UN Expert Panel Urges Governments to Regulate Advanced AI Before Risks Are Fully Understood

A UN independent international scientific expert panel on AI has released its first special briefing, urging governments to regulate increasingly powerful AI agents before risks are fully understood. The briefing is the UN's first major assessment document on the OpenAI intrusion into Hugging Face earlier this year.

The panel said the world does not need to wait until researchers fully understand how such incidents happen before putting stronger safeguards in place. The risk of losing control is exactly what the precautionary principle addresses: potential harm may be catastrophic and irreversible, but the probability cannot yet be determined scientifically.

This principle was first written into the 1992 UN Rio Declaration on Environment and Development, stating that scientific uncertainty should not delay control measures. Since the Hugging Face intrusion was first reported, multiple similar incidents have been recorded involving OpenAI, Anthropic, Google, and Meta.