Story Highlight
- UK government accepts all 44 AI recommendations for the NHS.
- Full implementation plan to be published by spring 2027.
- Healthcare AIs to have staged authorisations for controlled testing.
- MHRA starts developing framework for AI-enabled medical devices next year.
- AI Airlock regulatory sandbox opens for new applications immediately.
Full Story
The UK government has announced its acceptance of all 44 recommendations from a recently released report regarding the use of artificial intelligence tools in the National Health Service. A full implementation plan is set to be published by spring 2027, with an initial draft due before the end of this year.
The report by the National Commission into the Regulation of AI in Healthcare advocates for staged authorisations for healthcare AI applications. This approach is likened to the learner driver ‘L plates’, permitting testing in controlled environments prior to broader NHS deployment. The findings highlight inadequacies in the current regulatory framework, which is primarily designed for traditional medical devices. The government stated that, by adopting these recommendations, the UK will establish “a world-leading agile approach to AI device regulation that supports innovation while maintaining patient safety and public trust.”
Further commitments include developing a method for managing ongoing changes to AI-enabled medical devices as they advance, aimed at ensuring safety while promoting timely access to new innovations. The report suggests exploring staged authorisation pathways, which would allow for earlier use of promising AI tools in the NHS, under close supervision while additional real-world evidence is collected.
Professor Alastair Denniston, who led the review, remarked, “This response shows that government has listened, not just to the Commission, but to the more than 12,000 patients, clinicians, and members of the public who shaped our recommendations. Our approach to AI should be people-centred, and be underpinned by a regulatory system that is safe, fast, and trusted.”
In conjunction with the government’s announcement, the Medicines and Healthcare products Regulatory Agency (MHRA) revealed it is now accepting applications for the third round of AI Airlock, its regulatory sandbox for AI-enabled medical devices. Following funding for an additional three years, this new phase will directly apply the recently recommended regulations. AI developers are expected to collaborate with regulators and healthcare partners to confirm how AI devices can be effectively monitored post-deployment.
Applications for participation in the third phase are open, with selections to be finalised by November. A webinar for prospective applicants is scheduled for 22nd October. Lawrence Tallon, chief executive of the MHRA, stated, “The MHRA’s AI Airlock is already opening its doors to more innovators who want to help us understand how these technologies can be implemented and monitored safely throughout their working life. This is about making sure patients can benefit from safe, trusted AI more quickly, while never losing sight of the safeguards that maintain public confidence.”
Source: read the original report.
What this means for your site
The introduction of AI in healthcare presents significant challenges regarding safety and regulation. The inadequacy of the existing regulatory framework signals a need for organisations to reassess their compliance with current legislation. The Health and Safety at Work etc. Act 1974 and the Management of Health and Safety at Work Regulations 1999 are particularly relevant, requiring a systematic approach to risk assessment and management of new technologies. Furthermore, ensuring the safety of AI medical devices under RIDDOR 2013 is crucial, as any incidents involving these devices must be reported to ensure accountability and learning.
To prevent the potential pitfalls illustrated by the need for new regulatory measures for AI in the NHS, organisations should implement rigorous internal monitoring frameworks that align with proposed guidelines and continuously assess the effectiveness of AI tools. This includes establishing robust processes to gather real-world evidence and ensure compliance with health and safety legislation. Conducting regular training sessions for staff on navigating these new technologies with a clear focus on safety protocols will further support the integration of AI solutions while safeguarding patient welfare.
