A new AI model can forecast a person’s risk of more than 1,200 diseases

What if you could walk into a doctor’s clinic and receive bespoke health advice - put together using AI that has modelled your personal disease risks over the coming decade? That’s the idea behind the AI model Delphi-2M, developed by the European Molecular Biology Laboratory, the German Cancer Research Centre and the University of Copenhagen. Delphi-2M is trained on millions of anonymised health records, spotting patterns which it can apply to forecast the likely health outcomes for any given individual. The model’s predictions aren’t perfect, and it still has to go through trials and refinements before it’s used in any health system. But Delphi-2M is already providing insights for scientists, by identifying clusters of diseases which hint at unknown links, and predicting overall disease burdens up to 20 years in advance, which may allow health systems to prepare.