The MHRA has issued 44 recommendations to update medical device regulations as AI integration into the NHS increases.
Key facts
- •The MHRA report was developed by an independent commission with input from over 12,000 people.
- •Approximately 40% of UK-based GPs currently use AI-powered scribe tools to record patient consultations.
- •The MHRA's 44 recommendations aim to address the unique challenges of AI models that adapt and change after deployment.
- •A University of Edinburgh study found that patients may be less forthcoming with sensitive information when AI is used in consultations.
- •Arm CEO Rene Haas stated in September that he believes AI could help discover a cure for cancer within our lifetime.
The Medicines and Healthcare Products Regulatory Agency (MHRA) has published 44 new recommendations to update UK medical device policies. As AI becomes more integrated into the NHS, the watchdog aims to ensure patient trust and safety. MHRA chief Lawrence Tallon stated that current regulatory frameworks, originally designed for physical devices like hip replacements, are insufficient for complex, adaptive AI models.
By the numbers
Challenges of Regulating Adaptive Technology
The MHRA notes that while existing guidelines may cover simple AI tools trained to identify specific symptoms on scans, they do not address more complex systems. Unlike traditional medical products, AI models continue to evolve after authorization as they ingest new data, learn, and adapt. Tallon acknowledged that regulating this technology is a global challenge, noting that no single country has yet established a definitive regulatory framework.
AI in Clinical Practice and Patient Concerns
AI-powered note-takers, or scribes, are currently used by 40% of UK-based GPs to record consultations. While some patients welcome this technology, a study from the University of Edinburgh indicated that some individuals may withhold sensitive personal information, such as substance abuse history, if they know an AI is processing the conversation. GP and commission member Professor Henrietta Hughes noted that patients have the right to opt out of AI use during consultations. She emphasized that while AI scribes may make errors, the responsibility for correcting medical notes remains with the doctor.
Potential and Risks of AI in Healthcare
Experts see significant potential for AI in medical research and healthcare delivery, with ophthalmologist Professor Alastair Denniston comparing the technology's impact to that of antibiotics or MRI scans. However, the report also highlights concerns regarding potential biases in training data that could lead to incorrect decisions, as well as the risk of chatbots providing inaccurate medical advice.
Advertisement
This article was independently rewritten by ManyPress editorial AI from reporting originally published by BBC Health.
