WIREDGOV

In February 2024 the Secretary of State for DSIT (the Department for Science, Innovation and Technology) wrote to several regulators, including the MHRA (Medicines and Healthcare Products Regulatory Agency), requesting the publication of their respective strategic approaches to artificial intelligence by 30 April 2024.

The approach set out by MHRA on 30 April is aligned to the government’s AI White Paper’s five principles; safety, security and robustness; appropriate transparency and explainability; fairness; accountability and governance; and contestability and redress.

The MHRA has reaffirmed that the current Medical Devices Regulations from 2002 still stand, and that where artificial intelligence is adopted for use as a medical device it must conform to these requirements across a product’s whole lifecycle. The organisation also highlights the legal obligations for manufacturers in terms of their responsibility and accountability to mitigate risks and address product safety and performance.

The MHRA’s strategic approach to AI includes a variety of new updates, including:

• The MHRA’s regulatory sandbox AI-Airlock, created with the aim of supporting the development and deployment of AI and software-based medical devices, and being developed with funding from the Department of Health and Social Care’s AI Lab, will launch in pilot form this Spring;
• Several AI products will be “up-classified” from lower-risk classifications (as set out in the Medical Devices Regulations’ risk-based classification system) to higher risk classes, which will entail greater scrutiny;
• Plans for new guidance, including:
  • on cyber security for software and AI as medical devices will be published by Spring 2025;
  • on applying human factors to medical devices using AI (due for release in Spring 2025);
  • On good practice in relation to machine learning in medical device development (release timescale unclear);
  • on best practice for the development and deployment of AI as a medical device.

The strategic approach to AI publication also states there are currently “approximately three full-time equivalent employees” working on workstreams in relation to the use of AI as a medical device, and that this is expected to rise over the next 12 months to approximately 7.5 full-time employees through the £10 million governmental funding provided to MHRA as part of the government’s response to AI regulation.

Looking inward to the MHRA’s own operations as an organisation, the MHRA’s strategic approach to AI notes there are approximately 15 full-time staff working on the organisation’s approach to expanding and optimising AI in the delivery of regulatory services. The publication also several further organisational updates, including:

• An MHRA data strategy is in development, which is due to focus how the organisation will apply artificial intelligence in a safe and responsible manner;
• a “deliverable” on the role of large language models and Generative AI use in the organisation;
• MHRA is exploring the use of “supervised machine learning” to support assessors to review initial documents submitted as part of market authorisation and approval applications;
• MHRA are “working with a developer of a generative AI tool that helps users write documents or analyse data by providing suggestions and feedback".