Artificial intelligence is making its way deeper into healthcare, and the UK is taking a practical approach to figuring out what actually works.
The UK’s Medicines and Healthcare products Regulatory Agency (MHRA), together with Manchester University NHS Foundation Trust, has launched the Manchester Sandbox, a programme designed to test new healthcare technologies, including AI, in real NHS settings.
Instead of allowing promising technologies to move straight from the laboratory into hospitals, the programme gives developers an opportunity to test them alongside doctors, healthcare workers and patients. This means researchers can see how these systems perform when faced with the complexities of everyday healthcare.
Some of the technologies being considered include AI-enabled medical devices that can identify patients who may be at greater risk of complications from long-term illnesses. In theory, spotting those warning signs earlier could give healthcare professionals more time to intervene.
But that’s also where the importance of testing becomes clear. An AI system might perform impressively in a controlled environment and behave very differently when confronted with incomplete information, unusual cases or the realities of a busy hospital.
The programme will therefore look beyond whether an AI system simply works. Regulators and healthcare professionals will examine how safely it operates, how it affects patient care and whether it can genuinely fit into existing NHS services.
If successful, the approach could make it easier to identify useful technologies and introduce them more widely, while filtering out systems that aren’t ready for patients.
And as AI becomes increasingly involved in decisions that affect people’s health, that distinction matters. The question is no longer simply “Can AI do this?” but “Can it do this safely enough to trust with someone’s care?”
