Person-led, AI-enhanced: what the next phase of AI in healthcare means for patients, communicators and public affairs professionals

There is certainly no shortage of ambition for AI in healthcare. From supporting earlier diagnosis and more personalised care, to reducing administrative burden and giving clinicians more time with patients, its potential applications are significant. In fact, the Government has set an ambition for the NHS to become the most AI-enabled healthcare system in the world. But the latest report from the National Commission into the Regulation of AI in Healthcare provides a useful reminder that successful adoption will depend on more than the technology itself.

The Commission found strong support for AI in healthcare, but describes that support as “conditional, rather than automatic”. People want to see evidence of benefit, but also meaningful human oversight, accountability, fairness and transparency.

For those working in communications and public affairs, that is also worth paying attention to. As AI becomes increasingly embedded within healthcare, questions of public understanding, patient involvement and trust are likely to sit alongside questions of efficacy, regulation and implementation.

Person-led, AI-enhanced

One useful principle we have encountered through our work with organisations at the forefront of health innovation is a simple one: healthcare should remain person-led and AI-enhanced. This is also an interesting lens through which to read the Commission’s recommendations.

The report itself states that AI should augment rather than replace healthcare professionals. It envisages technology automating routine tasks and supporting care, while enabling healthcare professionals to focus more of their time on communication, compassion and shared decision-making.

And rightly so, that focus on people extends beyond clinicians.

The Commission recommends a system-level approach to informing patients about the use of AI in their care, including recognising a reasonable expectation that people know when AI is being used and, where possible or appropriate, have the ability and knowledge of that to opt out.

It also calls for patient and public perspectives to be incorporated continually into the regulatory system through mechanisms including workshops, public engagement sessions and a Patient Engagement Advisory Committee. Perhaps most interestingly, it proposes a periodic “AI sentiment census” to understand how patients and healthcare professionals feel about AI as its role in healthcare evolves.

Taken together, these recommendations suggest that engagement will not simply be something that happens once an AI technology is ready to deploy. Understanding the expectations and concerns of the people using and affected by it is likely to become part of how responsible adoption itself is approached.

But where are patients getting their information?

There is another side to the relationship between AI and patients that health organisations will increasingly need to consider: not only how AI is used within healthcare, but how people are using it to understand their own health.

In our conversations with a number of patient advocacy groups, this is already emerging as a concern. Organisations that have invested significant time and expertise in developing trusted, clinically reviewed patient information are seeing changes in how people access it. Rather than necessarily arriving at a charity or patient organisation’s website, people may increasingly ask an AI tool a question directly, receiving an answer without always knowing where that information has come from, how current it is, or whether important context has been lost along the way. That creates a different challenge. The question is no longer simply how trusted organisations communicate accurate information, but how that information remains visible, attributable and reliable in an AI-mediated information environment.

The Commission recommends clearer public-facing information about AI, alongside greater AI literacy and ongoing engagement with patients and professionals. But as the way people search for health information changes, there is likely to be a wider conversation about what good health information looks like in an AI-enabled world, and the role that regulators, technology companies, health systems and trusted patient organisations each play within it.

For communicators, this is an important shift to watch. Ensuring credible health information exists may no longer be enough; we will increasingly need to consider whether, where and how that information is being found and used.

From potential to proof

Another important feature of the Commission’s thinking is that AI-enabled products can evolve over time and perform differently depending on the population, setting, data and workflows around them. The Commission therefore argues for a more lifecycle-based approach to regulation, with greater emphasis on real-world evidence and ongoing monitoring, rather than relying too heavily on assessment at a single point in time.

There is a communications implication here too, with much of the public conversation around AI understandably focusing on its potential. As technologies move into real healthcare settings, there will be greater opportunity and arguably greater need to communicate what happens in practice.

Where is AI improving care? What are patients and healthcare professionals experiencing? Are the expected benefits being realised? What has been learnt through implementation?

For organisations working in this space, real-world evidence, clinician experience and patient perspectives can help move the conversation from what AI could achieve towards a clearer understanding of what it is achieving.

Who benefits will matter too

The Commission also gives considerable attention to health equity, describing it as both an ethical imperative and an important aspect of safety and performance. Its report notes the relationship between trust and issues including poor design, digital exclusion, bias and unequal access. If technologies underperform for particular groups, or people cannot access or understand them, confidence in AI-enabled care may be affected. This is likely to be an important consideration as AI adoption expands.

For policymakers and health systems, assessing an innovation may increasingly involve not only whether it works, but whether it works consistently across different populations and settings and whether its benefits are being distributed equitably.

For public affairs teams, these questions are therefore likely to become increasingly relevant to conversations about adoption, access and health inequalities.

The next phase is implementation

The Commission is clear that its recommendations are intended to enable innovation, not simply add additional controls around it. Its vision is for a framework that is safe, fast and trusted, with proportionate regulation, greater use of real-world evidence and more flexible approaches capable of responding to technologies that continue to evolve.

But it is equally clear that regulation cannot deliver successful adoption on its own.

The report describes AI as a system-wide responsibility, requiring healthcare organisations to have appropriate governance, workforce capability and AI literacy. It also proposes an “AI readiness” toolbox to help organisations assess whether they have the capabilities and controls needed to introduce particular technologies safely.

This perhaps points towards the next phase of the UK’s healthcare AI debate.

Alongside asking what AI can do, attention is likely to turn increasingly towards whether the system is ready to use it effectively: whether professionals are equipped to work with it, whether patients understand and trust its role, whether appropriate governance is in place and whether benefits are being realised across different communities.

For communications and public affairs teams, there is an important role to be taken within that process. Not simply communicating AI once decisions have been made, but supporting organisations to understand stakeholder expectations, involve patients and professionals appropriately, communicate evidence and uncertainty clearly, and demonstrate what innovation is delivering in practice.

Person-led, AI-enhanced is a relatively simple principle, but the Commission’s recommendations suggest it may be a useful lens through which to consider the next stage of AI adoption in healthcare: not technology for technology’s sake, but how technology can support a health system that works better for the people within it.