The Pulse
Lancet Commission Names Malicious AI an Extinction Risk
A Lancet Commission identifies malicious use of artificial intelligence and nuclear conflict as threats with potential to cause human extinction by 2100. Its assessment says action to reduce malicious-AI risk could avert an estimated 10.4 b
AI.info Team · Written with AI ·
A Lancet Commission published on October 10 places malicious use of artificial intelligence in the same extinction-potential category as nuclear conflict, while saying the likelihood of both remains low. The pairing sets a stark upper bound on the possible consequences, not a forecast that either catastrophe is likely. The Commission also cautions against reading its threat categories as a fixed ranking of policy priorities.
The report, launched at the World Health Summit in Berlin on October 11, examines 17 threats to health and human survival between 2025 and 2100. Its authors say governments, funders and industry have practical ways to reduce the risks, but the assessment frames malicious AI as a danger whose possible scale warrants attention even amid deep uncertainty.
Extinction potential, not a probability ranking
The Commission examined hundreds of possible threats and identified 17 that met its criteria for catastrophic risk: the threat could account for at least one billion disability-adjusted life-years by 2100, enough data existed to model it, and policy action could make a difference. The list spans technology and conflict as well as familiar health and social risks, including naturally occurring pandemics, climate change, air pollution, childhood malnutrition, obesity and income inequality.
For each threat, the analysis compares a plausible best-case future with a plausible worst-case future. The gap, measured in disability-adjusted life-years, represents health loss that targeted action could avoid. A DALY is a year of healthy life lost through illness, disability or premature death; the metric gives decision-makers a way to compare very different threats, but it does not mean all are equally likely or equally urgent.
“The striking finding is not only the scale of these threats, but that humanity still has the agency to act. Across all 17 catastrophic threats, the decisions made by leaders today will shape the health and wellbeing of the next generation. But delaying action will only serve to shift and increase the costs into the future.”
Christopher JL Murray, Commission co-chair and director of the Institute for Health Metrics and Evaluation, University of Washington
What the 10.4 billion figure measures
The Commission estimates that action to reduce malicious-AI risk could avert 10.4 billion DALYs by 2100, accounting for both potential impact and probability. That figure is an estimate of avoidable health loss across the assessment period, not a prediction that AI will cause a single event costing that number of healthy-life years. The release does not give a separate probability estimate for an AI catastrophe.
That distinction matters: the report labels malicious AI a threat with extinction potential while describing its likelihood as low. Its accompanying editorial says the Commission does not offer a fixed ranking of threats or policy priorities, and that findings may change as evidence accumulates. The assessment is presented as a framework for action and reassessment, rather than a league table of dangers.
Readiness, investment and monitoring
The Commission proposes three broad responses: build health systems prepared for a range of threats, increase investment in areas where current tools fall short, and monitor emerging risks. Its examples of innovation include vaccines, antibiotics, obesity treatments and climate technologies, paired with appropriate governance and equitable access.
For ongoing oversight, the authors recommend a Lancet Monitor of 21st Century Global Threats to Health, an institutional mechanism to reassess risks and bring evidence to decision-makers in health, finance and security. The Commission was informed by 35 leaders across health, economics, security, governance and technology, and drew on risk analysis covering 204 countries and territories. Its central claim about AI is narrower than a prediction of imminent catastrophe: malicious use belongs among the threats that could end human survival, even as the Commission describes that outcome as unlikely.