UK Government Accepts 44 Recommendations to Monitor Medical AI Throughout Its Use

Medical AI should face scrutiny throughout its working life, not mainly at a single approval checkpoint. On October 6, 2026, the UK government accepted all 44 recommendations from the National Commission into the Regulation of AI in Healthcare, backing a framework intended to maintain patient safety as devices change and improve.

The commission, whose recommendations were published on September 10, called for a proportionate, lifecycle-based framework: oversight across a device’s working life. Its central concern was that relying too heavily on one-time assessments can leave regulators without a full understanding of performance in use. Monitoring after deployment is meant to provide that real-world evidence.

The response also commits to exploring staged authorisation. That would allow promising AI tools into the NHS earlier, under close supervision, while further evidence is gathered. It is an approach the government will explore, not a newly available approval route. The stated goal is faster access without abandoning safeguards as technologies evolve.

But innovation must never come at the expense of patient safety.

James Frith, UK Health Innovation Minister

Alongside the decision, the Medicines and Healthcare products Regulatory Agency, or MHRA, opened applications for phase three of AI Airlock. This regulatory sandbox brings developers, regulators and healthcare partners together to test regulatory questions using real products and technologies. The new phase focuses on post-market surveillance—monitoring devices after deployment—and oversight throughout their use.

Its testing themes come directly from the commission’s recommendations. Participants will examine how devices can be monitored and managed safely, generating evidence for future MHRA guidance and policy. AI Airlock has secured three additional years of government funding following its pilot and second phase. Eligible medical-device developers can apply now.

Acceptance of the blueprint starts an implementation process. The government has set out several near-term milestones for developers and the health system:

• October 22, 2026: a webinar for prospective AI Airlock applicants. Applications close in November, when the first wave of innovators will also be selected.
• By December 2026: draft MHRA guidance on managing changes to AI-enabled medical devices as they adapt and improve.
• By spring 2027: a full implementation roadmap assigning timelines and responsibilities across all 44 recommendations, followed by regular progress updates.

The MHRA will also begin 2027 by consulting on how to qualify and classify AI-enabled devices. A newly established Programme Board will coordinate work across government, the regulator, NHS partners and devolved authorities on nine priority areas.

The commitments extend beyond device performance. The government pledged to strengthen patient involvement in decisions, keep patients informed when AI is used in their care, and improve access to redress when care falls below expectations. It also committed to working with health-system partners so benefits reach underserved communities.