MHRA opens London AI sandbox to speed NHS tech adoption

TL;DR:

  • The MHRA has launched a regulatory sandbox to fast-track AI-enabled healthcare technology across London NHS organisations.
  • Up to 10 AI medical-device makers will deploy in live clinical settings under regulator oversight, running until December 2028.
  • The aim is to close the gap between regulatory approval and real-world adoption in the NHS.

The Medicines and Healthcare products Regulatory Agency has launched a regulatory sandbox designed to accelerate the adoption of AI-enabled healthcare technologies across London NHS organisations. Known as London Region I, the programme was developed with NHS England (London) and the capital’s three Health Innovation Networks. It creates a controlled environment in which AI-powered medical devices can be deployed in live clinical settings while generating real-world evidence on safety, effectiveness and patient benefit.

Regulation as enabler

The scheme will initially support up to 10 AI medical-device manufacturers working alongside NHS providers under MHRA oversight, and is expected to run until December 2028. It targets one of healthcare innovation’s most persistent problems: the gap between regulatory approval and actual deployment in services. “This initiative demonstrates that regulation can be an enabler for innovation, not a barrier,” said MHRA chief executive Lawrence Tallon, adding that regulators “need to work faster to keep pace with developments in AI”. Dame Caroline Clarke, NHS England’s regional director for London, said the goal was to help patients access new technologies “quickly, safely” while making the capital more attractive to the life-sciences sector.

The sandbox sits within a broader MHRA digital overhaul; the regulator recently awarded consultancy TXP a £1.5m contract for architecture, infrastructure and cybersecurity support. The MHRA will invite expressions of interest from NHS bodies and manufacturers next month.

Looking forward

The “evidence in live settings” model is the interesting part. Rather than approving devices in the abstract and hoping for uptake, the sandbox generates the procurement-grade evidence trusts actually need to buy and deploy — addressing adoption, not just authorisation. For UK health-tech firms, a structured route into live NHS environments could shorten commercialisation timelines that have long frustrated the sector. The risk is scale: a 10-company, London-only pilot is a test bed, not a national rollout. Whether the lessons translate into faster, safer adoption across the wider NHS will determine if this becomes a model or a one-off.