The MHRA has published the findings of its most extensive public engagement programme on artificial intelligence in healthcare, drawing on contributions from patients, clinicians, industry and civil society across the UK.

UK Regulator Sets Terms For AI Health Governance With Most Comprehensive Evidence Base Yet

The question of how to regulate artificial intelligence in healthcare has attracted strong opinions across every part of the UK health and social care system. What has been harder to establish, until now, is what the public actually thinks. 

The MHRA has moved to close that gap. Published today, two major reports set out one of the most extensive evidence bases ever assembled on AI perception, use and governance in UK health and their findings will directly shape the recommendations of the independent National Commission into the Regulation of AI in Healthcare, expected later this summer.

What The Evidence Shows

There is broad recognition of AI’s potential to improve care but that recognition comes with conditions attached. Participants across the public engagement programme expressed consistent support for ongoing post-market surveillance of AI tools, ensuring that performance, safety and efficacy continue to be monitored after deployment rather than assessed once at the point of approval. 

That distinction matters enormously in a clinical environment where AI systems can drift, degrade, or behave differently across patient populations than they did in controlled trials.

Equally consistent was the demand for human oversight. Respondents did not want AI operating autonomously in high-stakes clinical decisions. They wanted transparency about when and how AI was being used, accountability when things went wrong, and assurance that safety checks were rigorous rather than perfunctory. These are not unreasonable expectations. They are, in fact, the minimum standards any credible regulatory framework should already be delivering.

The Regulatory Gap The Commission Must Address

The more consequential finding for policymakers is the broad consensus that existing regulatory approaches are no longer fit for purpose. Current UK frameworks for medical device regulation were designed for a world of static technologies. A diagnostic algorithm that learns and updates is a fundamentally different object than a pulse oximeter, and the MHRA’s existing conformity assessment pathways struggle to account for that difference.

That pressure has been building for years. The MHRA’s Software and AI as a Medical Device Change Programme, launched in 2022, acknowledged the need for reform. The National Commission represents the most substantive attempt yet to translate that acknowledgement into a concrete regulatory architecture. The evidence published today will inform what that architecture looks like and the weight of public expectation sitting behind it means the Commission cannot afford to produce something vague.

Who Was Consulted And Who Was Not Forgotten

The engagement programme was notably broad in scope. The MHRA worked with National Voices, a coalition of health and social care charities, specifically to reach groups at risk of health inequalities or with complex needs communities that are frequently underrepresented in regulatory consultations but disproportionately affected by the deployment decisions that follow. The Health Foundation delivered a series of public deliberative sessions, and a public Ask Me Anything session gave patients direct access to Commission leadership.

The Call for Evidence drew 760 individual and institutional responses. For health and social care technology companies, that level of engagement is itself a signal: the days of deploying AI tools without meaningful public accountability are numbered.

What This Means For Social Care

The reports focus primarily on healthcare regulation, but the implications for social care technology are direct. AI-enabled tools are increasingly embedded in domiciliary care from predictive risk monitoring platforms to automated care scheduling and electronic care planning systems with AI-assisted features. Many operate in a regulatory grey area: not clearly medical devices, but making consequential decisions about vulnerable people’s lives.

The Commission’s recommendations are likely to inform how analogous frameworks develop in social care. For registered providers, integrated care boards commissioning digital services, and technology companies selling into the sector, the summer report will be essential reading.

The Summer Report And What Comes Next

The Commission’s recommendations, expected before the end of the summer, will feed into the MHRA’s ongoing regulatory reform work and inform wider government thinking on AI governance. The UK is operating in a competitive international environment: the EU AI Act is already in force, and the United States is developing its own federal AI regulation agenda. How the MHRA balances proportionality with rigour will determine whether the UK can position itself as a trusted destination for AI health innovation.

The public has spoken with unusual clarity. It wants AI in healthcare to be beneficial, safe and accountable. Whether the Commission translates that clarity into enforceable regulation or produces another framework that sounds robust until it meets reality is the question the sector should be holding in mind this summer.