Sky News Daily Is the NHS ready for an AI revolution?
Sep 17, 2025
Thomas Moore, Sky's science and medical correspondent, delves into how AI tools like Delphi could revolutionize disease prediction in the NHS. He discusses the potential of AI to predict chronic conditions years in advance, but raises concerns about patient reactions and implementation hurdles. Dr. Ben Maruthappu shares insights on how AI can assist frontline staff and support home care. They explore the balance between technological advancement and the essential human touch in healthcare, emphasizing the need for oversight and consistent systems.
AI Snips
Chapters
Transcript
Episode notes
AI Learns Disease Trajectories From Data
- Researchers trained an AI (Delphi) on large health datasets to spot early signals of disease before symptoms appear.
- The model finds statistical patterns in records to predict likely future diagnoses across many conditions.
Large Datasets Improve Predictive Power
- Delphi was trained on 400,000 UK Biobank records and validated on 1.9 million Danish patient registry records.
- It predicts many conditions well (eg heart disease) but struggles with infectious disease and psychiatric conditions.
Early Prediction Enables Prevention — With Caution
- Early prediction could enable lifestyle or medical interventions to prevent or slow chronic disease progression.
- Researchers warn psychological impact and public reaction need assessment before clinical rollout.