Key Takeaways Delphi-2M AI model can forecast risk for more than 1,000 diseases up to 20 years in advance with accuracy comparable to single-disease models. Trained on 400,000 UK Biobank participants...
European researchers have developed a groundbreaking artificial intelligence model that can predict a person's risk of developing more than 1,000 diseases decades before symptoms appear, marking a significant advance in predictive medicine.
The AI system, called Delphi-2M, was trained on anonymized medical records from 400,000 participants in the UK Biobank and successfully validated using data from 1.9 million patients in the Danish National Patient Registry.
The research, published in the journal Nature on September 17, 2025, represents one of the most comprehensive demonstrations of AI's potential in modeling human disease progression.
Delphi-2M employs a modified version of the transformer architecture used in large language models like ChatGPT, but instead of processing text, it analyzes sequences of medical events over time.
The system learns to identify patterns in healthcare data, including medical diagnoses, lifestyle factors such as smoking and obesity, age, and gender.
"Understanding a sequence of medical diagnoses is a bit like learning the grammar in a text," explained Moritz Gerstung, an AI expert at the German Cancer Research Center and head of the division of AI in oncology. "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."
The model works by assessing probabilities rather than making definitive predictions. "So, just like weather, where we could have a 70% chance of rain, we can do that for healthcare," said Professor Ewan Birney, interim executive director of the European Molecular Biology Laboratory (EMBL). "We can do all diseases at once and over a long time period."
Delphi-2M predicts the rates of more than 1,000 diseases, conditional on each individual's past disease history, with accuracy comparable to that of existing single-disease models. The system can provide meaningful estimates of potential disease burden for up to 20 years into the future.
The model performs particularly well for conditions with clear and consistent progression patterns, such as certain types of cancer, heart attacks, and septicemia (blood poisoning).
However, it is less reliable for more variable conditions, including mental health disorders and pregnancy-related complications that depend on unpredictable life events.
"Medical events often follow predictable patterns," said Tom Fitzgerald, Staff Scientist at EMBL's European Bioinformatics Institute (EMBL-EBI).
"Our AI model learns those patterns and can forecast future health outcomes. It gives us a way to explore what might happen based on a person's medical history and other key factors. Crucially, this is not a certainty, but an estimate of the potential risks."
The research team envisions the technology being deployed in clinical settings within five to ten years, fundamentally changing how doctors approach preventive care.
"The future, and this is five to 10 years away – is when clinicians are enhanced and supported by these sophisticated AI tools," Birney explained. "You walk into the doctor's surgery and the clinician is very used to using these tools, and they are able to say: 'Here's four major risks that are in your future and here's two things you could do to really change that.'"
Beyond individual patient care, the system could have broader healthcare system implications. On a larger scale, such tools could help with "optimization of resources across a stretched health care system," Fitzgerald noted.
"Our AI model is a proof of concept, showing that it's possible for AI to learn many of our long-term health patterns and use this information to generate meaningful predictions," said Birney. "By modeling how illnesses develop over time, we can start to explore when certain risks emerge and how best to plan early interventions. It's a big step towards more personalized and preventive approaches to health care."
Despite its promising capabilities, researchers emphasize that Delphi-2M requires further testing before clinical implementation. The current datasets from Britain and Denmark are biased in terms of age, ethnicity, and health outcomes, potentially limiting the model's applicability to more diverse populations.
"This is still a long way from improved health care," Gerstung acknowledged, stressing that both datasets used for training and validation have inherent limitations.
Independent experts have noted these concerns. Peter Bannister, a health technology researcher and fellow at Britain's Institution of Engineering and Technology, commented that "both (British and Danish) datasets are biased in terms of age, ethnicity and current health care outcomes."
However, the research has received positive reception from the AI medical community. Gustavo Sudre, a professor at King's College London specializing in medical AI, described the work as "a significant step towards scalable, interpretable and—most importantly—ethically responsible predictive modeling."
"This is the beginning of a new way to understand human health and disease progression," said Gerstung.
Key takeaways OpenAI has signed a multi-year, $10 billion agreement with AI chipmaker Cerebras Systems to secure computing infrastructure. The deal will deliver 750 megawatts of computing power throug...
KEY TAKEAWAYS: xAI implemented restrictions preventing Grok from editing images of real people in revealing clothing after global backlash California Attorney General Rob Bonta launched investigation...
Key takeaways TSMC reported fourth-quarter net income of NT$505.74 billion, beating analyst expectations and setting a new record. Fourth-quarter revenue reached NT$1.046 trillion, up 20.5% year over...
