AI Forecasts UK's One-in-Four Chance of Major War by 2036
Despite ongoing ceasefire discussions and market optimism, behind-the-scenes calculations paint a stark picture of Britain's future security landscape. According to a leading artificial intelligence firm, the United Kingdom faces a significant probability of being embroiled in a major war within the next decade.
The Startling Probability Assessment
Dr Keith Dear, a former military intelligence officer and co-founder of British AI company Cassi, has presented sobering calculations to the Commons Defence Committee. His firm's analysis indicates Britain currently has a one-in-four chance of involvement in a major conflict before 2036. This assessment was revised upward from a one-in-five probability following recent US-Israel strikes on Iran.
A major war is defined as a conflict resulting in either more than 500 British service personnel deaths or 250 British casualties in a conflict with at least 10,000 combatant deaths overall. The last time Britain experienced such a conflict was during the Korean War.
The AI Company Outperforming Tech Giants
Remarkably, Cassi – with just eleven employees – consistently outperforms industry giants in forecasting accuracy. According to AI industry-wide standard rankings, the British firm leads competitors including Anthropic, Google, Elon Musk's xAI, OpenAI, and Meta in predictive capabilities.
This achievement comes despite the vast resources and determination of these technology behemoths to dominate AI forecasting. Musk himself has declared that "the ability to predict the future is the best measure of intelligence."
Defence Spending as a Critical Factor
Dear presented compelling evidence to parliamentarians about how defence investment could dramatically alter these probabilities. He explained that if the UK consistently spends 3% of GDP on defence over a five-year average basis by 2036, the probability of British involvement in a major conflict would drop by 50% in relative terms – from 20% to 10%.
"The forecasts suggest that if the UK consistently spends 3% of GDP on a five-year average basis by 2036, then the probability of the UK being involved in a major conflict drops by 50 per cent in relative terms," Dear told the committee.
Russia Identified as Primary Threat
The analysis provides specific probabilities for different geopolitical scenarios. Conflict with Russia before 2036 is forecast at 17%, making it the most likely trigger for British involvement in a major war.
"Conflict with Russia before 2036 is forecast at 17 per cent – sufficiently close to the likelihood that there will be a major conflict to imply that if the UK were to be involved in a major conflict, it is most likely to be with Russia," Dear stated.
Taiwan Scenario Less Concerning for UK Direct Involvement
Regarding tensions in the Indo-Pacific region, Cassi calculates a 26% chance of China-Taiwan conflict. However, Dear noted that such an outbreak would only increase the likelihood of UK involvement in a major conflict by 11%.
This suggests Britain would be unlikely to suffer sufficient casualties to meet the "major conflict" threshold in a Taiwan scenario, though regional instability would undoubtedly have global consequences.
The Man Behind the Predictions
Dr Keith Dear brings substantial credentials to his forecasting work. He spent eighteen years in RAF intelligence, serving in Afghanistan, Iraq, Mali, as a peacekeeper in Georgia, and on exchange with the US Air Force. Following his military career, he served as a Downing Street expert adviser, leading on science and technology in the Integrated Review of Security, Defence, Foreign & Development Policy.
Before founding Cassi, Dear also served as managing director of the Centre for Cognitive & Advanced Technologies for Fujitsu. His company describes itself as "a superstrategy company" and a "truth-engine" that addresses "modern pathologies in decision-making" across society, government, and business.
The Methodology: AI Meets Moneyball
While protective of his precise methodology, Dear reveals his team was inspired by the "Moneyball approach" popularised by Michael Lewis's book and subsequent film about American professional baseball. However, applying this approach to geopolitics presents unique challenges.
"Moneyball works because baseball has abundant historical data and essentially stable rules," Dear explains. "War and geopolitics largely don't. Historical statistics are not just insufficient; they can actively mislead."
Cassi's innovation involves making expert judgement explicit and testable rather than relying on buried intuition. The company doesn't score geopolitical "players" but rather evaluates the track records of experts at predicting outcomes, then trains its AI to mimic the most successful approaches.
Quantifying Uncertainty for Better Decisions
Dear emphasises the power of quantification in decision-making processes. "If you say the probability of something bad occurring has moved from 'substantial' to 'severe', nobody listens," he observes. "If you say it's gone from 60 per cent to 80 per cent or higher, then they do."
This approach proves particularly valuable in Whitehall, where departments routinely present dire consequences without funding increases. Cassi's probabilistic assessments enable Treasury officials and Number 10 to evaluate the actual likelihood of such outcomes, informing more rational resource allocation.
Beyond Warfare: Broader Applications
Cassi's forecasting capabilities extend beyond military conflicts. The company applies similar methodologies to predict pandemic recurrences, commodity price movements, and political developments. One customer reportedly achieves substantial returns trading on Cassi's energy price forecasts.
Regarding specific current events, Cassi currently assesses a 25% probability of a popular uprising against the Iranian regime before 30 April 2026. Dear argues such specific, quantified assessments prove more useful than general political statements from world leaders.
The Growing Demand for Predictive Intelligence
As anxiety mounts over Britain's preparedness for potential conflicts, Cassi's services are increasingly sought by both governments and corporations. The company's ability to provide nuanced, probabilistic assessments of complex geopolitical scenarios fills a critical gap in traditional intelligence and forecasting.
While Dear emphasises that his company provides information rather than prescriptions – comparing it to house insurance where individuals choose their risk tolerance based on probabilities – the stark numbers his firm produces demand serious consideration from policymakers facing an increasingly uncertain global landscape.
