
MHRA expands its pioneering AI Airlock programme with £3.6m funding boost
pharmafile | April 16, 2026 | News story | Research and Development |
The Medicines and Healthcare products Regulatory Agency (MHRA) has secured new funding to expand its AI Airlock programme, aimed at supporting the safe introduction of AI-powered medical devices to patients.
Following a second phase, the Department of Health and Social Care will provide £1.2m per year from 2026 to 2029, enabling longer-term testing of technologies and development of regulatory pathways.
James Pound, MHRA Executive Director of Innovation and Compliance, said the programme is helping to bring new technologies to patients more quickly while maintaining safety.
He added: “This additional investment will allow us to scale up and ultimately strengthen our ability to ensure that AI‑powered medical devices can reach patients safely, efficiently and with the confidence of robust regulatory oversight.”
Billed as the MHRA’s first regulatory sandbox for AI as a Medical Device (AIaMD) products, AI Airlock is led by the MHRA, whose partners include the Department of Health and Social Care, the NHS AI Team and the UK Association of Medical Device Approved Bodies (Team-AB).
It forms part of wider efforts to adapt regulation for AI in healthcare and aligns with government plans to support the safe adoption of new technologies across the NHS.
Since launching its first phase in 2024, the programme has examined challenges specific to AI medical devices.
These include managing risks of errors and inaccuracies, the need for transparency in system recommendations to maintain clinician confidence and the importance of ongoing monitoring after deployment.
The second phase has explored areas such as AI-powered diagnostics, pre-determined change control plans (PCCPs) and how devices might expand in scope or intended use.
Published reporting of this phase is due in summer 2026 and, combined with the phase one pilot findings, will inform the third phase of the project and future regulatory approaches.






