Artificial intelligence has been making quiet inroads into UK healthcare for the better part of a decade. What began as ambitious government investment and cautious experimentation has gradually become something more tangible, tools that screen for cancer, summarise GP appointments, and monitor patients at home.
This is a look at how AI adoption in UK health clinics has unfolded year by year, what’s actually being used, and what the evidence is beginning to tell us about its long-term place in British healthcare.
2019–2021: Laying the Groundwork
The story of AI in the NHS starts formally in 2019, when the government launched the NHS AI Lab with a £250 million investment. Its stated goals were ambitious: earlier cancer detection, better dementia care, and more personalised treatment.
By early 2021, the Parliamentary Office of Science and Technology published a briefing on AI and healthcare, and NHS England Digital began sharing early research, including work on producing synthetic data from UK primary care patient records.
At this stage, AI in UK health was largely experimental: isolated pilots, academic research, and early feasibility studies.
2022: Regulation Catches Up
May 2022 brought the first major regulatory signal. The Care Quality Commission (CQC) published a review of machine learning applications being used for diagnostic purposes in NHS healthcare services. This marked a shift from “should we use AI?” to “how do we make sure AI is safe?”
2023: Momentum Builds - and the Gaps Show
By 2023, real-world AI tools were starting to reach clinical settings in more meaningful numbers. The Health Foundation published a major long read in November 2023 exploring what AI could mean for the NHS workforce, reflecting the growing interest at a national level.
Specific clinical use cases were beginning to prove themselves:
- Breast cancer screening: A 2023 study in The Lancet Oncology found that AI-supported mammogram screening was safe and reduced radiologist workloads.
- Diabetic retinopathy screening: AI tools were already being used in NHS clinics to analyse eye images for early signs of diabetes-related damage.
- COVID-19 diagnostics: AI had already been used during the pandemic to analyse chest imaging, giving the sector its first taste of AI at pace.
The GP surgery experience, however, was far patchier. Adoption was sporadic and largely driven by individual practices rather than any coordinated national push.
2024: The Year of Reckoning
2024 was a pivotal year, one in which the sector had to confront a growing tension between the enthusiasm for AI and the messiness of real-world implementation.
Investment and Regulation
The government committed £21 million to the AI Diagnostic Fund (AIDF) to accelerate AI adoption across NHS Trusts, particularly in imaging. The MHRA and NHS AI Lab jointly launched the AI Airlock, a regulatory sandbox allowing companies to trial AI health solutions in a supervised environment before wider NHS rollout.
Staff Attitudes
A survey published by the Health Foundation on 31 July 2024, covering 1,292 NHS staff and 7,200 members of the public, found that 76% of NHS staff supported the use of AI to help with patient care, and 81% specifically favoured AI for administrative tasks. This was a significant finding: the workforce wasn’t resistant, they were welcoming. The problem was access and implementation, not attitude.
But Adoption Was Lagging Behind Europe
A report produced by healthcare AI company Corti and YouGov, based on fieldwork conducted 26–28 November 2024, offered a sobering comparison. It found that 73% of UK healthcare professionals had never used AI at work, a higher non-adoption rate than any other nation in the study, which also included Denmark, France, and Germany. The biggest barrier cited by UK clinicians was fear of errors (62%), compared to just 44% in France.
Less than one in four UK healthcare professionals felt comfortable with their ability to use AI tools, well below the European average of 31%. UK burnout rates were also reported as among the highest in Europe at 64%.
UK healthcare was simultaneously one of the most interested in AI and one of the least equipped to actually use it.
2025: Scale, Evidence, and a System Under Scrutiny
2025 became the year the NHS moved, from pilots toward something approaching deployment.
The Numbers Jump Dramatically
According to SOTI’s report Healthcare’s Digital Dilemma, AI adoption in UK healthcare rose from 47% in 2024 to 94% in 2025. AI had moved beyond administrative tasks: 52% of healthcare professionals were using it for diagnosis and 57% to personalise treatments. These are striking numbers, though they reflect broad organisational adoption rather than uniform frontline use.
Virtual Wards and Remote Monitoring
AI-enabled remote monitoring was producing striking results in community care. A company providing virtual ward technology to NHS Trusts, reported that its AI-backed model which uses machine learning to identify patients at risk of deterioration via clinical-grade wearables resulted in a 61% reduction in bed days, an 89% reduction in GP appointments, and a 39% drop in non-elective admissions.
The Transcription Moment
One of the most visible changes in UK clinics in 2025 was the rapid spread of ambient voice technology (AVT) AI tools that transcribe and summarise GP appointments. The government made AI transcribing tools a central feature of the 10-Year NHS Health Plan, published in 2025, committing to rolling out “validated AI diagnostic tools” and deploying “AI administrative tools” NHS-wide within two years.
NHS England published guidance specifically for Chief Information Officers on how to implement ambient scribing products, emphasising that staff should be trained and that outputs must always be reviewed by the clinician.
2026: Public Attitudes and What's Coming
A Health Foundation survey conducted between July and October 2025 (published March 2026) offered the most recent snapshot of where public and staff opinion stands:
- 55% of the public believes technology improves care quality, though the number saying it makes care worse has risen from 8% to 13% since 2024.
- 38% of the public says AI will improve care quality (up from 33% in 2024), but 19% think it will make quality worse.
- Support for AI in administrative tasks (66%) outstrips support for AI in clinical decision-making (54%).
- Between 2024 and 2025, the modest increase in public support for AI in healthcare was largely confined to administrative applications.
The picture is one of cautious but growing acceptance, highest where the risks feel lowest.
