‘From taboo to tool’: 30% of GPs in UK use AI tools in patient consultations, study finds | Health

Nearly three in ten GPs in the UK use AI tools such as ChatGPT when consulting with patients, even though it could lead to them making mistakes and being sued, a study reveals.

The rapid adoption of AI to ease workloads is accompanied by a lack of regulation of the technology, which leaves GPs in the dark about which tools can be used safely. That’s the conclusion of a study by think tank Nuffield Trust, based on a survey of 2,108 family doctors by the Royal College of GPs on AI and focus groups of GPs.

Ministers hope AI can help reduce the delays patients face in seeing a GP.

The study found that more GPs were using AI to produce summaries of patient appointments, helping them diagnose patient conditions and carry out routine administrative tasks.

In total, 598 (28%) of the 2,108 respondents said they were already using AI. More male (33%) than female (25%) GPs have used it and many more use it in affluent areas than in poorer areas.

Its use is rapidly becoming more and more widespread. However, a large majority of GPs, whether they use it or not, are concerned that practices adopting it could face “professional liability and medico-legal issues”, and “risks of clinical errors” and “patient privacy and data security” issues, the Nuffield Trust report said.

“The government is pinning its hopes on the potential of AI to transform the NHS. But there is a huge gulf between policy ambitions and the current disorganized reality of how AI is deployed and used in general practice,” said Dr Becks Fisher, GP and director of research and policy at the think tank.

“It is very difficult for GPs to use AI with confidence when they are faced with a Wild West of tools that are not nationally regulated in the NHS,” she added.

While some regional NHS integrated care boards support GPs using AI, others ban it.

In a blow to ministerial hopes, the survey also reveals that GPs are using the time it saves to recover from the stress of their busy days rather than seeing more patients. “While policymakers hope this saved time will be used to offer more appointments, GPs reported using it mainly for self-care and rest, including reducing overtime to avoid burnout,” the report adds.

A separate study of how UK family doctors are using AI, published last month in the journal Digital Health, found similar results. The proportion using AI was found to have increased from 20% to 25% compared to the previous year.

“In just 12 months, generative AI has gone from taboo to tool in British medicine,” said lead author of the study Dr Charlotte Blease of Uppsala University in Sweden.

Like the Nuffield Trust, it highlighted the lack of regulation as a major concern, particularly given the speed with which GPs are integrating AI into their clinical practice. “The real risk is not that GPs use AI. It’s that they do so without training or oversight,” Blease said.

“AI is already used in everyday medicine. The challenge now is to ensure it is deployed in a safe, ethical and open way.”

A growing number of patients are also using AI to improve their healthcare, particularly when they cannot get an appointment with a GP, Healthwatch England said.

“Our recent research shows that although patients continue to trust the NHS for information about their health, around one in 10 (9%) use AI tools to obtain information about their health,” said Chris McCann, deputy chief executive of the patient watchdog.

“There are a variety of reasons why people may turn to AI tools, including when they cannot access the services of a GP. However, the quality of advice provided by AI tools is inconsistent. For example, one person received advice from an AI tool that mistook shingles for Lyme disease.”

A commission launched by the government in September, into how to ensure AI is used safely, effectively and properly regulated, will make recommendations in its report.

The Department of Health and Social Care has been contacted for comment.