All 193 UN member states gathered at the Palexpo arena on Monday for the opening of the first Global Dialogue on Artificial Intelligence Governance, a two-day convening co-coordinated by the ITU and UNESCO that drew more than 2,600 registered participants and over 4,500 written submissions to the co-chairs. It’s the broadest state-level forum on AI ever assembled, and pointedly, not a treaty-making process.

Secretary-General António Guterres set the frame in his opening remarks, warning that the technology is “being deployed faster than anyone, including the people building it, can keep up.” He laid out four priorities led by safety, pressed for “common baselines for frontier systems” and risk-evaluation methods, and argued that systems with global reach must be “worthy of global trust.” His line for the headlines: “If AI is to be powerful, it must be governed.”

Guterres then walked through the harms he wants on the table, from protecting children online to civilian-designed AI chips migrating to the battlefield, where autonomous “killer robots” are, in his telling, already the norm. General Assembly President Annalena Baerbock added the data point that lands hardest: 99 per cent of deepfakes are sexual in nature, and 96 per cent target women and girls.

The asymmetry underneath the room was documented in an independent UN scientific report released alongside the Dialogue. The United States holds 75 per cent of the computing power across the world’s top 500 AI supercomputers. China holds 15 per cent. Every other state at Palexpo, including most of the 193, shares what’s left.

A second Dialogue is scheduled for New York in May 2027. The governance venue is now multilateral; the compute isn’t.

Sources