The National Commission into the Regulation of AI in Healthcare has published the findings of its Call for Evidence, revealing a striking divide between patients and industry over whether the current framework adequately protects health data but clear consensus that the rules on liability, transparency and post-market oversight need significant reform.
The summary of findings, produced by the Medicines and Healthcare products Regulatory Agency (MHRA), draws on 761 responses to the Call for Evidence, which ran from 19 December 2025 to 2 February 2026. Around two thirds of responses came from individuals and a third from organisations, including medConfidential, Privacy International, the Open Data Institute, Understanding Patient Data and the Ada Lovelace Institute.
The Commission, an independent advisory group announced in September 2025 to advise the MHRA on a new regulatory framework for AI in healthcare, is chaired by Professor Alastair Denniston with Patient Safety Commissioner for England Professor Henrietta Hughes as deputy chair. Its recommendations are due in the summer of 2026, supporting the Government's ambition for what its foreword describes as a framework that is "safe, fast and trusted".
Asked whether the current regulatory framework is sufficient on data governance and data privacy, 82% of patients and the public said it was not with 57% strongly disagreeing that it is adequate. Industry respondents took the opposite view: 48% considered the framework sufficient, against 37% who did not, making data governance the domain with the sharpest divergence between respondent groups.
In qualitative responses to the survey, patients and the public expressed strong concerns about consent arrangements for data access and the use of NHS data by commercial entities, including unease about private companies extracting long-term value from health data without fair public return. Industry respondents, by contrast, identified governance and compliance burdens and fragmented data infrastructure as barriers to development, calling for shared data governance templates and clearer guidance on data standards. Respondents also insisted that post-market surveillance arrangements must not give manufacturers inappropriate access to confidential patient data.
Across all groups, more than 80% of respondents wanted some level of change to the regulatory framework, with "significant reform" the majority position (54% to 59%) among healthcare professionals, providers and industry. Patients and the public went further, splitting almost evenly between significant reform (34%) and a complete overhaul (35%) - 77% of them rated the current framework as too loose for patient safety, while industry leaned the other way, with 65 per cent finding it somewhat or too restrictive of innovation.
On liability - the subject of the Medical Protection Society's intervention earlier this week calling for AI systems to be classified as products - the findings record between 77% and 88% respondents across groups saying the current legal framework is insufficient or contains gaps. Three quarters of patients and the public deemed existing laws unfit for AI altogether, while most industry respondents preferred the milder formulation that gaps create uncertainty in some scenarios. Respondents consistently called for clarity on how liability is allocated across manufacturers, deploying organisations and clinicians, with many favouring shared or distributed liability models.
Among the report's ten key findings are calls for continuous, lifecycle-based post-market surveillance capable of detecting performance drift; retained human oversight of clinical decisions; transparency and explainability requirements, including plain-English information for patients; and improved AI literacy and training.
Respondents also proposed a centralised national reporting system for AI incidents - modelled on the MHRA's Yellow Card scheme - noting limited awareness among healthcare professionals that the existing scheme already covers AI-enabled medical devices.
The MHRA stresses that the document is a neutral summary of respondents' views rather than a statement of policy. The findings will feed into the Commission's final recommendations, expected later this summer, which will inform the new regulatory framework for medical devices including AI committed to in the 10 Year Health Plan for England.
