UK medicines watchdog sets out 44 rules for AI in the NHS

The UK’s Medicines and Healthcare Products Regulatory Agency (MHRA) published 44 recommendations on 10 September 2026 for new rules covering AI products used in the NHS and other healthcare settings. The regulator, which licenses treatment drugs and regulates medical devices in the UK, said the existing framework was not designed for software that keeps changing after it is approved.

What does the MHRA want to change?

The recommendations were compiled by an independent commission with input from more than 12,000 people, including patients and clinicians. They include continuous monitoring of AI products, with regulatory approval withdrawn if a product malfunctions or becomes less effective over time. Patients would gain the right to know whether AI is involved in their care, plus easy access to information about the products used. Regulators would also get the power to penalise developers whose AI fails to meet required standards, and an AI “L plate” system would let healthcare professionals trial new models under close supervision.

Why does the MHRA say existing rules fall short?

MHRA chief Lawrence Tallon said the medical devices regulatory framework “predominantly dates from a period where we were thinking about things like hip replacements and knee replacements, or smaller things like stethoscopes and plasters”. He said current guidelines may work for simple AI trained to spot known symptoms on scans, but not for more complex models. “Unlike most of the medical products we’re used to regulating, these products continue to change after the point of authorization,” he said. “As new data gets fed in, they learn, they adapt, they drift.” Tallon said no single country or regulatory framework had yet solved the problem.

How is AI already used in UK healthcare?

AI note-takers known as scribes, which are powered by large language models, are reportedly used by 40% of UK-based GPs to record consultations and generate reports. A recent study from the University of Edinburgh found patients could be less likely to share personal information such as substance abuse history if they knew a conversation was being processed by AI.

What do doctors and researchers say about the technology?

Henrietta Hughes, a GP on the report commission, said many of her patients were happy for AI to be used during their consultations, while some chose to opt out. “Some say, ‘I don’t want to talk to a robot’, and that is also fine,” she said. Hughes acknowledged that scribes could make mistakes in their notes but said correcting them was the responsibility of doctors. Alastair Denniston, an ophthalmologist who also worked on the report, called the technology “an exceptional opportunity” for healthcare that was “likely to rank alongside step-changes such as antibiotics and MRI”. In September, Rene Haas, chief executive of chip designer Arm, told the BBC he believed the technology would produce a cure for cancer “within our lifetime”.

Frequently asked questions

How many recommendations has the MHRA published on AI in healthcare?

The Medicines and Healthcare Products Regulatory Agency published 44 recommendations to update its policies as AI use in the health sector rises.

What is the MHRA?

The Medicines and Healthcare Products Regulatory Agency regulates all medical devices and licenses treatment drugs in the UK.

How many UK GPs use AI scribes?

AI note-takers known as scribes, powered by large language models, are reportedly used by 40% of UK-based GPs to record consultations and generate reports.

Who contributed to the MHRA review?

An independent commission compiled the report with input from more than 12,000 people, including patients and clinicians, among them GP Henrietta Hughes and ophthalmologist Alastair Denniston.

With reporting from BBC News.

Stock image, not from the events described. Doctors using transparent tablet with hologram medical technology (50812636116) — Vu Hoang / Wikimedia Commons (CC BY 2.0)