Is the NHS ready for the AI-powered patient?

How a consumer health revolution is reshaping the system

The NHS is under pressure. A mismatch in care needs and care capacity has created long waits, undetected care needs and overburdened staff who are pressured to make quicker decisions.

At the same time, there is a growing wave of direct-to-consumer health innovation. Weight-loss drugs, AI health companions, home diagnostics and private care options are increasingly available. These developments are reshaping people’s relationship with their own health and care experiences and, by extension, how they relate to the NHS.

The NHS can choose to ignore, react or shape the direct-to-consumer innovation boom. The first two options leave the NHS with little control and patients risk being frustrated with poor care experience and outcomes. But if the NHS were to play a more active role in shaping the wider health care system, it could unlock more responsive person-centred care and help address the gap between demand and capacity.

In February 2026, The King’s Fund convened a roundtable with NHS leaders, technology developers, patient representatives, clinicians and policymakers to discuss what direct-to-consumer care means for the NHS. This long read draws on that discussion to suggest what is needed to navigate this rapidly evolving environment.

Three forces reshaping health behaviour and systems

Direct-to-consumer innovation

In 2025, weight-loss drugs (GLP-1 agnostics) hit mainstream public consciousness and became available through both NHS and private providers. But 9 in 10 people using GLP-1s are buying these drugs privately, largely because of ease of access: approximately 1.5 million people going through private routes compared to 200,000 through the NHS. This shows how speed and convenience matter to the public.

The advent of large language models (LLMs), such as ChatGPT and Copilot, is having an impact on how the public seek and receive health information. During the roundtable one participant highlighted how some consumer health AI tools are responding to 10 million health related queries per day, compared to the NHS providing 1 million interactions in 36 hours.

Research indicates that health information seeking increases at times when health care services are closed. The most frequent types of health queries are seeking advice and information on symptoms, specific medical conditions, medical jargon and navigating health care services. And the use of AI for health information could increase sharply in 2026 as leading generative AI developers have announced dedicated health modes, making it easier for the public to upload data from health care interactions and personal devices like wearable technology.

Generational change

Direct-to-consumer innovations are arriving at a time when increasing life expectancy means a greater range of different generations are drawing on health care services simultaneously. Digitalisation has transformed all sectors to such an extent that younger generations are accustomed to the majority of interactions, from banks and travel to retail and fitness, that are digital first if not digital only. These interactions are rapidly becoming personalised, data driven and increasingly mediated by AI. For these younger and digitally confident generations NHS exceptionalism, the idea that health services can operate by different rules to all other interactions, is unlikely to be accepted or understood. As more interactions with health services start digitally, the question becomes: how can the NHS adapt to this reality to stay relevant, responsive and accessible to large segments of the population?

A greater role for employers

The Keep Britain Working review aims to reduce economic inactivity occurring from ill health or disability. The review proposes that health at work is a shared responsibility between employers, employees and health services (which are deliberately defined as broader than the NHS, encompassing occupational health partnerships and other providers); whereas prior to the report health at work was conceptualised as the responsibility of the individual and the NHS alone.

The review suggests that a multi-provider marketplace of services should be cultivated for employers to select services to better support employees with emerging health needs. This could incorporate many of the direct-to-consumer innovations already becoming available for early intervention, early treatment and a focus on common conditions such as mental health and musculoskeletal conditions.

The NHS is not standing still

The NHS is not standing still in the face of these forces. The NHS App is the most visible example of a response to changing public behaviours and expectations. The 10 Year Health Plan accelerates this by cementing the NHS App as the front door of the NHS, intended to reduce the fragmentation of health care apps that is a source of public frustration. There is also an aspiration for the NHS App to incorporate AI triage, AI-mediated health information and navigating services with signposting tailored to the user.

But the reality today is beneath the surface, the NHS remains a cluster of silos, both digital and physical. Patients can experience fragmented care with repeated tests across NHS organisations, often for unclear reasons. This fragmentation is not limited to the NHS–private divide; it happens routinely when moving between NHS providers.

The NHS and government must take a more active role of shaping the holistic care experience across the workplace, direct-to-consumer support, private services and the NHS. A response that will require changes to NHS services, technology, workforce culture and skills.

The NHS is part of a multi-service health system

Meeting the consumer revolution including AI means the NHS taking three steps.

First, there needs to be pragmatism to meet people where they are. This means acknowledging that a multi-service system is emerging where patients can engage partially or wholly through different routes: drawing on care in the NHS, using private and direct-to-consumer services (whether through workplace health provision or self-pay), or drawing on AI tools to self-manage. In such a model, people may flow in and out of interactions with the NHS carrying a partially complete collection of tests and information from various sources, including consumer technologies like wearables.

Second, the NHS will need to be able to critically appraise health information from a range of sources and then build on this to continue the patient’s health journey. Staff will need to become familiar with shared decision-making conversations that use the health information patients bring to the consultation. One roundtable participant illustrated what this could mean in practice. They shared their experience of having a close friend who is a dermatologist and realising that much of their communication over years had involved asking medical advice about their children’s rashes and skin concerns. They observed that 99 per cent of people do not have access to that kind of expertise. An AI tool could make this kind of easy, informal clinical interaction more available to the public, making the care experience dramatically different.

Third, more work is needed to understand and actively direct how consumer-led AI innovations will complement, replace or add to the activity the NHS already does. During the roundtable NHS leaders voiced concerns that easier access to information might drive unnecessary consultations and widen inequalities. However, providers deploying digital triage tools in other health care markets shared their view that while this has been a big concern, it hasn’t materialised. Instead, when tools are designed intentionally, they can shift demand from being focused on sickness to prevention and engage populations that may have previously been excluded due to access barriers or cost.

How can the NHS meet the consumer health revolution?

Deep system change across digital and physical services

Too often digital and physical services are considered separately. A strong message from the roundtable was that large-scale digital change means physical services need to adapt too. The NHS App, for example, sits on top of existing infrastructure and processes. When these are not well aligned, patients in different locations have dramatically different experiences of the same app. Misaligned digital and physical processes increases the likelihood that things go wrong, for example, the accidental release of patient test results. Participants shared how large-scale digital change in other countries has required subsequent redesign of service configurations.

Digital change without service redesign results in a fractured patient experience and safety risks. When digital and physical services are not complementary it can add pressures through increased workload when patients move across settings and increased safety risk through insufficient data. The necessary deep transformation is unlikely to be a sequential process but an iterative and adaptive one, cycling between digital innovation and local service redesign, so that the AI revolution leads to genuinely better services, rather than simply digitising existing physical services.

Support patient voice and empowerment

At the roundtable, many participants shared stories of poor health care experiences, spanning clinical errors and missed diagnoses to being dismissed by clinicians. While AI is not perfect, we need to recognise harms are already occurring in the existing system, and we should be aiming for better, not necessarily perfect.

A multi-service system would mean a member of the public could choose to use direct-to-consumer AI tools to improve their health literacy and awareness. For example, receiving plain-language explanations of test results from an AI tool to support self-care and informed decision-making.

This could potentially mean the public would be better able to advocate for themselves: describing symptoms, articulating treatment preferences and navigating the system more effectively. The ease of access and flexibility of consumer AI could mean people with low health literacy and limited ability to convey their health needs stand to benefit most.

However, being able to advocate for yourself requires the right tools and support. Consumer AI tools are routinely criticised for being sycophantic, having a tendency to agree with the user rather than offering a critical perspective.

We heard a real example of what this means for patients. One participant shared the story of an elderly relative who had a suspected heart attack. He had conferred with an LLM and was more confidently able to explain and self-advocate when he arrived at A&E. But the tool did not reinforce the need to stay, so he self-discharged because of the AI’s tendency to agree with the user. Evidence is starting to show that the sycophantic nature of conversational AI can distort judgement and make it difficult for opinions to be changed. In health care, there is a risk that people using these tools could be exposed to poor decision-making and interactions that create, or entrench, incorrect suppositions that clinicians would then need to unpick.

Focus on system culture and workforce

The proliferation and use of consumer AI tools is likely to change the culture of the NHS workforce as well as their knowledge and skills. When medical knowledge is easily accessible and understandable by members of the public through consumer AI, it changes what patients can do. This combined with access to direct-to-consumer services and private and workplace health provision, it will be harder for the NHS to default to only accepting NHS-originated tests. Staff will need to access and adapt care plans with the patient based on their various interactions within and beyond the NHS.

This requires a shift away from paternalism and exceptionalism towards recognising other sources of medical information within a culture of meeting people where they are. Patients may arrive with significant amounts of information, some correct and some with high uncertainty or outright incorrect. It will require different skills, education and training with changes to workforce roles, responsibilities and professions to navigate good care provision in a different landscape of health information enabled through AI.

Overcome data barriers

Any multi-service model will need data flows underpinning its capabilities. Unfortunately, the NHS still has challenges with the basics of providing patients access to their own data. And data within the NHS is siloed and fragmented, which makes adding data flows to and from other service providers seem a far more difficult task.

In the US, a radically different approach is emerging: patients have the final say if they consent, their medical information must be shared regardless of health provider concerns or whether the recipient is a consumer technology company or a health care provider. Such a radical approach could be seen as the solution to some of the NHS data woes, but there are understandably a range of levels of trust in data-sharing amongst the public. In particular, there are legitimate concerns about data sharing and data use among some groups.

Addressing these concerns requires delivering on previous engagement with the public while building on this with additional engagement and co-production on the concept of multi-service care.

The unique opportunity the UK and the NHS have is the single patient record. In this rapidly changing world data flows must be accelerated. The opportunity is for the single patient record to become a truly holistic record of all care interactions, to meet the needs of a person who draws on care through multiple services. The NHS Modernisation Bill aims to facilitate the creation of a single patient record by mandating data flows across all health and social care providers and IT system providers. However, data sharing barriers are a complex mixture of technical and societal trust challenges, which the bill doesn’t address. For example, interoperability and data standards are foundational for a health care system that can coexist with workplace, private and consumer provision. And yet the system has been unable to create standards and harmonise, understand where data is not being shared and use financial penalties effectively. Furthermore, the patient control of data remains immature.

Consider risk and accountability

From the roundtable discussion, it was clear that the changing landscape would also create changes to where risk and accountability sit. These could be significant barriers to creating a system that enables multiple care services to wrap around an individual. Clinicians can be reluctant to divest themselves of risk: they feel professionally responsible for patient safety, and GPs as independent contractors face additional concerns about personal liability for harm caused.

But assessing risk is harder when the pace of evidence generation is far slower than the pace of technology development. And it can be a challenge to ascertain how effective generative AI is at certain tasks – a 2026 publication found that current AI models get over half of triage recommendations wrong for emergency conditions, for example.

Some generative AI companies are developing their own safety benchmarks, such as HealthBench. But to build trust we need to ensure evidence generation and safety assessment are not solely led by suppliers. Government has an essential role in bringing together academics, health systems and regulators to provide firmer views on consumer AI regulation. A clear approach to regulation is essential to give confidence to staff, patients and industry that there is an approach to ascertaining and managing the acceptable risk.

At the roundtable participants discussed how a different culture and approach to informed risk-taking using shared decision-making could help to shift how risk and accountability are navigated in a multi-service care model. This would require a shift away from the current paternal risk decisions approach by health providers where all the risk sits with clinicians and providers. Instead, we should explore how models of shared risk between patient and clinician could support different care options from AI enabled self-care to clinical mediated care. In this approach clinicians would support patients to take shared responsibility, by being articulating risk based on robust evidence and benchmarks. In some instances patient may not feel confident in making these decisions, and so should default to the current approach.

Ensure equity and build trust

At the roundtable there was no consensus on if direct-to-consumer AI innovations will widen or narrow inequalities.

AI offers the opportunity for people to be more confident in advocating for themselves, accessing information and being more informed. Technology can address gaps in health care provision by enabling access for sensitive health needs (eg, sexual health), providing access without travel and enabling care outside core hours.

But there are still significant numbers of people digitally excluded or at risk of digital exclusion. And these are not static figures: in times of increasing costs, people can be forced to choose between essentials like food or digital devices and connectivity.

Conclusion

The NHS is facing a consumer AI revolution and has at least three options on how it could respond.

First, it could ignore what’s happening in consumer tech and maintain a walled garden. It’s likely this will create more disgruntled patients and generational divides as digital approaches become more pervasive while the NHS remains an outlier.

Second, it could react to consumer health tech innovation as it happens. But this leaves the NHS always playing catch-up as the world changes around it. Potentially dealing with increasing care demand generation from other suppliers and consumer AI that is insufficiently regulated.

A third approach makes the most sense to me. The NHS could be a leading voice at the table, bringing different parts of the new health system together. The NHS does not need to do everything itself. But it does need to work out how to exist alongside, integrate with and learn from the consumer health revolution that is already under way.