UK Medicines and Healthcare products Regulatory Agency outlines AI strategy

As requested in the government's response to its white paper, "A pro-innovation approach to AI regulation", the Medicines and Healthcare products Regulatory Agency (MHRA) has published its strategic approach to AI. The strategy outlines how the MHRA is, and will be, implementing the principles of the white paper: safety, security and robustness; appropriate transparency and explainability; fairness; accountability and governance; and contestability and redress.

The MHRA's AI strategy focuses on three perspectives: as a regulator of AI products, as a public service organisation delivering time-critical decisions, and as an organisation making evidence-based decisions that impact public and patient safety.

The MHRA recognises the opportunities that AI offers and the regulatory reform that is required to ensure safety while embracing the advantages of AI, such as through the launch of the AI Airlock this week. 

Regulation of AI products

The MHRA's strategy looks to the regulation of AI as a medical device (AIaMD). While the UK Medical Devices Regulations 2002 already apply to software, and its subset AI, the MHRA understands these regulations require updating to address the advancements in AI technology.

The current regulations encompass a risk-based classification system, but many AI products currently fall within the lower categories of risk and can be put onto the market without an independent assessment of conformity. The MHRA in its reform (due in 2025) will aim to ensure a proportionate approach to regulating AIaMD, taking into account the unique challenges of each product and assessing risks accordingly. 

As well as legislative changes, the MHRA in its strategy also pin points a number of guidance documents available to manufacturers and those it will be publishing in the near future. Recognising the importance of security, the MHRA will publish guidance in spring 2025 to address cybersecurity risks associated with AIaMD. To ensure transparency and explainability, the MHRA highlights guidance it has published to support manufacturers in meeting labelling and information requirements for AIaMD products.

AI Airlock

A further initiative by the MHRA referred to in the strategy is the AI Airlock, which was launched in pilot form on 9 May. The AI Airlock aims to identify and address regulatory challenges for AIaMD and is a collaborative project involving key partners such as UK approved bodies, the NHS, and other regulators.

This sandbox will assist businesses in navigating the regulatory landscape and help them comply with the regulatory framework while bolstering innovation. As well as this national approach, the MHRA is also making sure it is at the forefront international best practice via its collaboration with the International Medical Devices Regulators Forum.

By addressing these regulatory aspects, the MHRA aims to ensure a balanced approach to regulating AIaMD, promoting innovation while safeguarding patient safety and cybersecurity.

Delivery of regulatory services

The MHRA appreciates the opportunities that AI offers to enhance the efficiency of services across various regulatory functions, from market access and post-market surveillance. By leveraging AI, the MHRA aims to improve the timeliness of access to medical products for patients in the UK. This will enable the MHRA to allocate its resources more effectively, focusing on priority activities such as innovation and patient engagement. 

In addition the MHRA aims to expand the use of AI, machine learning and advanced analytics to improve how it delivers regulatory decisions in the future. The MHRA will ensure these decisions remain transparent and explainable, as well as making security, including cyber-security, central to how they operate.

Evidence-based decisions

As a regulator, the MHRA relies heavily on evidence provided by manufacturers and other third parties. With the increasing use of AI in generating evidence, the MHRA recognises the need to understand the impact of AI on evidence generation and regulation. For example, the impact this could have on clinical trial design and precision medicine. Moreover, collaborating with international regulatory bodies and industry partners, the MHRA aims to develop best practices in the use of AI and avoiding unnecessary burdens.

Osborne Clarke comment

For businesses seeking to bring AIaMD to the UK market, the principle-based regulatory approach offers flexibility and opportunities for innovation. The MHRA's regulatory sandbox also provides a unique chance for stakeholders to collaborate with the regulator and address the challenges in getting AIaMD to market.

In comparison to the EU's regulatory framework under the new AI Act, the UK's approach may be more appealing to businesses, as it avoids excessive hurdles and potential market access issues. However, it is crucial to address the risks associated with AI in healthcare, such as bias and misdiagnosis, as well as issues with patient data, which can negatively impact patient safety and healthcare delivery. As such the requirement for a regulatory framework to manage the risks of AI is necessary to some extent without stifling innovation.

The MHRA's strategy aims to strike this balance of managing the risks by ensuring proportionate regulation, as evidenced in its current reform of the Medical Devices Regulations 2002, while also demonstrating a readiness to embrace AI. How the MHRA balances the competing demands of innovation and safety in AIaMD will become much clearer when new medical device regulations are introduced next year.

An additional factor is the MHRA's collaboration with other international regulators in this area, coupled with its recently launched International Recognition Procedure. This could have an impact on the life sciences and healthcare sector if those recognised countries begin authorising AIaMD or AI platform-enabled precision medicines. An example of this is the US Food and Drug Administration's consideration of how to deal with AI platform-generated precision vaccines (although its plans are yet to be formalised).

If these recognised international regulators begin authorising new AI-based products for the market, it may prompt the UK regulator to do the same, including adopting similar regulatory review processes. Indeed, the MHRA has noted that it is working with international regulatory bodies to develop best practice in the use of AI and with a view to international alignment to reduce the burden on both industry and regulators.

As the MHRA implements its strategy, manufacturers of AIaMD should engage with the regulator to help shape the UK's regulatory framework. As AI continues to advance, its use across sectors, including life sciences and healthcare, will inevitably grow, and this collaboration between businesses and the MHRA could lead to the UK becoming an appealing destination for AIaMD products.