UK review urges provisional approval and monitoring for AI health tools

In short: A UK review says new AI tools used in healthcare should get provisional approval at first, with ongoing checks after they start being used.

What happened

A national UK review of how to regulate AI in healthcare has recommended a new approach that treats some AI tools like learner drivers. It suggests giving them provisional approval, similar to putting L-plates on a car, until companies can show the tools are safe and effective in real-world use.

The recommendations also call for continuous monitoring after approval. This is because some AI systems can change over time as they take in more data, which can be helpful but can also lead to “drift” (when a tool slowly starts making different decisions than expected).

The work came from the National Commission into the Regulation of AI in Healthcare, launched last year by the UK medicines regulator, the MHRA. The commission was led by two NHS doctors and drew on evidence from about 12,000 people, including patients, clinicians, health leaders, and technology developers. The MHRA said it will consider the recommendations and respond.

The commission also recommended a searchable public source of safety information about AI-enabled medical devices, including reports of harmful incidents. It said many people want to be told when AI was used in their care, although it noted that telling patients about every single use may not always be practical.

Why it matters

AI is moving into more parts of patient care, from helping read scans to supporting clinical decisions. Clear rules, live monitoring, and easy-to-find safety reports can help catch problems sooner and maintain trust, especially when tools can change after they are approved.

Source: Financial Times