AI Airlock is a regulatory "sandbox", a type of study where manufacturers can explore how best to collect evidence that could later be used to support the approval of their product. This is done under MHRA supervision in a virtual or simulated setting. Doing so will help the manufacturer and the MHRA better understand the challenges of regulating AI in medical devices, leading to a more bespoke and enabling regulatory framework, a clearer route to market for industry and, most importantly, paving the way for quicker NHS and patient access to potentially transformative AI technologies.

The five selected technologies are:

Using AI to target at risk patients with COPD
Lenus Stratify is a medical device, developed by Lenus Health, that uses AI to analyse health data and predict serious outcomes from COPD, such as the risk of hospital admission. These predictions could allow multi-disciplinary care teams to intervene earlier, adjust treatment plans and significantly reduce the multiple hospital admissions that are unpleasant for patients and costly for the NHS.

Using Large Language Models to improve the efficiency and accuracy of radiology reporting
Philips aims to improve the workflow for radiologists by integrating AI into its existing systems. Typically, when radiologists review patient results, they write a summary called the 'Impression', which is a section of the radiology report that includes only information that the radiologist deems most important for the referring physician. By automatically summarising this section using AI, Philips aims to make radiology reporting more efficient and accurate: reducing administrative loads, errors, omissions, and miscommunications, to ultimately benefit patients and public health.

Using AI performance monitoring platforms in hospitals
AI learns by analysing large amounts of data. However, real life continually changes, and no dataset can capture every possible situation. This means that over time the AI's performance may decline. It may decline because new types of patients are seen, or new medical scanners are used, or something else in the environment has changed. This is called drift and presents a significant barrier to AI safety, and therefore its uptake. Federated AI Monitoring Service is part of an AI platform developed by Newton's Tree, that helps hospitals, AI developers and regulators monitor AI performance in real time. This proactive approach identifies and allows issues like drift to be resolved early, preventing potential risks and ensuring AI applications remain reliable. 

Using AI to improve the efficiency of cancer care
OncoFlow uses AI to help healthcare professionals involved in cancer care create personalised management plans for cancer patients. This has the potential to reduce waiting times for cancer appointments, leading to earlier treatment which in turn significantly increases the chances of survival. Initially, OncoFlow will focus on breast cancer patients due to the high number of cases and waiting times. However, the platform can be adapted for other types of cancer in the future.

Using AI to facilitate clinician decision-making
Large language models (LLMs) are a type of AI designed to generate normal language. However, it is often unclear what data was used to teach the AI and what information the AI used in its answer. This can lead to biased or inaccurate information being produced. SmartGuideline is an AI-powered medical device that allows clinicians to smart-search national guidelines with normal questions. It does this using a verified knowledge base (NICE guidelines) with a specially trained LLM. This helps doctors give patients the safest and most reliable treatments by using the most accurate and up-to-date information.

This collaborative project is led by the MHRA, in partnership with the NHS AI Lab and Team AB, the consortium of UK Approved Bodies. Also involved are subject matter experts across the healthcare sector, government and academia, the Information Commissioner's Office and other regulators.

Laura Squire, MedTech regulatory reform lead and chief officer at the MHRA, said: ‘We need to be confident that AI-powered medical devices introduced into the NHS are safe, and stay safe and perform as intended through their lifetime of use.

‘By examining the technologies announced today in a safe setting, in partnership with technology specialists, developers and the NHS, we can test and improve the rules for AI-powered medical devices, helping get products like these to the hospitals and patients who need them sooner.'

Karin Smyth, minister of State for Health (Secondary Care), added: 'AI has the power to revolutionise care by supporting doctors to diagnose diseases, automating time-consuming admin tasks and reducing hospital admissions by predicting future ill health allowing targeted, preventative action.'

Being selected for AI Airlock does not constitute a regulatory approval. The findings from the pilot, due to be announced in 2025, will inform future AI Airlock projects and influence future UK AI Medical Device guidance.