NHS AI tools to get ‘L-plates’ under new UK rules

New AI tools used in the NHS would run under supervision before earning full approval, in the same way learner drivers carry L-plates, under recommendations published on Thursday (10 September) by the National Commission into the Regulation of AI in Healthcare.

The Medicines and Healthcare products Regulatory Agency (MHRA) set up the independent commission in September 2025 and published its findings, which draw on evidence from more than 12,000 patients, clinicians, healthcare leaders and technology developers.

The commission wants the MHRA to introduce staged authorisations for new AI models. New models would go into use under close supervision and tight guardrails, then prove their real-world safety and performance before regulators grant fuller authorisation.

The commission said this approach would give UK patients world-first access to promising new models while keeping risk under control at every step.

The NHS already uses AI to spot strokes and skin cancers early, and to save clinicians’ time through voice-enabled technologies.

The report also calls for AI-enabled medical devices to face continuous, real-world monitoring for the whole of their working life, rather than a single approval at launch.

The Commission said this reflects how AI systems learn and change after deployment, and would let regulators act quickly if problems emerge.

“Generative AI is not like other medical technologies we’ve regulated before. It can behave differently in different circumstances, and it can evolve after its initial deployment. That means we can’t rely on a single point of approval and assume the job is done,” said Neil Lawrence, Chair of the Commission’s Technology Working Group and DeepMind Professor of Machine Learning at the University of Cambridge.

Patients would also gain the right to search for safety information about specific AI-enabled medical devices, including any adverse incidents.

The Commission said this would build on existing MHRA tools such as its interactive Drug Analysis Profiles.

The report sets out a proportionate approach to telling patients when AI is being used in their care, after patients told the commission they want to know.

The Commission also recommends stronger enforcement powers for the MHRA so the regulator can act decisively where AI systems fall short of the standards patients expect.

“People are open to AI improving their care, but only if it is safe, overseen by humans, and if they know when it is being used,” said Alastair Denniston, NHS consultant and Chair of the Commission.

“Patients have told us that to trust AI is safe, they want to know when it’s used, that it supports rather than replaces clinicians, and that there is clear accountability if things go wrong. That’s exactly what our recommendations deliver,” said Henrietta Hughes, NHS general practitioner, Deputy Chair of the Commission and Patient Safety Commissioner for England.

Research partner the Health Foundation published separate deliberative research on the same day, based on workshops with 78 members of the public in Cardiff, Milton Keynes and York between March and April 2026.

The work found accuracy is the public’s top priority, human oversight is a critical condition for using AI, and AI should not result in worse care for any group.

“The real test will be whether the NHS has the capacity, skills and systems in place to implement and monitor AI applications safely and effectively at the scale now needed,” said Jennifer Dixon, Chief Executive of the Health Foundation.

The Royal College of Radiologists urged the government and the MHRA to put the recommendations into practice without delay, and said it will support AI education for all doctors through a national lecture series.

The government and the MHRA will now consider the recommendations, with a formal response to follow in due course.

Now read: New ‘hash matching’ tool will stop certain images being shared in the UK