UN AI panel urges governments to act on risks without full certainty
The brief pushes AI safety up the diplomatic agenda, calling for stronger safeguards and coordination despite differing national laws.
A new UN scientific panel on AI has issued its first thematic brief, urging governments to implement robust safeguards for advanced AI systems now rather than waiting for full scientific certainty about their risks. The report uses OpenAI’s hack of Hugging Face earlier this year, along with subsequent incidents at OpenAI, Anthropic, Google, and Meta, as evidence that increasingly capable AI agents can already be misused at scale. It argues that loss‑of‑control scenarios fit the “precautionary principle,” originally codified in the 1992 Rio Declaration, meaning regulators should act even when the probability of catastrophic harm is not well quantified. The panel calls for more resources devoted to managing emerging AI risks and for stronger international coordination on safety and accountability, while accepting that national legal frameworks will differ. The brief lands ahead of the UN General Assembly in New York and parallel US‑China talks on AI, signaling that AI governance is becoming a central diplomatic issue for states and global companies building these systems.
Why it matters
The panel’s intervention raises pressure on governments and major AI developers by arguing they should not wait for complete scientific clarity before tightening safeguards for advanced systems. By grounding its case in recent security incidents at leading AI companies and invoking the precautionary principle, it strengthens the policy footing for earlier and more coordinated regulation. Landing alongside UN General Assembly meetings and planned US–China talks, it signals AI safety is becoming a core subject of international diplomacy and corporate scrutiny.