Artificial intelligence and nuclear conflict pose potential extinction risks, alongside 15 other major threats to global health and wellbeing, according to a new report published in The Lancet.

The report, produced by a commission of 35 leaders from various industries, identifies 17 interconnected threats that could collectively cause tens of billions of years of healthy life to be lost worldwide by 2100 if current trends continue. The risks range from obesity, high blood pressure and mental health problems to environmental change, economic instability, technological risks and conflict.

The commission estimates that the malicious use of AI could lead to around 10.4 billion years of healthy life lost globally by 2100, a figure comparable to the projected impact of obesity and naturally occurring pandemics. However, researchers stressed that AI-related estimates remain theoretical, as no malicious AI event of truly global magnitude has yet occurred. Forecasts of AI catastrophe also varied widely among experts consulted by the commission.

The report warns that obesity rates are rising worldwide and could affect 60% of adults and one-third of children by 2050. Researchers say addressing these threats will require targeted action by governments, stronger international cooperation and greater public trust. They argue that early intervention could help prevent worst-case scenarios and protect future generations, particularly vulnerable communities.