Delegates from every U.N. member state gathered Monday at the Palexpo convention centre in Geneva for the opening of the Global Dialogue on AI Governance, the first United Nations forum in which every government holds a seat at the table on artificial intelligence rules. The two-day meeting, established by General Assembly Resolution A/RES/79/325 and mandated under the Global Digital Compact and the Pact for the Future, is co-chaired by Ambassador Egriselda López of El Salvador and Ambassador Rein Tammsaar of Estonia, and follows six months of consultations that drew more than 1,500 written submissions from governments, civil society and the private sector.

The proceedings are shadowed by a preliminary report from the Independent International Scientific Panel on Artificial Intelligence, issued the week before the Geneva opening. Its findings read less like a policy brief than a warning shot. Adoption has, in the panel’s phrase, “compressed from decades into months.” More than a billion people now use conversational A.I. tools each week. The United States accounts for roughly 75 percent of global A.I. supercomputing capacity; China holds about 15 percent. And on the frontier, the panel concludes, “no known technical guarantee currently exists” that the most advanced agent systems will follow the instructions they’re given, with rising evidence of deceptive model behavior and potential misuse for fraud, cyberattacks and biological threats.

“AI is approaching or surpassing human capabilities in many domains. It is outpacing both scientific understanding and governments’ ability to adapt,” Yoshua Bengio, the panel’s co-chair, told UN News ahead of the meeting.

The co-chairs are framing the dialogue around distribution rather than restriction. Tammsaar said A.I. “could be a great equalizer” for many countries, supporting economic development and health systems. López said the technology could help governments “better improve their work and the delivery of services.” It’s the language of a forum that knows it can’t outpace the labs, and has instead chosen to argue over who benefits.

The choreography around the venue matters. The Geneva session runs alongside the World Summit on the Information Society Forum and the ITU AI for Good Global Summit, and on July 8 the ITU convenes a separate AI for Good Global Commission co-chaired by Salesforce chief executive Marc Benioff and President Paul Kagame of Rwanda, with Amazon’s Andy Jassy, Microsoft president Brad Smith and Nvidia chief executive Jensen Huang among the participants. One track seats every government; the adjacent track seats the executives whose companies own most of the compute the panel just finished counting.

A second session of the Global Dialogue is scheduled for New York in May 2027. That gives the multilateral system roughly ten months to translate 1,500 submissions into something the frontier labs can’t simply outrun between meetings.

Sources