September 17, 2025

New AI model for predicting individual risk of disease over decades

Researchers have developed a new generative AI model (Delphi-2M) that uses anonymised large-scale health records to estimate how a person’s health may change over time. It can forecast the individual risk and timing of over 1,000 diseases and predict health outcomes over a decade in advance. While it’s not ready for direct clinical use just yet, the AI model currently offers new ways to study disease and inform healthcare strategies.

The AI model was trained on anonymised patient data from 400,000 participants from the UK Biobank. The researchers also successfully tested the model using data from 1.9 million patients in the Danish National Patient Registry, with those results being published in a paper in Nature.

Journalists came along to this SMC briefing to hear the experts who developed this model discuss their results, as well as the wider context and potential impact of this innovative technology, including:

• How does this new AI model work and what exactly can it predict about an individual’s future health risks?
• What are the potential uses and benefits for patients, healthcare professionals, and researchers over currently available tools?
• How can patients be reassured that their data is safe?
• How could this technology be used to address healthcare inequalities and support healthcare planning in the NHS and beyond?
• What are the next steps for this AI model in research and in healthcare?

Speakers included:

Dr Ewan Birney, Interim Executive Director, EMBL

Prof Moritz Gerstung, Visiting Group Leader, EMBL

Dr Tomas Fitzgerald, Research Staff Scientist, EMBL