MHRA Secures £3.6 Million to Enhance AI Regulatory Sandbox

Key Takeaways

The MHRA has received £1.2 million annually for three years to expand its AI Airlock programme.
This funding aims to establish a sustainable regulatory framework for AI medical technologies.
Phase two of the programme will address regulatory challenges specific to AI diagnostics and is set for reporting in 2026.

Expansion of AI Airlock Programme

The Medicines and Healthcare products Regulatory Agency (MHRA) has secured an annual funding increase of £1.2 million over the next three years to enhance its AI Airlock programme. This initiative serves as a regulatory sandbox for AI applications in healthcare, aimed at fostering the responsible deployment of AI as a medical device.

The funding boost from the Department of Health and Social Care (DHSC) is a significant step, allowing the programme to transition from a yearly budget framework to a long-term approach. James Pound, executive director of innovation and compliance at the MHRA, emphasized the importance of this support, stating that it represents a crucial advancement for both AI Airlock and the safe integration of AI technologies in healthcare.

The AI Airlock programme operates in collaboration with DHSC, the NHS AI Team, and Team AB, a consortium of UK Approved Bodies. Its purpose is not just to facilitate the development of AI medical technologies but also to align with broader governmental priorities including the AI Opportunities Action Plan and the Life Sciences Sector Plan.

Dr Dom Pimenta, CEO of TORTUS, an AI medical scribe, highlighted the programme’s cross-educational benefits, saying it helps deepen understanding of regulatory expectations and fosters the exchange of industry insights at a time of rapid AI advancements. He considers the extension of AI Airlock an opportunity to create a global standard for the deployment of safe and effective clinical AI solutions.

The first phase of the AI Airlock pilot, conducted in 2024, revealed several regulatory challenges unique to AI medical devices. Key findings indicated that effective risk management must address specific AI-related issues, such as minimizing errors through techniques that ensure model outputs are supported by verified clinical data. Additionally, the pilot emphasized the necessity for AI systems to provide clearer explanations for their recommendations to bolster clinician confidence.

As the programme advances into its second phase, it will address specific regulatory challenges relating to AI diagnostic tools. This phase includes developing pre-determined change control plans and exploring how AI devices may evolve in functionality or application. Technologies examined in this phase include large language models, voice recognition tools, and specialized diagnostic solutions for conditions like cancer and rare diseases.

Insights from the AI Airlock programme are expected to contribute significantly to the ongoing efforts of the National AI Commission, particularly regarding future regulations for AI in healthcare. Coverage from phase two is anticipated to be published in Summer 2026, adding to the growing body of knowledge necessary to navigate the evolving landscape of AI in healthcare.

The content above is a summary. For more details, see the source article.