AI Adoption Accelerates in UK General Practice Despite Safety and Legal Concerns

Key Takeaways: 

• Nearly three in 10 GPs in the UK are already using AI tools, including generative systems such as ChatGPT, during patient consultations and for administrative tasks.
  
• Widespread concern remains about clinical risk, liability and data security, with most GPs warning that the current lack of national regulation leaves them exposed.
  
• Time saved through AI is largely used to reduce burnout rather than increase appointments, challenging policy assumptions about productivity gains.
  

Growing use of AI in GP consultations

Almost 30 per cent of general practitioners in the UK are now using artificial intelligence tools in consultations with patients, according to new research, despite concerns that such use could lead to clinical errors and legal action.

The findings highlight how rapidly AI has moved into everyday general practice, largely as a response to intense workload pressures. Tools such as ChatGPT are being used to support tasks including appointment summaries, elements of clinical reasoning and routine administrative work.

However, this expansion is taking place in what researchers describe as a largely unregulated environment, leaving many clinicians uncertain about which tools are safe, appropriate and compliant with NHS standards.

Findings from the Nuffield Trust and RCGP survey

The research was conducted by the Nuffield Trust thinktank and is based on a survey of 2,108 family doctors carried out by the Royal College of General Practitioners, alongside focus groups involving GPs.

In total, 598 respondents, representing 28 per cent of those surveyed, said they were already using AI in their work. Usage varied notably across demographic and geographic lines, with 33 per cent of male GPs reporting use compared with 25 per cent of female GPs. Uptake was also significantly higher in more affluent areas than in deprived communities.

The study found that AI is most commonly being used to:

• generate summaries of patient consultations
• assist with aspects of diagnosis
• support routine administrative and documentation tasks
  

Ministers have expressed hopes that AI could help reduce waiting times and improve access to general practice. However, the report suggests that the reality on the ground is more complex.

Concerns about risk, liability and data security

Despite the pace of adoption, large majorities of GPs, including both users and non-users of AI, expressed serious concerns about the risks involved. According to the report, clinicians fear that practices adopting AI could face “professional liability and medico-legal issues”, alongside “risks of clinical errors” and challenges relating to “patient privacy and data security”.

Dr Becks Fisher, a GP and director of research and policy at the Nuffield Trust, said the current situation falls far short of national ambitions.

“The government is pinning its hopes on the potential of AI to transform the NHS. But there is a huge chasm between policy ambitions and the current disorganised reality of how AI is being rolled out and used in general practice”, she said.

Dr Fisher added that uncertainty around regulation is undermining confidence among clinicians.

“It is very hard for GPs to feel confident about using AI when they’re faced with a wild west of tools which are unregulated at a national level in the NHS.”

Inconsistent guidance across the NHS

The report also highlights variation in local policy. While some NHS integrated care boards actively support GPs in using AI tools, others prohibit their use altogether. This inconsistency adds to confusion and reinforces concerns about accountability and governance.

Productivity gains do not translate into more appointments

In a setback for policymakers, the research found that time saved through AI adoption is not typically used to see more patients. Instead, GPs reported using that time to manage exhaustion and prevent burnout.

“While policymakers hope that this saved time will be used to offer more appointments, GPs reported using it primarily for self-care and rest, including reducing overtime working hours to prevent burnout”, the report states.

Evidence from wider academic research

Similar conclusions were reached in a separate study published last month in the journal Digital Health. That research found that the proportion of UK family doctors using AI rose from 20 per cent to 25 per cent over the course of a single year.

Dr Charlotte Blease of Uppsala University in Sweden, the study’s lead author, described the pace of change as striking.

“In just 12 months, generative AI has gone from taboo to tool in British medicine”, she said.

Like the Nuffield Trust, Dr Blease emphasised the risks of adoption without adequate safeguards.

“The real risk isn’t that GPs are using AI. It’s that they’re doing it without training or oversight.”

She added: “AI is already being used in everyday medicine. The challenge now is to ensure it’s deployed safely, ethically and openly.”

Patients increasingly turning to AI

The growing role of AI is not limited to clinicians. Healthwatch England reports that increasing numbers of patients are also using AI tools to support their healthcare, particularly when they struggle to access GP appointments.

“Our recent research shows that while patients continue to trust the NHS for health information, around one in 10 (9%) are using AI tools for information on staying healthy”, said Chris McCann, deputy chief executive of Healthwatch England.

He noted that access barriers and convenience are driving this trend, but warned about variable quality.

“There are various reasons people may turn to AI tools, including when they cannot access GP services. However, the quality of the advice from AI tools is inconsistent. For example, one person received advice from an AI tool that confused shingles with Lyme disease.”

Government response and next steps

In September, the government launched a commission to examine how AI can be used safely, effectively and within an appropriate regulatory framework across healthcare. The commission is expected to publish recommendations on governance, oversight and implementation when it reports.

Until then, the research suggests that AI will continue to spread in general practice faster than the systems designed to regulate and support its use.