Lancet assessment warns AI attacks and nuclear conflict could end humanity by 2100
A new global risk assessment from The Lancet Commission identifies 17 catastrophic threats to human health that could emerge over the next 75 years. According to the report, malicious use of AI and nuclear conflict pose the most severe existential risks, capable of driving humanity to extinction by 2100. While the experts stressed that these threats interact and must be addressed together, they emphasised that practical solutions exist for each challenge.
The commission modelled potential health losses (measured in disability-adjusted life years, or Dalys) from each threat. Nuclear conflict could cause extinction through nuclear famine triggered by atmospheric particles blocking sunlight, whilst AI-created biological weapons present another existential pathway. Other major threats include air pollution, antimicrobial resistance, childhood malnutrition and climate breakdown, each potentially causing over 1 billion Dalys. The report found that limiting global heating to 1.8°C rather than 4.4°C could prevent 1.89 billion years of healthy life being lost, whilst proposed cuts to international health aid could result in up to 1.5 billion Dalys by 2100, with nearly 70 per cent of losses concentrated in sub-Saharan Africa.
- Lancet Commission identifies AI and nuclear war as top extinction threats by 2100
- 17 major threats to human health identified, but solutions exist for each
- Aid cuts could cause 1.5bn Dalys, mostly in sub-Saharan Africa
New here? Start with this
Nuclear weapons have threatened civilisation since their development in the 1940s, capable of causing massive destruction and secondary effects such as atmospheric cooling that could trigger global famine. Artificial intelligence is a rapidly advancing technology that presents a newer concern: whilst offering many benefits, it risks being misused to cause widespread harm in ways existing institutions may struggle to control. Both represent dangers that experts believe could threaten human survival itself.
Beyond these dramatic risks, health and environmental experts track multiple other serious threats to human wellbeing, including pandemics, climate change, air pollution, antimicrobial resistance and malnutrition. These dangers do not operate independently; they interact and can amplify each other. A warming climate, for instance, can facilitate disease spread and worsen food security, while antimicrobial resistance undermines our ability to treat infections—challenges that reinforce each other.
Over recent decades, scientists and policy experts have developed frameworks to assess and compare such varied threats, attempting to quantify their potential harm and identify priorities for global action. The Lancet Commission, which brings together leading health experts, is one of many organisations that periodically undertake comprehensive risk assessments. Such evaluations aim to inform government decision-making and direct international resources towards the most pressing challenges.
Both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
Expert systematic assessment of catastrophic risks is essential precisely because the consequences of inaction are irreversible—even modest probabilities of extinction warrant rigorous analysis and prevention. These identified threats represent documented realities: advanced AI systems exist, nuclear weapons remain deployed, and climate tipping points are scientifically established. The Lancet Commission's approach provides a crucial framework for comparing risks and identifying intervention points before scenarios become unmanageable, making such assessment a responsible foundation for policy.
The case against
Whilst individual threats merit attention, prioritising speculative extinction scenarios over demonstrable present-day health crises risks systematic misallocation of limited resources. Historical catastrophe predictions frequently fail, and the report itself notes solutions exist, suggesting these represent significant but manageable risks rather than true extinction-level threats. Evidence-based public health has succeeded through targeted intervention on current disease, malnutrition and poverty; resources diverted to preparedness for low-probability scenarios might yield more lives saved by addressing the certain, measurable suffering affecting billions today.
Read the full article at the source →
Originally published by The Guardian as “Nuclear war and malicious use of AI threaten end of humanity by 2100, says Lancet assessment”.