Doctors and the NHS may face medical negligence claims for harm caused by artificial intelligence tools used in patient diagnosis and treatment recommendations, according to warnings issued to ministers. Current law holds physicians and the health service responsible for patient injuries or deaths, even when AI systems produce the faulty outputs. The Medical Protection Society, which defends doctors in misconduct cases, states in a new report that clinicians risk becoming the main target for lawsuits over AI mistakes unless regulations are updated. The NHS increasingly relies on AI for tasks such as interpreting scans and X-rays, summarizing patient consultations, and preparing correspondence. Dr Sarah Townley, the society’s deputy medical director, noted that legislation has difficulty matching the speed of technological advances, creating a growing divide. The report highlights examples of possible harm, including an AI system overlooking a lung tumour on a chest X-ray, leading to delayed treatment and disease progression. Another scenario involves AI incorrectly advising a higher dose of the blood thinner warfarin, resulting in severe bleeding that requires surgery and intensive care. In these situations, the society warns, doctors could be held fully accountable under existing product liability rules. It calls for reclassifying AI tools as products under the Consumer Protection Act 1987 to shift responsibility away from clinicians. Medical professionals in the UK express growing concern about being held liable for AI-related mistakes, fearing this could erode public confidence in healthcare. Dr Ragit Varia, president-elect of the Society for Acute Medicine, stressed that innovation and safety must advance together and that legislation must not lag behind. NHS Resolution is preparing guidance on AI liability, the Department of Health and Social Care confirmed, adding that it will consider the report’s suggestions to maintain safe and rapid adoption of the technology. Ahmed Binesmael of the Health Foundation observed that public trust in AI depends on strong oversight and clear accountability measures.
