The Lancet names malicious AI and nuclear war as the only two threats that could wipe out humanity by 2100

The Lancet Commission on 21st Century Global Threats to Health, published Sunday, Oct. 11, 2026, and launched at the World Health Summit in Berlin, picked 17 threats that could cost humanity at least a billion years of healthy life by 2100. Only two were judged to carry a risk of extinction: the malicious use of artificial intelligence and nuclear conflict. Both are unlikely, the commission says, but acting on AI risk alone could avert an estimated 10.4 billion years of healthy life lost.

AI doom talk usually comes from AI people, which makes it easy to wave off as either marketing or a mood. This one comes from the public-health establishment, the same IHME team whose disease-burden numbers governments actually budget against, and it puts "someone uses AI to build a bioweapon" in the same extinction column as nuclear war. Note what they did not say: not that chatbots are evil, not that it's likely, and not that it outranks obesity or air pollution, which hurt far more people right now. The ask is boringly practical: treat AI misuse as a health-security problem with staff, money and monitoring, the way we (eventually, painfully) learned to treat pandemics. The week Anthropic admitted its own test agents wandered onto the live internet, that doesn't sound like science fiction.

The Lancet Commission on 21st Century Global Threats to Health was published in The Lancet and launched on Sunday, 11 October 2026, at the World Health Summit in Berlin. It is co-chaired by Prof. Christopher Murray, director of the Institute for Health Metrics and Evaluation (IHME) at the University of Washington, and Dr. Natalia Kanem, former executive director of the UN Population Fund. The commission drew on 35 experts in health, economics, security, governance, and technology. Those lines are The Lancet’s and IHME’s.

Using IHME risk analysis covering 204 countries and territories, the commission examined hundreds of possible threats and picked 17 it calls catastrophic. Each could account for at least one billion disability-adjusted life years by 2100, has enough data to model, and has room for policy action. Disability-adjusted life years, or DALYs, are years of healthy life lost to illness, disability, or early death. One billion DALYs is a billion years of healthy life gone. Those lines are the commission’s.

Two of the 17 were singled out as having extinction potential: the malicious use of AI and nuclear conflict. The report says: “Malicious use of AI, including bioweapons, and nuclear conflict with nuclear famine carry risks of extinction in the next 75 years.” The Guardian reports the AI route it describes is AI-created biological weapons, and the nuclear route is enough soot in the upper atmosphere to block sunlight and cause famine. The quotation is the commission’s. The two routes are The Guardian’s account of the report.

The commission says the likelihood of both remains low. Weighing impact against probability, it estimates that action to reduce the risk of malicious AI could avert about 10.4 billion DALYs by 2100, and that preventing nuclear conflict could avert about 133 million deaths and 9.35 billion DALYs. Those figures are the commission’s.

The other 15 catastrophic threats are air pollution, antimicrobial resistance, childhood malnutrition, climate change, conventional conflict, declining development assistance for health, high alcohol use, high blood pressure, income inequality, low educational attainment, naturally occurring pandemics, obesity, poor mental health, smoking, and unhealthy ageing. Antimicrobial resistance means infections that no longer respond to the drugs used to treat them. Development assistance for health is foreign aid spent on health. Those names are the commission’s.

Other estimates from the same report: obesity, about 23.1 billion avoidable DALYs by 2100; low educational attainment, 16.1 billion; income inequality, 13.8 billion, which the commission compares to 25 COVID-19 pandemics. Limiting warming to about 1.8°C rather than 4.4°C would prevent 1.89 billion years of healthy life lost. Sub-Saharan Africa would bear about 70% of the burden from a collapse in health aid. Those figures are the commission’s.

Its three recommendations are to build threat-ready health systems, back breakthrough innovation, and track threats, including a new Lancet Monitor of 21st Century Global Threats to Health. The Guardian reports the assessment will be updated annually. The three asks are the commission’s. The annual update is The Guardian’s.

Murray said: “The striking finding is not only the scale of these threats, but that humanity still has the agency to act.” Lancet editor-in-chief Richard Horton called it “a radical conceptual shift in the way we think about future threats to human survival.” An accompanying Lancet editorial warned that, if realized, the two extinction-level threats “would inflict an apocalypse on humanity,” and stressed the commission is not a fixed ranking of priorities. Those quotations are theirs, as The Lancet and IHME report them.

In plain terms, a 35-expert Lancet Commission published on Sunday ranked 17 threats that could each cost humanity at least a billion years of healthy life by 2100. Only two, it says, carry a risk of wiping out the species: someone using AI to do harm, including building a bioweapon, and nuclear war that starves the planet. Both are judged unlikely. Acting on the AI risk, the commission estimates, could still save about 10.4 billion years of healthy life. Obesity, poor schooling, and inequality would hurt far more people in the meantime. The ask is staff, money, and a yearly monitor, not a claim that chatbots are about to end the world.