AI leaders warn of serious risks

Top AI CEOs are now publicly calling for slower development, citing threats from loss of control to cyberattacks to bioterrorism. The Bank of England warns that AI poses a new threat to an already fragile global financial system.

“one exploited weakness can spread rapidly”

170 countries found common ground

At the first UN Global Dialogue on AI Governance in Geneva this July, nations alongside industry and civil society agreed the world needs common rules for AI development and deployment.

Connect scientists and policymakers

Right now, AI researchers studying risks and diplomats negotiating rules work on separate tracks with no structured communication. The UN's scientific panel needs a formal channel into the negotiating process before next May.

Build a shared safety baseline

The UN should develop a common reference baseline for AI safety grounded in human rights law, drawing from existing frameworks like the OECD AI Principles and G7 Hiroshima code of conduct.

Fund participation for all nations

Governments and civil society with less financial capacity need dedicated funding to participate in global AI governance conversations, similar to how the climate panel is funded.

Test cross-border AI regulation

A small group of countries should pilot cross-border regulatory sandboxes to prove different systems can work together, similar to how financial regulators test new products across borders.

Three deliverables by May 2026

By the time the Global Dialogue reconvenes in May, three things must be in place: an evidence assessment from the scientific panel, a Dialogue agenda on safety baseline, and a first cross-border sandbox pilot.

Act now or lose control

If a handful of powerful actors decide how AI is governed, the rest of the world loses out. The window to establish inclusive global rules is closing fast.

“We do not have time to relitigate issues that should already be settled.”