Skip to main content
Back to blog

Public and NHS Staff Attitudes to AI in Healthcare: Key Findings from The Health Foundation 2025 Survey

We attended a TechUK Briefing earlier this week where The Health Foundation presented some of their key findings from their landmark 2025 survey exploring public and NHS staff attitudes to technology and AI in health care. Based on responses from 8,000 members of the UK public and 2,000 NHS staff, this research offers the most comprehensive picture to date of how patients, citizens and the health workforce view the growing role of digital tools and artificial intelligence in the NHS, and where caution remains.

Public and NHS Staff Attitudes to AI in Healthcare: Key Findings from The Health Foundation 2025 Survey

The Desire for Digital: Technology Wins but AI Still Faces Skepticism

The public still sees technology as a route to better care. According to the report, 55% of the public say technology improves the quality of care. Only 13% think it makes care worse. Patients are definitely not nostalgic for clunky outdated systems with paper letters, phone queues and disconnected systems and they are ever more accepting of digital solutions. However, when it comes to AI only 38% of the public say AI will improve care quality, up from 33% in 2024. This small increase shows that the public is becoming more familiar with AI, and perhaps more curious, but still not entirely convinced.

This is entirely understandable. AI in health care is not the same as AI recommending a playlist on Spotify or CoPilot fixing the grammar in an email. The implications are of course vastly more important. It’s full of uncertainty, emotion, risk and consequence. A wrong answer can at best delay treatment and at worst cause significant harm. A poorly designed workflow can increase pressure on already overworked staff. A biased model can widen existing inequalities.

Administrative AI Leads the Way: High Support for Back-Office Efficiency

This is where the difference in use case I mentioned earlier comes into play. Around 66% of the public support AI being used for administrative purposes, while 24% are unsupportive. This relates to AI that helps with documentation, scheduling, triage support, coding, back-office processing and the repetitive work that drains staff time from care.

Unsurprisingly support is lower when AI moves closer to clinical care but it’s still higher than you might think. The survey found that 54% of the public support AI being used in patient care, while 33% are unsupportive.

Clinical Decisions: The Critical Need for a Human in the Loop

This number, however, rises sharply when people know there’s a human in the loop, a member of staff who will be checking its output. In the survey, respondents were asked how comfortable they would be with the NHS using AI in six clinical and non-clinical situations. One group was told that any AI-supported decisions would be reviewed by health care staff. The other was told that the AI would operate without routine staff checks.

The difference was striking. Across the use cases, people were much more likely to support AI when there was human supervision. For example, when asked about using AI to help choose a medicine or course of treatment, 50% of the public said they would be happy with this if the decision was checked by staff, while 35% said they would not. That leaves a net support margin of 15 percentage points.

In plain terms: people are not necessarily opposed to AI making recommendations. But they want reassurance that a human is still responsible for reviewing the decision before it affects their care.

NHS Staff: More Positive about AI, Driven by Operational Relief

The survey also looks at the views of NHS staff and they are consistently more positive than the public about technology and AI. The report found that 80% of staff surveyed supported the use of AI for patient care, compared with 54% of the public. This makes sense when you consider that staff bear the brunt of operational inefficiencies and they know first-hand how much valuable time is lost to administration. They deal on a daily basis with the burden of documentation and duplicated effort caused by fragmented systems and data lost in different silos. With the right digital tools they can use their time and expertise where it really matters the most.

The Path Forward: Safe, Supervised AI Grounded in Human Judgement

To conclude, The Health Foundation’s report doesn’t suggest that people are opposed to AI in healthcare, it shows that people are open to it when it feels safe, supervised and grounded in human judgement.

The survey findings should shape not only how we talk about AI, but more importantly, how we build and implement it. The strongest case for AI in healthcare is not that it can replace people but that it can give people back the time, clarity and capacity they need to deliver better care. It can reduce administrative burden, surface useful information, support decision-making and make services easier to navigate.

But it works best when it sits around clinical judgement, not above it and this aligns closely with how we see AI at Blüm Health. As we explored in our recent article AI in Healthcare: The Most Valuable Innovation Still Starts With People, healthcare remains a human-first industry. So the opportunity ahead as we see it is not simply to deploy more AI but to deploy the right AI, in the right places, with the right safeguards, and with clinicians and patients involved from the start.

You can read the full Attitudes to technology and AI in Healthcare Survey by the Health Foundation here.