The UK government said on 6 October 2026 that it had accepted all 44 recommendations from the independent National Commission on the Regulation of AI in Healthcare. The commission published its recommendations on 10 September, so the October announcement is the government’s implementation response, not a new clinical result or a launch of a new AI product. The government and the Medicines and Healthcare products Regulatory Agency (MHRA) want regulation to place greater weight on safety and performance throughout a device’s working life instead of relying mainly on a one-time assessment. That approach addresses the possibility that software systems can behave differently as they are updated and used beyond the conditions represented by evidence at initial approval.

Alongside the response, the MHRA opened applications for the third phase of AI Airlock, its regulatory sandbox for AI-enabled medical devices. The phase is intended to test with developers, regulators and health partners how devices can be monitored after deployment, how changes can be managed and how evidence from real-world use can be gathered. The government says the programme has secured three years of additional public funding. The MHRA committed to issue draft guidance by December 2026 on managing changes to AI-enabled devices as they adapt, and to begin a consultation in 2027 on qualifying and classifying such devices. Accepting all recommendations does not mean each measure is already in force or complete: the government plans to publish a full implementation roadmap in spring 2027.

The design also has economic implications. Clearer rules on evaluation and monitoring could help medical software developers anticipate the evidence they need and help hospitals assess safety responsibilities when buying products. Continuous monitoring can also add costs for data quality, clinical oversight, staff training and incident reporting. Airlock is a testing and policy-learning mechanism; participation is not general certification that a product is effective or safe. Each system still requires evidence on accuracy, performance across patient groups, human oversight and the effects of software changes.

The commission’s 44-point package and the government response cover assessment, market access and post-market oversight together. Near-term steps include phase three of AI Airlock, draft guidance on maintaining safety as devices change, and a consultation on device classification. The government said work had begun across ten priority areas involving government, the MHRA, NHS partners and the devolved administrations. A detailed roadmap with responsibilities and timelines is expected in spring 2027. The announcement therefore sets a regulatory direction, but it does not provide a final compliance timetable for companies or a measured patient-benefit outcome.