Delegations from 193 nations convened in Geneva on 6 and 7 July for the first session of the U.N.’s Global Dialogue on Artificial Intelligence Governance, a forum mandated by the 2024 Global Digital Compact and formally established under General Assembly resolution A/RES/79/325. The Dialogue isn’t designed to produce a treaty. It’s designed to produce a shared vocabulary, which is a more modest ambition and, on the evidence of the opening remarks, still an ambitious one.

Secretary-General António Guterres set the tempo. Artificial intelligence, he said, is “being deployed faster than anyone, including the people building it, can keep up,” a technology already positioned to “reshape economies, transform the world of work, sway elections and tilt the balance of security.” Per Reuters, he reached for the comparison that every diplomat in the room already knew by heart: “The internet took 15 years to reach a billion people. A.I. got there in two.” His prescription was terse. “Innovation needs guardrails. If A.I. is to be powerful, it must be governed.”

The empirical backdrop arrived on 1 July, when the Independent International Scientific Panel on Artificial Intelligence, drawing members from all five U.N. regions, published its preliminary report. Its verdict on the current regime was blunt: existing safeguards “cannot keep pace with the growth of A.I.’s capabilities,” with the technology “outpacing both scientific understanding and governments’ ability to adapt.” A June assessment from the same panel had warned the systems could “cause catastrophic harm, either on its own or due to malicious users.”

The report also mapped the terrain of power. More than a billion people now use conversational A.I. weekly, while adoption in developing countries lags. The United States controls roughly 75 percent of the compute among the world’s top 500 A.I. supercomputers; China holds 15 percent. Everyone else is a rounding error, negotiating governance for an industry they don’t host.

Panel co-chair Yoshua Bengio told U.N. News the technology “is approaching or surpassing human capabilities in many domains,” and that the trajectory “doesn’t look like it’s stopping.” Assembly President Annalena Baerbock cited a starker cultural indicator: 99 percent of deepfakes are sexual in nature, and 96 percent target women and girls.

The session’s co-chairs, El Salvador’s Egriselda López and Estonia’s Rein Tammsaar, gaveled the meeting closed with a second session scheduled for New York in May 2027. That’s a ten-month gap in a field that just compressed the internet’s first decade into twenty-four months. The delegates know the arithmetic.

Sources