Archive · 10 Sep 2026
NHS Rules Face Challenge From Changing Healthcare AI
AI is becoming more common in NHS healthcare.
The medical regulator says current rules were mostly designed for devices that do not change after approval.
Some AI systems can learn from new data and behave differently later.
The regulator has suggested 44 ways to improve oversight.
These include checking systems continuously and removing approval if they stop working well.
Patients might also be told when AI is involved in their care.
AI scribes can save doctors time by creating consultation notes, but they can make mistakes.
Some patients may share less personal information if they know AI is listening.
Doctors, patients, and regulators must balance AI’s possible benefits with its risks.
The Medicines and Healthcare products Regulatory Agency has published 44 recommendations for regulating healthcare AI.
The proposals include continuous monitoring, possible removal of approval, and penalties for developers whose products fail standards.
Patients could gain rights to know when AI is used in their care and how it operates.
AI systems can change after approval as they learn from new data, unlike many traditional medical devices.
AI scribes are reportedly used by around 40% of UK GPs, but studies and clinicians have raised concerns about privacy, errors, and trust.
- Who
- The Medicines and Healthcare products Regulatory Agency, an independent commission, patients, clinicians, and healthcare AI developers are involved.
- What
- The MHRA has proposed 44 recommendations to update regulation of AI products used in healthcare.
- When
- The recommendations were published as AI is increasingly being used in NHS healthcare; the articles do not give a publication date.
- Where
- The proposals concern healthcare in Britain, including the NHS and UK general practices.
- Why
- Existing medical-device rules were designed for products that generally remain stable after approval, while AI can learn, adapt, or drift after deployment.
Healthcare AI Opportunity
Healthcare AI Risks
Efficiency and medical progress
Healthcare AI Opportunity
Supporters say AI could save doctors time, improve healthcare, and become a major advance comparable to antibiotics or MRI.
Healthcare AI Risks
Critics point to incorrect advice, biased training data, and confident but wrong conclusions as potential harms.
Patient participation
Healthcare AI Opportunity
Patients could benefit from AI-assisted consultations and greater transparency through a right to know when AI is being used.
Healthcare AI Risks
Some patients may refuse AI involvement or disclose less sensitive information when they know a system is processing their consultation.
Responsibility and oversight
Healthcare AI Opportunity
Continuous monitoring, supervised trials, and human checking could help keep AI products safe as they change.
Healthcare AI Risks
AI systems can evolve after approval, raising questions about whether doctors can realistically remain responsible for every clinical decision involving AI.
- Regulator
- Medicines and Healthcare products Regulatory Agency (MHRA)
- Recommendations
- 44 proposed recommendations
- Consultation
- More than 12,000 people, including patients and clinicians, were consulted
- Potential patient right
- To know when AI is involved in their care and access information about the products used
- Proposed oversight
- Continuous monitoring, possible removal of approval, and penalties for developers
- Current NHS use
- AI-powered medical scribes are reportedly used by around 40% of UK GPs
- Proposed testing measure
- An AI “L plate” system for supervised trials by healthcare professionals