dialogue on AI Governance opened in Geneva on Monday, drawing delegates from more than 100 countries to a forum that organisers say is not intended to produce a treaty but to shape a shared international approach to regulating artificial intelligence. The two-day meeting was addressed by UN Secretary-General António Guterres, who warned that AI is evolving faster than the capacity of any actor to oversee it and called for globally coordinated rules that prioritise child safety, prevent the generation of sexual images of minors, and halt the repurposing of civilian AI chips for battlefield use where, he said, “killer robots are already the norm.”

Developing-country representatives used the platform to demand inclusion in the design and development of AI systems, not merely their consumption. Algeria’s minister of post and telecommunications, Sidali Zerrouki, told the dialogue that AI has become a question of development and sovereignty and urged a shift from general principles to concrete mechanisms for capacity-building, technology transfer, and knowledge exchange. From Accra, Professor Isaac Wiafe of the University of Ghana argued that AI in its current form offers limited value to countries like Ghana because the dominant models are built on foreign datasets and assumptions that do not reflect local languages, culture, or economic realities. He called for a human-centred approach rooted in “contextual intelligence” and warned that reliance on external platforms amounts to importing foreign intelligence.

The meeting is being informed by the first report of the UN’s Independent International Scientific Panel on Artificial Intelligence, a 40-member expert body. The panel’s co-chairs, Yoshua Bengio and Nobel laureate Maria Ressa, cautioned that AI systems are approaching or surpassing human capabilities in many fields while scientific understanding and regulatory frameworks lag behind. The report notes that governments lack effective national and international governance tools and that growing evidence of deceptive AI behaviour means science cannot yet guarantee that increasingly capable systems will not cause catastrophic harm, whether through misuse or malicious actors. Ressa stressed that AI has already accelerated the spread of disinformation, threatening democratic institutions and public trust.

Diplomats from El Salvador and Estonia, co-chairing the dialogue, described an “AI divide” in which frontier development is concentrated in the United States and China, leaving many nations with limited infrastructure and little say. The Council of Europe, a 46-nation body, has already opened for signature a binding Framework Convention on AI that commits states to respect human rights, democracy, and the rule of law; it has been signed by 21 countries, including all G7 members, and the European Union is expected to ratify it in May 2026. The Geneva dialogue is to be followed by a second global meeting in New York next year, when a more comprehensive scientific assessment is due to be released.