KEY POINTS
- Delphi-2M trained on UK Biobank data
- Uses transformer architecture like ChatGPT
- Early but promising for preventative medicine
- Needs more testing before clinical use
ISLAMABAD: Scientists announced that they have developed an AI model capable of predicting medical diagnoses several years in advance, using the same technology that powers consumer chatbots like ChatGPT.
AI CAN NOW PREDICT 1,000 DISEASES
A new AI tool can predict a person’s risk of more than 1,000 diseases, according to experts. The tool, which was developed by a team of researchers at a major academic institution, is a powerful example of how AI can be used to enhance medical… pic.twitter.com/xq8b534Yek
— The AI Showdown 24×7 (@ai_24x7) September 21, 2025
Drawing on a patient’s medical history, the Delphi-2M AI “predicts the rates of more than 1,000 diseases” years ahead, according to a team from British, Danish, German, and Swiss institutions in a paper published in the journal Nature.
The researchers trained the model using data from the UK Biobank – a vast biomedical research database containing information on around half a million participants.
AI learns medical patterns like language
Neural networks based on the so-called “transformer” architecture – the “T” in “ChatGPT” – are well-known for handling language-related tasks, as seen in chatbots and their numerous imitators and competitors.
New AI model predicts susceptibility to over 1,000 diseases https://t.co/s3rotM7D04
— Financial Times (@FT) September 17, 2025
However, understanding a sequence of medical diagnoses is “a bit like learning the grammar in a text,” explained Moritz Gerstung, an AI expert from the German Cancer Research Center.
Delphi-2M “learns the patterns in healthcare data, preceding diagnoses, in which combinations they occur and in which succession”, enabling “very meaningful and health-relevant predictions,” he added.
Validated on millions, but further testing needed
Gerstung presented data suggesting the AI could identify individuals at significantly higher or lower risk of a heart attack than predicted by age and other factors alone.
AI forecasts chances of disease more than 20 years ahead ✅
An algorithm built with ChatGPT-like technology predicts over 1,000 diseases from information provided by more than 400,000 @uk_biobank participants. pic.twitter.com/xbV3nX9DDS
— UK Biobank (@uk_biobank) September 17, 2025
The team validated Delphi-2M’s accuracy by testing it on nearly two million individuals from Denmark’s public health database.
Nonetheless, Gerstung and his colleagues emphasised that the Delphi-2M tool requires further validation and is not yet ready for clinical application.
Potential to transform preventative healthcare
“This is still a long way from improved healthcare as the authors acknowledge that both (British and Danish) datasets are biased in terms of age, ethnicity and current healthcare outcomes,” noted Peter Bannister, a health technology researcher and fellow at Britain’s Institution of Engineering and Technology.
New AI Tool Predicts Which of 1,000 Diseases Someone May Develop in 20 Years
A large language model called Delphi-2M analyzes a person’s medical records and lifestyle to provide risk estimates for more than 1,000 diseases
– Scientific American https://t.co/Gmk8CJyyEg pic.twitter.com/SMh18fm6lh
— Evan Kirstel #B2B #TechFluencer (@EvanKirstel) September 19, 2025
Looking ahead, systems like Delphi-2M could “guide the monitoring and possibly earlier clinical interventions for effectively a preventative type of medicine,” said Gerstung.
On a broader scale, such tools could aid in the “optimization of resources across a stretched healthcare system,” added Tom Fitzgerald, co-author from the European Molecular Biology Laboratory.
Broader impact on healthcare resource management
Currently, doctors in many countries use computer tools like the QRISK3 programme, which British GPs employ to estimate heart attack or stroke risk.
Delphi-2M, in contrast, “can do all diseases at once and over a long time period,” said co-author Ewan Birney.
Gustavo Sudre, a professor specialising in medical AI at King’s College London, described the research as “a significant step towards scalable, interpretable and – most importantly – ethically responsible predictive modelling.”
“Interpretable” or “explainable” AI remains a key focus in the field, since the inner workings of many large AI models remain opaque even to their developers.
