GPs in AI Wild West: Urgent Call for Oversight

Dr. Anya Sharma gazed at the glowing screen, the faint hum of her old laptop a familiar companion.

Another long day was winding down, but the administrative burden felt relentless.

She had heard whispers, seen colleagues experimenting, and even tried it herself – using a publicly available AI tool, like ChatGPT, to draft referral letters or summarise patient notes.

The speed was undeniably tempting, a lifeline in an era of overwhelming paperwork.

Yet, a knot of unease tightened in her stomach.

The room felt heavy with unspoken questions.

Was this AI use safe?

Was it ethical?

Who was accountable if the AI made a mistake, if a nuance was lost, or if patient data, sacred and confidential, inadvertently strayed?

She was navigating uncharted waters, driven by necessity but without a compass.

It felt like standing alone on the precipice of a new frontier, one that promised efficiency but whispered of unforeseen risks for patient safety AI.

The Unregulated Reality of AI in General Practice

Dr. Sharma’s experience is not unique; it is a daily reality for a growing number of general practitioners across the UK.

The promise of AI in general practice is vast: reducing administrative burden, streamlining clinical documentation, and even supporting professional development.

However, the path to realising these benefits is currently fraught with peril.

UK GPs are increasingly using AI tools, often independently sourced, but face a significant lack of national guidance and regulatory oversight.

This AI wild west environment raises serious concerns about patient safety, professional liability, and widening health inequalities, necessitating urgent national policy.

Recent research highlights the urgency of this situation for GP AI use.

More than 1 in 4 GPs (28%) are actively using AI tools in their practice, according to the Royal College of General Practitioners (RCGP) and Nuffield Trust’s GP Voice survey 2025.

This adoption, while promising for primary care innovation, is happening largely in a vacuum of clear, national healthcare AI regulation.

GPs Navigating an Ethical Minefield

The core problem, articulated clearly by experts, is that national policy on AI is simply failing to keep pace with rapid technological change within general practice.

This creates a challenging landscape where dedicated healthcare professionals, keen to leverage technology, are left to make critical decisions about patient care with insufficient guidance.

Rather than a reluctance to adapt, GPs are leaning into AI, but without the foundational support needed for safe and ethical deployment.

Imagine Dr. Anya Sharma trying to decide if an AI-generated summary of a complex mental health consultation is robust enough for a referral.

An AI scribe might be efficient, improving the quality of patient interactions by freeing her from note-taking during the appointment.

Yet, without clear standards or validation, the responsibility for its accuracy and the potential for bias falls squarely on her shoulders.

This scenario plays out daily, as GPs grapple with the implications for patient safety AI, data privacy healthcare, and their own professional liability.

The current postcode lottery of local guidance leaves many feeling exposed, described as flying blind by Professor Victoria Tzortziou-Brown, RCGP chair (RCGP & Nuffield Trust, 2025).

What the Research Really Says About AI in Primary Care

Findings from comprehensive studies paint a vivid picture of both the potential and the pitfalls of AI in general practice:

• Nearly one-third of UK GPs (28%) are now using AI tools.

  However, guidance from local NHS oversight bodies remains incredibly patchy, as noted by Dr Becks Fisher, GP and Nuffield Trust director of research and policy (RCGP & Nuffield Trust, 2025).

  GPs are actively seeking solutions for administrative burdens, even without formal support, underscoring the need for a strong, unified national strategy to guide this proactive adoption towards safe and effective medical AI tools.

• The use of AI varies significantly across the UK, highlighting potential digital health inequalities.

  England reported the highest use (31%), while Northern Ireland lagged at just 9%.

  There is also a disparity between male (33%) and female (25%) GPs, and a notable gap between affluent areas (35%) and socioeconomically deprived areas (27%), according to the RCGP & Nuffield Trust 2025 survey.

  This means the benefits of AI are not being equally distributed, risking widening health disparities.

  Any national NHS AI policy must include targeted support and resources to ensure equitable access and prevent further marginalization of vulnerable populations.

• A substantial 11% of GPs are using tools they obtained independently, such as ChatGPT, rather than practice-provided solutions (RCGP & Nuffield Trust, 2025).

  This shows GPs are innovative and resource-conscious.

  However, independent sourcing bypasses critical validation and oversight processes.

  There is an urgent need for nationally approved, validated medical AI tools and clear guidelines for the appropriate and safe use of general-purpose AI platforms in clinical settings.

• A staggering 84% of GPs are concerned about the lack of regulatory oversight for AI, a sentiment echoed by 83% of doctors in Medscape’s UK Doctors and AI Report 2024.

  The regulatory vacuum is a significant deterrent for many GPs and undermines confidence among current users.

  Robust, clear, and consistent healthcare AI regulation is not just desirable but essential for building trust and enabling widespread, safe AI integration.

A Playbook for Responsible AI Adoption in Primary Care

Navigating this complex landscape requires a clear, actionable approach for stakeholders, from practices to policymakers, to ensure that AI in general practice truly serves patients and practitioners.

• Champion Consistent National Guidance: Advocate for and implement clear, consistent national guidelines from bodies like the MHRA and NHS England.

  This directly addresses the 84% of GPs concerned about the lack of regulatory oversight (RCGP & Nuffield Trust, 2025).

• Invest in Targeted, Practical Training: Develop and roll out accessible training programs for GPs on how to use approved medical AI tools effectively and safely.

  This is crucial given that 11% of GPs are independently sourcing tools, indicating a knowledge gap and a desire for solutions (RCGP & Nuffield Trust, 2025).

• Prioritise Ethical AI Procurement and Validation: Ensure that any AI tools integrated into primary care are rigorously tested, validated for clinical accuracy, and adhere to strict data privacy and consent standards.

  This is key for ethical AI in medicine.

• Foster Collaborative Learning Networks: Encourage the creation of primary care networks or communities of practice where GPs can share experiences, best practices, and challenges with AI adoption.

  This helps mitigate the postcode lottery of local guidance.

• Bridge the Digital Divide: Implement specific initiatives and funding to support AI adoption in socioeconomically deprived areas and among underrepresented GP groups.

  This is vital to prevent widening digital health inequalities (RCGP & Nuffield Trust, 2025), addressing health technology policy.

• Maintain Human Oversight and Accountability: Always position AI as a powerful assistant, not a replacement for human clinical judgment.

  GPs must remain the ultimate arbiters of patient care, with a clear understanding of AI’s limitations.

• Engage Patients in the Conversation: Foster transparency with patients about AI use in their care, obtaining informed consent where appropriate, to build trust and address concerns about the doctor-patient relationship.

Risks, Trade-offs, and Ethics in the AI Era

While the promise of AI in general practice is compelling, ignoring the potential downsides would be a grave error.

The primary risks include patient safety AI incidents due to unvalidated or poorly understood algorithms, potential breaches of data privacy healthcare, and ambiguities around professional liability if AI tools contribute to misdiagnosis or suboptimal care.

There is also the very real risk of widening digital health inequalities if adoption continues unevenly, further disadvantaging patients in already underserved areas.

Mitigation demands a multi-faceted approach.

Robust data governance, ensuring patient data is handled with the utmost security (healthcare data security), is non-negotiable.

Transparency in AI use, coupled with clear informed consent processes, helps maintain the integrity of the doctor-patient relationship.

Continuous monitoring of AI tool performance, and a clear framework for auditing and reporting incidents, are crucial.

Ultimately, ethical AI deployment must be at the forefront of every decision, ensuring that technology serves humanity, not the other way around.

Tools, Metrics, and Cadence for AI Integration

To effectively integrate AI, practices and oversight bodies need a clear framework for tools, metrics, and review cadence.

While specific brand recommendations are beyond this scope, the focus should be on categories of medical AI tools:

• AI Scribes/Clinical Documentation Assistants: For efficient note-taking and record generation.

• Administrative Automation Platforms: For managing appointments, patient queries, and referral pathways.

• Validated Clinical Decision Support Systems: Tools that assist with diagnosis or treatment planning, only after rigorous national validation.

Key Performance Indicators (KPIs) to monitor include:

• Patient Safety Incidents Related to AI: Target a year-on-year decrease.

• GP Administrative Time Saved: Aim for a measurable reduction in hours spent on non-clinical tasks.

• Approved AI Tool Adoption Rate: Track the percentage of GPs effectively using nationally approved tools.

• GP Confidence in AI Use: Measure through regular surveys, aiming for an increase in reported confidence.

• Data Privacy Compliance: Maintain 100% compliance with all relevant regulations.

A regular review cadence is critical: a quarterly review for technical performance and user feedback, alongside an annual strategic review to assess policy effectiveness, ethical implications, and overall impact on primary care innovation.

FAQs on AI in General Practice

Q: How many GPs in the UK are currently using AI tools?

A: Approximately 28% of GPs in the UK are using AI tools in their practice, though there is significant variation across regions and demographics (RCGP & Nuffield Trust, 2025).

Q: What are GPs primarily using AI for?

A: GPs most commonly use AI for clinical documentation and note-taking (57%), professional development (45%), and administrative tasks (44%), aiming to reduce overtime and administrative burden (RCGP & Nuffield Trust, 2025).

Q: What are the main concerns GPs have about using AI?

A: GPs are primarily concerned about patient safety, professional liability, data privacy and consent, the impact on the doctor-patient relationship, and digital exclusion, especially due to the lack of national regulatory oversight (RCGP, 2025; RCGP & Nuffield Trust, 2025).

Q: Is there consistent guidance for GPs on AI use?

A: No, guidance is described as incredibly patchy, with GPs often feeling they are flying blind due to inconsistent local NHS oversight bodies and a lack of national regulation (Dr Becks Fisher and Professor Victoria Tzortziou-Brown, RCGP & Nuffield Trust, 2025).

Balancing Innovation with Patient Safeguards

Dr. Anya Sharma’s reflections mirror the collective sentiment of GPs across the UK – a blend of hope for a more sustainable future in primary care and trepidation about the path ahead.

The potential of AI to revolutionise healthcare is undeniable, a powerful force that could alleviate the crushing administrative burden and free up invaluable time for patient interaction.

Yet, without a guiding hand, this force risks becoming a chaotic tide.

The call for clear, consistent national guidance and robust regulatory frameworks for RCGP AI guidance is not just about protecting GPs; it is fundamentally about safeguarding patients and ensuring that the promise of AI serves everyone equally.

It is very hard for GPs to feel confident about using AI when they are facing a wild west of tools which are unregulated at a national level in the NHS, as Dr Becks Fisher noted.

The time for piecemeal solutions is over; it is time for a national strategy that empowers GPs, protects patients, and truly harnesses the transformative power of AI with confidence and clarity.

References

• Medscape. (2024). UK Doctors and AI Report 2024.

• Royal College of General Practitioners (RCGP). (2025). RCGP warnings and recommendations on AI in primary care.

• Royal College of General Practitioners (RCGP) & Nuffield Trust. (2025). RCGP’s GP Voice survey 2025.