The NHS patient advocacy organization Healthwatch England has raised concerns about the accuracy and safety implications of artificial intelligence scribes being deployed across the health service. These AI systems transcribe conversations between healthcare providers and patients, but documented cases show they can make significant errors in recording medication names, medical diagnoses, and clinical recommendations.
One notable case involved a patient who discovered her AI-generated summary incorrectly stated she had demyelination, a serious neurological condition associated with multiple sclerosis, when the actual finding should have been “null demyelination.” The error caused considerable distress and was only corrected after the patient herself questioned the record. In another instance, an AI scribe confused a prescribed medication with a different drug having a similar name. Healthwatch highlighted that patients, rather than healthcare professionals, have frequently identified these errors, suggesting that inaccuracies may persist in medical records if patients do not catch them.
The technology is being rolled out rapidly as part of the government’s 10-year health plan to modernize the NHS and reduce administrative burdens on staff. However, Healthwatch emphasized the need for clearer mechanisms allowing patients to report and correct AI-related mistakes. The regulatory approach has also drawn criticism, as the Medicines and Healthcare products Regulatory Agency decided not to classify AI scribes as medical devices, meaning they lack England-wide oversight for safety and effectiveness. Currently, 27 different AI scribes are in use across GP surgeries and hospitals in England.
Experts have noted additional challenges with the technology. Research indicates that AI errors may increase when consultations involve multiple participants, patients with complex medical histories, or individuals whose first language is not English. Some doctors have also reported instances of AI “hallucinations,” where the system references medical discussions or medications that were never mentioned during the consultation. Despite these concerns, survey data showed that over half of GPs surveyed believed their AI records were more accurate than handwritten notes, though some medical professionals caution that the time required to review transcripts for errors may undermine efficiency gains.
Article Attribution | Read More at Article Source
Article summary produced by Claude AI