Nuclear war and malicious use of AI threaten end of humanity by 2100, Lancet say

A global Commission of experts mapped 17 interlinked catastrophic threats to health and human survival between 2025 and 2100, spanning disease, social and economic conditions, environmental change, technology and conflict. Using historical evidence and new risk modeling across 204 countries, threats were classified as catastrophic if they could account for at least one billion amenable disability-adjusted life years (DALYs) by 2100, had sufficient data to model, and offered scope for policy action. Two threats were singled out as having extinction potential: malicious use of artificial intelligence and nuclear conflict. The Commission modeled plausible best- and worst-case futures to quantify avoidable health loss, stressed that harms will be unequally borne by marginalized and low-resource populations, and argued that leaders’ decisions now will determine how many years of healthy life future generations gain or lose.

Key quantified findings include that preventing nuclear conflict could avert an estimated 133 million deaths and about 9.35 billion DALYs by 2100, while reducing malicious AI risk could avert roughly 10.4 billion DALYs. Limiting warming to ~1.8°C versus 4.4°C could save about 1.89 billion healthy life years; education improvements could address 16.1 billion amenable DALYs; income inequality and obesity correspond to ~13.8 billion and 23.1 billion DALYs respectively; a collapse in development assistance would disproportionately hit sub‑Saharan Africa (≈70% of that burden). The Commission issues a three-part roadmap: build threat-ready health systems with durable financing, back breakthrough innovations (vaccines, antibiotics, obesity and climate technologies) with equitable governance, and track emerging risks via an institutionalized monitor to inform cross-sector decision-making.