UN scientific panel warns science cannot yet rule out catastrophic AI risk

The UN’s Independent International Scientific Panel on AI published its first Preliminary Report on July 1, 2026, warning that science cannot yet rule out catastrophic harm as AI systems grow more capable. The panel calls it the first global, fully independent scientific assessment of AI’s risks, opportunities and impacts.

The report lands days before governments meet to weigh how, or whether, to govern the technology. It will be presented at the inaugural UN Global Dialogue on AI Governance in Geneva on July 6-7, 2026, with a full annual report due in 2027 ahead of a second dialogue in New York that May.

The 40-member panel, drawn from every UN region, examined seven domains spanning AI science, economic and societal impacts, security, human rights and governance. Co-chair Yoshua Bengio, the deep-learning pioneer, said “science currently cannot guarantee that as capabilities continue to increase, AI will not cause catastrophic harm, either on its own or due to malicious users.” Co-chair Maria Ressa, the Nobel Peace laureate, said risks to societies, security and the human species are already “too high.”

The assessment documents both the reach and the concentration of AI. More than one billion people now use conversational AI weekly, it says, while the United States accounts for 75 percent of the computing power among the world’s top 500 AI supercomputers, against 15 percent for China.

“The world cannot govern what it cannot understand,” UN Secretary-General Antonio Guterres said, describing the report as independent science available to every government.

The panel stops short of forecasting catastrophe; its finding is that current science cannot rule it out, a caveat, not a prediction. Nor does a UN assessment carry binding force. Whether its warnings shape the Geneva talks, or are noted and set aside, is the open question the report itself cannot answer.