Nuclear war and AI risk: what reports say now

According to international press reports, an assessment linked to the medical journal The Lancet deals with nuclear war and artificial intelligence, an unusual choice of subject in the context of a health journal. The Guardian’s October 2026 report says the Lancet commission assessment identified nuclear war and the malicious use of AI as the risks posing the greatest threat to humanity’s survival up to 2100. Below, based on the international coverage, we look at what the researchers examined according to the papers, why these two dangers in particular came out on top, and why the assessment should not be read as a prophecy of doom.

What did the assessment examine, according to the reports?

A 10 October report from CNN Health says the researchers identified 17 potentially catastrophic threats that could endanger humanity’s future. According to The Chosun Daily, the analysis covers a time horizon of about 75 years, that is, the period up to 2100.

A summary describing the Lancet commission’s work, linked to the Institute for Health Metrics and Evaluation (IHME), also refers to 17 global threats. Based on the reporting, the reason the topic appeared in the orbit of a medical journal is that the authors approach the risks primarily in terms of their impact on human health.

Why nuclear war in particular?

The destructive power of nuclear war is well known, but according to the Guardian’s report, the assessment also warns of consequences beyond the immediate explosions. Based on the paper’s account, particles reaching the stratosphere could lead to the collapse of agriculture.

This line of thinking explains why a nuclear conflict can also be seen as a public health issue. The victims of such a war would not only be in the directly affected regions, but also in places where famine could emerge because of a fall in food production.

The danger of AI: not the “rebelling machine”, but biological weapons

In public debate, the risks of AI often appear in the form of an autonomous system turning against humans. According to the reports, the assessment places the emphasis elsewhere. Articles by The Chosun Daily and bdnews24 both say that the malicious use of AI presents a direct danger chiefly because it could make it easier to design dangerous pathogens and biological weapons.

In this reading, then, the risk does not arise from the technology itself, but from who uses it and for what purpose. India’s Kalinga TV said that the combination of AI and biological weapons creates a new, asymmetric threat. This means that a smaller actor with limited resources could also cause disproportionately large harm if it gained access to the necessary knowledge.

How can catastrophe be measured? The DALY indicator

According to The Chosun Daily, the researchers used the disability-adjusted life years (DALY) indicator to rank the threats. The same method is also mentioned in the IHME-linked overview.

DALY is an established public health metric. It takes into account both the years lost because of premature death and the years people live with illness or impaired health. According to bdnews24, the analysis model also accounts for the decline in quality of life caused by ill health, so it is not only the number of deaths that matters, but also the condition in which survivors would continue to live.

Low probability, enormous stakes

The reports do not all show the same degree of nuance. bdnews24 highlights that the probability of events leading to total extinction is individually low, but their impact would be apocalyptic. Several shorter summaries, including a report by Ukraine’s UNN, present the scenarios up to 2100 more as a direct extinction risk, placing less emphasis on the question of probability.

It is not clear from the press reports whether the commission gave specific percentage probabilities. For that reason, any claim attributing precise odds to the assessment should be treated with caution. Based on the papers’ summaries, the key point is rather that rare but extremely severe events also deserve serious attention.

The risks may reinforce one another

According to CNN’s report, the researchers believe the risks should be treated as complex systems and regard targeted action as necessary. This approach matters because major threats often do not appear in isolation.

According to the reporting, AI could, for example, increase the risk posed by biological weapons, while based on the Guardian’s report, a nuclear conflict could turn into a health crisis through its effects on food production. If the threats were examined separately, it is precisely these interconnections that would remain hidden.

It is not the first time AI and nuclear war have appeared together

The linking of the two dangers is not new. According to the May 2023 statement on AI extinction risk published by the Center for AI Safety, “Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war.” According to Wikipedia’s summary, the statement was signed by hundreds of AI researchers and company leaders.

That statement at the time was primarily a warning from the technology sector. Based on the reports, the significance of the current assessment lies in the fact that it ranks the same dangers within a health-focused, metrics-based framework.

What is known for certain, and what is conveyed only by the papers?

What is certain is that nuclear war and artificial intelligence were already mentioned together in 2023 in a widely signed professional statement. The details of the recent Lancet-linked assessment — the 17 threats, the time horizon up to 2100, the DALY-based ranking and the two highlighted dangers — are conveyed by international press reports, so they are best treated as accounts given by those papers.

The reports consistently suggest that the assessment is not claiming catastrophe is inevitable, but rather that certain risks are so severe that they require targeted action even at low probability. Since both highlighted dangers depend on human decisions, prevention is also primarily a matter of policymaking and international cooperation. The practical lesson from news reports about such risk assessments is that it is worth checking who is presenting the data, whether probability and time horizon are mentioned, and treating critically articles that, without these, present the analysis simply as a prophecy of doom.