A new Lancet commission identifies malicious AI use as a threat with the potential to cause human extinction. The World Health Summit presented its report in Berlin on October 11. The researchers examine health risks through 2100; that year is the study horizon, not a predicted deadline for humanity.
The Institute for Health Metrics and Evaluation's detailed original summary includes the possible use of AI to develop biological weapons. The commission places malicious AI use alongside nuclear conflict in a separate threat category. Its assessment covers 17 particularly far-reaching threats overall.
For AI, however, historical data are insufficient for robust quantitative models. The official announcement says the assessment instead used structured expert judgement and superforecaster evaluations. The resulting estimates carry greater uncertainty and are intended as indicative rather than settled forecasts.
The commission, co-chaired by Christopher Murray and Natalia Kanem, calls for earlier preparedness, investment in protective tools and regular independent reassessment of threats. Its publication describes possible future harm and opportunities to act, rather than an AI attack that has already occurred.