‘Superhuman’ AI Could Predict Early Death Risk

A “superhuman” artificial intelligence (AI) system is being tested by the National Health Service (NHS) to assess its ability to predict the risk of disease and early death in patients. The AI tool, called AI-ECG risk estimation, or Aire, uses the results from a simple heart test to flag patients who may require further examination for heart conditions.

If the trials prove successful, Aire could become a regular feature in the health service within the next five years.

While AI has already demonstrated its capabilities in generating images, videos, and engaging in casual conversations, researchers are increasingly focusing on its potential to combat diseases and enhance our understanding of the human body. Aire does this by analyzing electrocardiogram (ECG) results, which measure heart rate, electrical activity, and rhythm to diagnose conditions such as heart attacks in adults and heart diseases in children.

The AI goes beyond standard diagnostic tools by analyzing genetic information and other hidden details that doctors cannot easily detect. It is capable of predicting not only the likelihood of death but also providing a “time-to-mortality” estimate, creating a “survival curve” from a single ECG test.

Trials for this advanced AI program are set to begin in 2025 at Imperial College Healthcare NHS Trust and Chelsea and Westminster Hospital NHS Foundation Trust, with hundreds of patients expected to participate. Additional trials will follow at other hospital sites.

Dr. Fu Siong Ng, a cardiac electrophysiology expert at Imperial College London and consultant cardiologist at Imperial College Healthcare NHS Trust, explained that the initial studies aim to confirm the accuracy of the AI in detecting diagnoses from ECG tests. If successful, the next phase will focus on using the tool to alter the treatment course for patients.