The UK is moving towards continuous monitoring of AI medical devices, replacing reliance on one-time safety assessments with lifecycle oversight.
LONDON: The UK government is moving towards continuous safety monitoring of AI medical devices throughout their working lives, marking a significant change in how artificial intelligence used in healthcare could be regulated.
The approach is intended to ensure that systems remain safe and effective even as their performance, software and real-world use evolve after approval.
UK accepts new AI healthcare recommendations
The government confirmed on October 6, 2026, that it had accepted all 44 recommendations from the National Commission into the Regulation of AI in Healthcare.
A central recommendation calls for regulators to move away from relying mainly on a single assessment before deployment and towards monitoring products throughout their lifecycle.
The Medicines and Healthcare products Regulatory Agency, or MHRA, is now opening the third phase of its AI Airlock programme.
The regulatory sandbox will bring developers, healthcare organisations and regulators together to test how AI-enabled healthcare tools can be monitored after they enter real-world use.
Why continuous monitoring matters
The latest phase will focus specifically on post-market surveillance and lifecycle regulation.
Evidence gathered through the programme is expected to shape future MHRA guidance and policies covering AI medical devices as they are deployed more widely across healthcare services.
One challenge for regulators is that some artificial intelligence systems can change or improve after deployment.
Traditional medical-device regulation was largely designed around products whose characteristics remain relatively stable. AI systems can behave differently as software is updated, data changes or systems encounter different patient populations.
Patients could get clearer AI information
The government said patients should also receive clearer information about when artificial intelligence is being used in their care.
Its plans include stronger safeguards, greater transparency and clearer routes for raising concerns if problems emerge.
Regulators will also pay greater attention to whether systems perform safely across different groups rather than simply measuring average performance.
New guidance expected from MHRA
The MHRA plans to issue draft guidance by December 2026 covering how changes to adaptive AI systems should be managed.
Further consultation on how AI-enabled devices should be classified is expected to begin in 2027, while a wider implementation roadmap covering all 44 recommendations is due by spring 2027.
The government is also exploring staged authorisation, which could allow promising technologies into NHS settings earlier under closer supervision while developers gather additional real-world evidence.
Impact to expect
Patients could gain earlier access to innovative healthcare technology while receiving stronger protection after deployment.
For developers, continuous oversight of AI medical devices could also mean clearer expectations around safety, updates and long-term performance, potentially helping the UK build greater trust in AI-powered healthcare.
