A new AI initiative in the UK is aiming to transform how healthcare decisions are made. The model, known as Foresight, is being trained on de-identified data from 57 million people across England, making it one of the largest health AI training efforts in the world.
Developed through a collaboration between University College London and King’s College London, Foresight uses routinely collected NHS data, such as hospital admissions and Covid-19 vaccination records, to predict health outcomes and identify opportunities for early interventions.
The data is stored and accessed through NHS England’s Secure Data Environment, ensuring privacy by removing personally identifiable information. With this massive and diverse dataset, researchers hope to train an AI model that can benefit all demographics, including groups often underrepresented in clinical research, such as minorities and those with rare conditions.
Dr. Chris Tomlinson, lead researcher from UCL’s Institute of Health Informatics, says the scale is key: «if we want a model that can benefit all patients, with all conditions, then the AI needs to have seen that during training.»
Foresight also aligns with the UK government’s broader push toward digital transformation in healthcare. With backing from the Wellcome Trust, the government has pledged up to £600 million to support national-scale use of NHS data.
Health Secretary Wes Streeting emphasized the potential impact: «AI will be central as we bring our analogue NHS into the digital age,» he said, aiming to cut hospital trips, accelerate diagnoses, and ease the burden on healthcare staff.
If successful, Foresight could be a major leap forward in using AI not just for innovation, but for practical, equitable care.
Article written by Robert Barrie
08/05/2025
Source:
Medical Device Network