UK regulation of AI in healthcare will move to lifecycle monitoring and system-wide responsibility

The MHRA has accepted all 44 recommendations from the National Commission into the Regulation of AI in Healthcare, confirming the future landscape for UK regulation of medical devices. The MHRA response sets 10 priority areas for delivering the shift to proportionate lifecycle regulation and system-wide responsibility. 

The National Commission into the Regulation of AI in Healthcare published recommendations in September 2026 which we wrote about here. The MHRA Response not only accepts all recommendations but begins to set out a practical roadmap for implementation with some initial timelines for delivery. 

For organisations developing, procuring or deploying AI-enabled healthcare technologies, the Response offers a clearer indication of the UK's future regulatory direction. While no immediate legal changes are introduced, the MHRA's acceptance of all recommendations suggests that the Commission's vision of a more proportionate, lifecycle-based and system-wide approach to AI regulation is likely to become the foundation of future reform. 

For manufacturers, healthcare providers and technology suppliers, now may be an appropriate time to assess governance frameworks, contracting arrangements and post-deployment monitoring processes against the future expectations emerging from both the Commission's recommendations and the MHRA’s Response. While the detail of many reforms remains to be worked through, the publication of the Response marks an important step from policy development towards implementation of a new framework for AI-enabled healthcare technologies in the UK.

The 10 priority areas for delivery

The MHRA has committed to focus on the following priority areas to begin delivering transformation over the next year:

Proportionate Lifecycle Regulation 

• Priority 1 – Modernise the regulatory framework for AI-enabled medical devices;

• Priority 2 - Delivering equitable access and outcomes for patients and members of the public, with a special focus on underserved populations;

• Priority 3 – Shift towards stronger post-market assurance;

• Priority 4 - Enable safe iteration and earlier patient access;

System-wide Responsibility and Safe Management

• Priority 5 - Creating the assurance conditions necessary for safe, real-world deployment, including risk controls;

• Priority 6 - Building system capability and readiness for widespread adoption of AI, including training of the workforce;

• Priority 7 - Strengthening accountability, including through a culture of safety and reporting;

Trust, Transparency and Predictability

• Priority 8 - Strengthening transparency and public confidence;

• Priority 9 - Improving regulatory clarity and navigability; and

• Priority 10 - Strengthening regulatory capability and the evidence base for AI in healthcare.

Implementation 

When the Commission's report was published last month, one of the key questions was whether the MHRA would endorse its recommendations and how quickly they might be implemented. The Response answers both of these questions.

The MHRA describes the National Commission’s report as a framework for the next phase of regulatory reform and links them directly to wider policy initiatives, including the NHS 10-Year Health Plan, the Life Sciences Sector Plan and the AI Opportunities Action Plan. It also expressly acknowledges that the current regulatory framework for medical devices was not designed with modern AI technologies in mind and is no longer fully fit for purpose. 

Additionally, and rather than treating AI regulation solely as a medical device issue, the Response adopts the Commission's broader systems-based approach. The MHRA emphasises that successful implementation will require coordinated action from regulators, healthcare providers, manufacturers, professional bodies and policymakers, with patient trust and public confidence remaining central objectives. 

Clear themes within the Response

Consistent with the Commission's recommendations, one of the clearest themes emerging from the Response is the move away from a regulatory model heavily focused on pre-market approval towards one based on ongoing assurance throughout an AI system's lifecycle.

The MHRA has identified strengthening post-market assurance as a priority. It intends to develop enhanced surveillance mechanisms, clearer evidence requirements and improved processes for monitoring safety, quality and performance in real-world settings. The aim is to create a more proportionate regulatory framework where evidence generation and oversight continue throughout deployment, rather than being concentrated at a single point prior to market entry. 

The Response also provides further detail on how adaptive AI systems may be regulated. In particular, the MHRA intends to publish draft guidance on Predetermined Change Control Plans (PCCPs) by December 2026, including proposals for adaptive AI-enabled devices and mechanisms allowing manufacturers to define boundaries for future changes while maintaining regulatory oversight.

Alongside PCCPs, the MHRA has committed to exploring staged authorisation pathways, expanded regulatory sandbox programmes and new approaches to enabling earlier patient access whilst continuing to generate real-world evidence. These proposals demonstrate a clear policy objective of supporting innovation without sacrificing ongoing oversight.

Greater clarity for AI developers & manufacturers

The Response contains a number of commitments that will be of particular interest to AI developers and manufacturers.

The MHRA intends to consult by Spring 2027 on a revised framework for qualification and classification of software and AI-enabled medical devices. Proposed reforms include:

• clearer rules on when software and AI products qualify as medical devices;

• potential carve-outs for certain software and AI-enabled products;

• an updated classification system better aligned with the risks and benefits of AI-enabled technologies; and

• a revised approach to assessing a product's intended purpose, taking account of functionality, design and context of use. 

The MHRA also plans to prepare further guidance covering areas such as direct-to-consumer AI devices, cybersecurity, usability and human factors, transparency measures, representative data and health inequalities, and the use of general-purpose AI models within regulated products. Most of these workstreams are expected to progress during 2027, with further details to be shared by Spring 2027. 

Notably, the Response recognises the growing importance of foundation and general-purpose AI models in healthcare. The MHRA has committed to exploring potential opt-in "master file" approaches and additional guidance on documenting model provenance, dependencies and associated risks within regulatory submissions. 

AI readiness, governance and accountability

The MHRA's Response reinforces the Commission's view that safe AI adoption cannot be achieved through product regulation alone.

A significant focus is placed on organisational readiness, workforce capability and governance. The Department of Health and Social Care and NHS England will work with devolved administrations to develop an AI Readiness Toolbox designed to help healthcare organisations assess and prepare for the safe adoption of AI-enabled technologies.

The Response also highlights the importance of AI literacy and staff training, with professional regulators, education bodies and healthcare organisations expected to work together to ensure healthcare professionals receive appropriate training and support. Alongside this, the MHRA plans to establish a working group involving DHSC, MHRA, CQC, NHS Resolution, professional regulators and other bodies to ensure that responsibilities are clearly understood across manufacturers, providers, healthcare professionals and regulators.

This aligns closely with the Commission's recommendation on contractual allocation of responsibilities and may increase scrutiny of contractual arrangements governing the deployment and operation of AI-enabled healthcare systems. Manufacturers and healthcare providers should therefore continue to monitor developments in this area and consider whether existing contracting models adequately allocate regulatory and operational responsibilities.

What happens next?

The MHRA will establish a cross-system Programme Board comprised of representatives from regulators, health bodies, research organisations and devolved administrations. The Board will meet quarterly to oversee delivery, manage risks and coordinate implementation across the four nations. 

Crucially, the MHRA has committed to publishing a detailed implementation plan and roadmap by Spring 2027, setting out roles, responsibilities, timelines and deliverables. Annual progress updates will follow thereafter.

This provides stakeholders with an early indication of when further detail can be expected and suggests that the next 6-12 months will be a particularly important period for the development of AI healthcare regulation in the UK.

If you would like to discuss further, please contact a member of our Health, Care and Life Sciences Team.

This article was written by Susannah Jury and Issy Peacock.