Regulation is usually what slows innovation down. This programme is explicitly designed to do the opposite. The UK’s Medicines and Healthcare products Regulatory Agency has announced an AI sandbox that will place AI-enabled medical devices directly into working NHS settings in London, with regulatory oversight present throughout, not applied after the fact. It’s a meaningful shift in how the system thinks about getting new technology to patients.
The programme is a three-way partnership between the MHRA, NHS England’s London region, and the three London Health Innovation Networks: Imperial College Health Partners, UCLPartners, and HIN South London. Up to ten AI medical device manufacturers will be selected for the first phase, working alongside NHS providers across London to deploy their technologies in live clinical environments under active MHRA supervision.
What the sandbox actually does
The core idea is to generate real-world evidence on safety and effectiveness while giving manufacturers a clearer, more predictable path toward wider NHS adoption. Right now, AI medical devices often face a long, opaque journey from regulatory approval to actual clinical use. The sandbox compresses that process by putting devices into practice earlier, under controlled conditions, and collecting the evidence that would otherwise take years to accumulate.
The programme also aims to match technologies with genuine health service needs. That matters. One persistent failure mode in health tech is products built without sufficient clinical input, which then struggle to find a foothold in practice. Bringing developers and NHS providers together from the start addresses that directly.
Why London, and why now
London is a logical location for a pilot of this kind. It has a concentrated cluster of NHS trusts, academic health science centres, and life sciences companies, and the Health Innovation Networks already exist as connective tissue between them. Dame Caroline Clarke, director of NHS London, framed the ambition clearly: faster, safer adoption of technology that improves care, while also strengthening London’s appeal to the life sciences sector.
The timing also connects to the NHS 10 Year Health Plan, which has placed technology adoption and reducing health inequalities near the top of its agenda. A programme that could widen access to AI-assisted diagnostics or monitoring tools, particularly in underserved areas, fits that framing well.
The regulatory signal this sends
Lawrence Tallon, chief executive of the MHRA, said the initiative shows that regulation can enable innovation rather than block it. That framing is worth taking seriously. For too long, regulatory caution and commercial speed have operated on different timelines, with patients waiting in the gap between them. A sandbox model, when well designed, closes that gap without abandoning the safety standards that justify public trust in medical devices.
- Up to ten AI medical device manufacturers will be selected for phase one
- Devices will be deployed in live NHS clinical settings across London
- MHRA oversight will be active throughout deployment, not retrospective
- Real-world safety and effectiveness evidence will be collected systematically
- The programme is designed to support a faster route to wider NHS adoption
The sandbox does not remove regulatory scrutiny. It repositions it, closer to practice, earlier in the process. If the first phase produces solid evidence and a workable model, the question will be whether it can be replicated beyond London. That would be the real test of whether this approach changes anything at scale.