Is the UK’s life sciences regulator, the MHRA, creating the right conditions to encourage businesses investing in artificial intelligence products?

In April 2024, the Medicines and Healthcare products Regulatory Agency (MHRA) set out its strategic approach to artificial intelligence (AI). Six months on, the UK has elected a new government that has earmarked the life sciences industry, including AI in healthcare, as key to achieving growth in the economy. One of the first steps the new government took was to commission an Independent Investigation of the NHS, by Lord Darzi, which included AI within its review and was published in September 2024. On 8 October 2024, the UK government launched the new Regulatory Innovation Office (RIO), with a remit to work with the MHRA to help quickly and safely deploy innovative healthcare products and services, including AI.

This article assesses the MHRA’s approach to AI, taking into account commercial and public policy factors that are driving the need for regulatory reform. It also examines some of the key regulatory issues that businesses are likely to need to take into account when bringing AI products to the market in the UK.

AI in healthcare

The publication of the MHRA’s April 2024 strategy, and Lord Darzi’s review of the NHS, have come at a timely moment for businesses investing in AI in healthcare. Biopharmaceutical companies are using AI and machine learning to speed up bringing products to the market (such as Pfizer using supercomputing to develop Paxlovid). AI is being used for more efficient and impactful drug research and development (such as Sanofi using AI to predict clinical trial results in drug candidates). AI is increasingly being used in the areas of diagnosis and prevention, medical device design and imaging (the UK government has predicted that cancer waiting times will fall through the use of AI detection). Large language models (LLMs) have the potential to assist with content analysis, data creation, post-market surveillance and compliance.

Governments the world over are talking up AI as a key factor in confronting the difficulties faced by a confluence of issues: the rising cost of healthcare, aging populations and the challenges associated with ever more complex treatment of chronic conditions.

Accordingly, the MHRA issued its April 2024 strategy on AI at an exciting time for the life sciences industry, but at time when that excitement is tempered by regulatory uncertainty. AI is developing at pace; even AI experts are being taken by surprise at the speed of change. Regulations are lagging behind technological developments, meaning that manufacturers take on added risk when drawing up their business plans. How far has the MHRA gone to fill gaps in the regulatory regime?

The MHRA’s strategic approach

The MHRA’s strategic approach to AI was published in April 2024. The MHRA has outlined how it will implement the principles of safety, security and robustness; appropriate transparency and explainability; fairness; accountability and governance; and contestability and redress.

For manufacturers of AI as a medical device (AIaMD), the MHRA has set out the detail of the existing regulatory regime and a summary of the likely new regulations and guidance. Regarding medicines, the MHRA has focused on using AI for administrative purposes, to improve the quality of applications for licences, protecting consumers from fraudulent products and to enhance post-market vigilance processes.

New medical device regulations are due to be brought in during 2025. The MHRA states that it wants to achieve a balance to both support innovation and promote patient safety. For example, the regulations affecting many AI products are going to be tightened up. Some products that are currently in the lowest risk category, Class I, will be ‘up-classified’ to Class II. Meanwhile, the MHRA launched its AI Airlock in Spring 2024, intended to bring together expertise to find solutions for regulatory challenges specific to AIaMD.

During 2025, the MHRA intends to publish guidance on cyber security, as well as further guidance on applying human factors to medical devices. Over fairness and bias, a particular risk area for AIaMD, the MHRA is encouraging manufacturers to adopt the ISO/IEC TR 24027:2021 standard, which sets out guidance for assessing bias within AI programs, and is working with the Standing Together Project to establish standards for data inclusivity and generalisability in health datasets and AI technologies.

On 25 September 2024, the MHRA supplemented its April 2024 strategy by updating its plans regarding new regulations that will be brought in during 2025. The MHRA will also consult industry on policy towards the recognition of medical devices authorised by overseas regulators, taking into account that products approved in the EU are currently permitted in the market in Great Britain until 2028 (2030 for some products).

The UK’s new government and Lord Darzi’s report on the NHS

As part of the new UK government's ‘first 100 days’ in office, the Secretary of State for Health and Social Care, Wes Streeting, commissioned an independent review by Lord Darzi into the National Health Service (NHS) in England. Lord Darzi reported on 12 September 2024 that the NHS is in a ‘critical condition’ and cited a litany of issues that need to be addressed, including tackling the overall health of the nation, as well as waiting lists, cancer mortality rates, low productivity and a lack of capital investment.

Lord Darzi argues that there should be a ‘major tilt towards technology to unlock productivity’, noting that ‘there is enormous potential in AI to transform care and for life sciences breakthroughs to create new treatments’.

In addition, Lord Darzi refers to the opportunities presented by the NHS being unusual in holding deep and broad datasets on the population, combined with the fact that the UK is a leader in developing AI technologies. Lord Darzi argues that there is an opportunity for the NHS to be ‘at the forefront of this revolution’.

Industry leaders have called for Lord Darzi’s recommendations to be acted upon, in order to improve patient outcomes. In response to the report, the UK’s Prime Minister, Sir Keir Starmer, has pledged to implement a ten-year plan based around ‘moving from an analogue to a digital NHS’ that leverages data and technology. The launch of the new RIO is intended to enhance regulatory capabilities and reduce barriers to market.

The life sciences industry can be forgiven for thinking that Lord Darzi’s report is long on diagnosis of the NHS’s problems, but short on the specifics of how to cure its ailments. Neither the NHS nor the UK government have provided much detail, so far, on the changes needed to laws, regulations and guidance that are required to integrate AI into the healthcare system. However, taken together, the MHRA’s strategy on AI, Lord Darzi’s report, and the government’s response, amount to developments that companies and their advisors can use in their business planning.

Brexit and developments in the EU

The regulations concerning life sciences in the UK are derived from EU law, but are frozen in aspic as of the end of the Brexit transition period, which ran until 31 December 2020. The UK in effect ‘cut and paste’ the EU law in force at that time and nationalised it. In the meantime, the EU regulatory regime has continued to develop, particularly with the Medical Devices Regulation 2017, which took effect in May 2021, the Artificial Intelligence Act 2024, which entered into force in August 2024, and the Data Act, which has been in force since the beginning of 2024.

For a time, following Brexit, the indications from the MHRA and industry were that there would be an intention to largely follow the EU model when the UK updated its regulations. As it became more apparent that the UK’s Labour Party would form the next government, this became more of a solid prediction, given the Party’s position on seeking closer regulatory alignment with the EU. However, over recent months, criticism of the EU system has become more vocal, with some arguing that too much red tape is affecting productivity and competitiveness, including in AI and life sciences. These points were expressed in a report by Mario Draghi, commissioned by the EU and published in September 2024.

Based on recent developments: what can businesses developing AI expect from the MHRA?

This table sets out some key expected MHRA actions, how they link to policy objectives, and when industry can expect more concrete information.

Conclusion: is the UK life sciences regulatory system creating the conditions to encourage investment in AI?

Whilst the potential for creating life sciences AI products, including AIaMD, appears almost limitless, business benefits from regulatory certainty and clarity. The pace of developments in AI and healthcare means that regulatory certainty is hard to deliver. The key is for regulations to be flexible and kept under constant review.

The MHRA’s strategic approach to AI, taken alongside policy objectives by the new UK government and Lord Darzi’s report mark a significant point for life sciences in the UK. The MHRA’s emphasis on balancing innovation with patient safety, coupled with its commitment to updating and aligning regulations with international standards, provides a promising framework for industry. However, the regulatory landscape remains in flux. Manufacturers must stay vigilant to navigate the evolving requirements.

Life sciences manufacturers will expect to see the MHRA keep to the timetable it has set itself to consult with the industry, work with the government to introduce new regulations and issue guidance throughout 2024 and 2025. The MHRA’s business-focused initiatives, such as the AI Airlock, and forthcoming reform of the medical device regulations, are crucial in addressing the novel challenges posed by AI in healthcare. As the UK government continues to prioritise AI and technology in its economic growth strategy, the life sciences industry stands at the cusp of a transformative era. Businesses that can adeptly manage regulatory risks, respond to the MHRA’s initiatives, and leverage the opportunities presented by AI, will be well-positioned to lead in this dynamic and rapidly advancing field.