MHRA · Life Science & Pharma: GxP - AI

Impact of AI on the regulation of medical products (MHRA)

Final AI · Guidance Last verified: not yet · Source last checked 3 Nov 2026

Issued by
MHRA
Category
AI
Sub Category
Guidance
Legal weight
Guidance & Q&A
Reuse
Open Government Licence v3.0

The UK regulator's paper on how it applies the government's AI principles to medicines and medical devices. It is policy, not a rule.

How to read this page

Details: Guide Definitions: every term and label explained

Lifecycle

Solid steps have happened. A dashed step is expected, and it is not a fact. Only dated steps in our record appear here.

  1. Done · 30 Apr 2024Published

In plain English

MHRA policy paper on implementing the UK AI White Paper principles in the regulation of medical products.

Summary drafted with AI assistance. Not yet reviewed by the SME Reviewer. Read the official text before acting.

Context and history

The MHRA published this policy paper on 30 April 2024, setting out how it is applying the UK AI White Paper's principles to medical products regulation. The paper divides the MHRA's role into three. It regulates AI products, it is a public service that might use AI in its own work, and it makes evidence-based decisions where evidence, often supplied by third parties, may involve AI.

It mentions the AI Airlock, a regulatory sandbox for AI as a medical device, then due to launch in pilot form. It says the existing device regulations map well to the White Paper principles, while noting reform plans.

Who and what it applies to

Readers following UK AI policy on medicines and devices, including developers of AI as a medical device and firms that use AI for evidence in submissions.

What it contains

Part 1 The MHRA as a regulator of AI products
Current regulations, regulatory reform, the AI Airlock, international work through IMDRF, and guidance planned for software and AI as a medical device.
Part 2 The MHRA as a public service organisation
Using AI in regulatory services, data foundations, and exploring large language models and generative AI.
Part 3 The MHRA as an evidence-based decision maker
AI in evidence, often supplied by third parties, that informs decisions on safety.
Conclusion
Summary of its direction.

Key points

  • It is policy, not a binding requirement.
  • It reflects 2024, so check later MHRA publications.
  • The AI Airlock is a sandbox for AI as a medical device.
  • The MHRA stresses international alignment, including IMDRF.

In practice

General reading, not advice. This is drafted with AI and may not fit your situation. It does not replace the official text, your own quality system or qualified advice, and an inspector may take a different view.

  • Use it to understand the MHRA's thinking, then check newer pages for current steps.
  • If you supply AI-derived evidence, expect questions on data quality and fairness.

Read from the source document on 1 Nov 2026. This is our own description, not the official text. Check the full dated history in the change log below.

Change log

Key dated events only, not a complete history. Status after is the status once that event happened. Event kinds explained

DateEventStatus afterSource
30 Apr 2024PublishedFinalSource for Published

Official source

The official text always prevails over anything on this site.

Read the official text

Source: MHRA. Contains public sector information licensed under the Open Government Licence v3.0.

Stay updated

Newsletter signup opens at launch.