The United Nations convened its first Global Dialogue on AI Governance in Geneva on July 6, bringing together delegations from all 193 member states in what organizers called a historic milestone: the moment the international community officially recognized that artificial intelligence has become a geopolitical challenge, not just a technological one.

Expert panel: no guarantee of control

The Independent International Scientific Panel on Artificial Intelligence — 40 global experts appointed by the UN General Assembly — presented a preliminary report at the conference’s opening. Its central finding was unsparing: AI capabilities are advancing faster than any government’s ability to understand or regulate them, and “science currently cannot guarantee that as capabilities continue to increase, AI will not cause catastrophic harm.”

The panel, co-chaired by Turing Award-winning computer scientist Yoshua Bengio and journalist Maria Ressa, flagged risks including deceptive behavior in AI systems, misuse for propaganda, and erosion of democratic institutions.

Four proposals from the Secretary-General

UN Secretary-General António Guterres opened the dialogue by identifying three defining features of modern AI: rapid reach (the technology reached a billion users in two years, compared to fifteen for the internet), extreme power concentration in “a handful of companies and a handful of countries,” and a deteriorating information environment where “a machine-enabled lie can now persuade as effectively as the truth.”

Guterres called for four concrete steps:

  • Common safety baselines for frontier AI systems, including an AI Child Safety Pledge requiring pre-release safety testing and zero tolerance for child sexual abuse material.
  • Human red lines in high-stakes decisions, with machines informing but humans deciding.
  • A Global Fund for AI to help developing nations build independent capacity rather than remain dependent on technology they cannot audit or adapt.
  • An Environmental Transparency Initiative requiring disclosure of data centers’ carbon, water, and land footprints.

The Secretary-General also called for an international treaty banning lethal autonomous weapons, stating the decision to take a human life “must remain forever human.”

What comes next

The Geneva event is the first of two planned sessions; a second dialogue is scheduled for New York in May 2027. The conference ran alongside the ITU’s AI for Good Global Summit, which also opened on July 7 at the same Palexpo venue in Geneva.