UN says AI safeguards can’t wait for certainty

A United Nations scientific panel warns that governments must implement stronger safeguards for AI agents without waiting for full scientific certainty about their risks. This advice comes as the UN General Assembly convenes, with AI safety on the agenda, including discussions between the US and China. The panel, in its first thematic brief, invokes the precautionary principle, stating that potential catastrophic or irreversible harm justifies action even amid uncertainty. It references recent incidents, such as hacks at OpenAI, Anthropic, Google, and Meta, to underscore the urgency. The report emphasizes the need for greater international coordination and resources to manage AI risks.

(Source：The Verge)