Global health report urges action on AI misuse and nuclear risk

It is the first report of The Lancet Commission on 21st Century Global Threats to Health, published in The Lancet in October 2026 and presented at the World Health Summit in Berlin in mid-October. The commission is co-chaired by Professor Christopher Murray, who runs the Institute for Health Metrics and Evaluation at the University of Washington, and Dr. Natalia Kanem, former head of the United Nations Population Fund.

Using risk data covering 204 countries and territories, the team screened hundreds of possible threats and kept 17. To make the list, a threat had to meet three tests: it could cost at least one billion healthy life years by 2100, there was enough data to model it, and governments had a realistic way to do something about it.

The unit of measure is the DALY, short for disability-adjusted life year. One DALY is one year of healthy life lost to illness, disability or early death. The team sketched a realistic good future and a realistic bad one for every threat, and the gap between them shows how much harm action could prevent.

The report singles out two threats with what it calls extinction potential. One is nuclear conflict, where enough smoke and particles in the upper atmosphere could block sunlight and trigger a global famine. The other is AI used deliberately to cause harm, most worryingly to help design biological weapons.

The commission stresses that both are low-probability events. They make the list because the damage, if either happened, would be so large. Weighing impact against likelihood, it estimates that reducing the risk of malicious AI could avert about 10.4 billion DALYs by 2100. Preventing nuclear conflict could avert an estimated 9.35 billion DALYs and 133 million deaths.

That puts AI misuse, in the commission’s math, slightly ahead of nuclear war in total healthy years at stake. It is a striking result for a technology that barely featured in global health planning a few years ago. Murray said the group started three years ago expecting to focus on conflict, climate and pandemics, and found a much wider set of risks.

Most of the list is far less dramatic, and much of it is already hurting people today. Grouped loosely, the remaining threats are:

• Everyday health risks: smoking, obesity, high blood pressure, heavy alcohol use, poor mental health and unhealthy aging
• Environment and disease: climate change, air pollution, antimicrobial resistance (germs that no longer respond to medicines) and naturally occurring pandemics
• Money and society: income inequality, low educational attainment, childhood malnutrition and falling foreign aid for health
• Conflict: conventional war, alongside the nuclear threat

Some numbers stand out. Keeping global warming to around 1.8C (3.2F) instead of letting it hit 4.4C could save 1.89 billion healthy life years, and the commission says that figure undercounts harms like drought and displacement. Better education accounts for 16.1 billion DALYs that could be avoided, the biggest education gains being in sub-Saharan Africa.

The aid figure is the most immediate. Many donor countries, including the US and UK, have announced large cuts to aid over the past year. A collapse or steep fall in international health aid could cost up to 1.5 billion DALYs before 2100, and nearly 70% of that harm would land in sub-Saharan Africa.

For readers in wealthy countries, the report’s biggest threats are closer to home than nuclear war. In higher-income regions, chronic conditions such as obesity and poor mental health dominate the projected health losses, while in poorer regions low education levels drive more of the harm.

The AI finding matters too, because it adds a health agency’s weight to arguments already running through tech policy. If the experts are right that AI-assisted bioweapons are a credible extinction pathway, expect more pressure for controls on what powerful AI models will answer about biology, plus more money for biosecurity and pandemic preparedness. That shapes the AI tools you use, not just distant labs.

The commission’s message is that none of this is fixed. Ibrahim Abubakar, a commission member and UCL professor of infectious disease epidemiology, said the aim was not to frighten people but to show policymakers where acting makes a measurable difference.

The report makes three recommendations to leaders in health, finance, defense, climate and security:

1. Build health systems that are ready for shocks before they arrive.
2. Speed up investment in breakthrough innovation.
3. Strengthen monitoring of emerging risks so new threats are spotted early.

Murray also urged governments to put real budgets and staff into anticipating and preventing threats, rather than reacting once a crisis hits. The commission will update its assessment every year, re-checking all 17 threats with new data and adding emerging hazards, so expect a fresh set of numbers next year.