Christopher Murray, an epidemiologist at the University of Washington who runs the Global Burden of Disease study at the Institute for Health Metrics and Evaluation, says healthy life expectancy will keep rising through 2100 — but that progress is now shadowed by two threats his Lancet Commission work flags as capable of ending humanity outright: malicious use of artificial intelligence and nuclear war, L’Express reported.
Murray co-chairs the Lancet Commission on global threats to health, which examines what could reverse the gains made since 1945 — including weaponized AI, nuclear conflict, climate change, obesity and inequality. His diagnosis is two-sided: on current trends, healthy life expectancy continues to climb, yet the predictability that defined the postwar era has evaporated, and AI is largely why.
Healthy life expectancy keeps rising — with a catch
Murray’s preferred metric is healthy life expectancy, which counts not just years lived but years in good functional health. That figure continues to increase, though more slowly than overall life expectancy, because people spend more years with illness at older ages. Some countries, notably the United States, have stagnated under the combined weight of opioids, obesity and a slight rise in suicides. Most countries returned to their pre-pandemic trajectory after COVID-19’s mortality shock.
By the end of the century, Murray’s reference scenario — current trends continuing — projects life expectancy of 85 to 90 years for most upper-middle and high-income countries. Low-income countries, particularly in sub-Saharan Africa, will likely retain a gap of 10 to 15 years. The gains will come mainly from better survival after age 70, not from growing numbers of centenarians: nothing in the data suggests more people will pass 105, and Murray is skeptical of visions of living to 200.
Where medicine will — and won’t — deliver
The biggest advances will likely come against cardiovascular disease and cancer, which concentrate enormous research effort. In vivo gene therapy could affect some diseases, and HIV research inspires hope, though feasibility remains uncertain. Monogenic diseases would benefit enormously at the individual level, but affect too few people to shift population-level outcomes.
Progress in preventing the conditions that degrade quality of life — dementias, neurological disorders, bone and joint disease, mental health problems — is moving much more slowly. For now, humanity is better at reducing mortality than at guaranteeing additional healthy years, and only major therapeutic innovations could change that.
AI: a 10% risk of catastrophe, and a much larger upside
The Lancet Commission report identifies malicious use of AI — including the design of biological weapons — as a risk capable of threatening human survival. Malicious actors already have access to tools that can facilitate the design or production of dangerous biological agents, simple DNA recombination, and even novel pathogens. The hardest scenario to grasp is loss of control: attacks on health systems or the power grid, lasting destruction of the internet, autonomous weapons.
Murray puts the risk of very bad AI outcomes at 10% or less, tilting firmly toward the positive: diagnostics are already improving, research will accelerate, and therapies should follow. Better regulation could push the balance further. He also flags the social disruption from employment displacement and its mental-health effects as a legitimate concern that models could not capture because the scale is unknown.
Commission warns of famine from nuclear dust
The Commission classifies nuclear conflict, and its possible nuclear winter, among extinction-potential threats. Any nuclear exchange would be gravely serious with lasting consequences where it occurs, likely among the great powers. The controversial nuclear famine scenario holds that a sufficiently massive exchange would loft enough dust into the stratosphere to cool the planet and slash global food production. Some argue part of humanity would survive through alternative food production, but Murray said it cannot be ruled out — which is why the Commission retained it as a risk to account for.
Obesity, education and inequality top the modifiable threats
Measuring the “modifiable” burden of disease — the gap between the best and worst plausible outcomes observed across countries over 35 years — the Commission found the largest addressable factor is obesity, followed by low education, then income inequality. Climate change is already killing through heat, spreading malaria and dengue, and threatening food security: under a high-emissions scenario, the Sahel, West Africa and South Sudan could lose up to three years of life expectancy by century’s end, and Arizona about one year. Cold countries like Norway and Scotland might see some winter-mortality relief, but that would not offset the losses elsewhere. Climate is serious but not an extinction threat this century.
One decision Murray would impose on world leaders
Murray remains broadly optimistic: humanity contained the nuclear threat for more than 75 years and has mastered other dangers. But that, he said, requires taking these threats seriously and acting before it is too late.
Worth a look
