Why AI Needs Proportionate Lifecycle Regulation
National Commission into the Regulation of AI in Healthcare — Recommendations Launch
Neil D. Lawrence
National Commission into the Regulation of AI in Healthcare — Recommendations Launch Event, MHRA
National AI Commission — commissioner framing
The Commission’s Question
- AI is already in healthcare
- Support is strong — and conditional
- People asked how it can be used safely, fairly and transparently
Designed for a Different Era
Static Products, Evolving Systems
- Much of today’s framework assumes a product assessed at a point in time
- AI-enabled products can iterate, drift and perform differently by setting
- A one-off pre-market snapshot is not enough
What the Technology Working Group Saw
Collaboration — and Confusion
- An enthusiasm across government, regulators, industry and the NHS to work together
- A landscape where it is often unclear who is in charge
- Gaps where no one acts; overlaps where everyone does
Proportionate Lifecycle Regulation
What Chapter 1 Proposes
- Clear qualification and classification — when is a product a medical device?
- Risk-proportionate oversight — including function-based regulation
- Evidence and change managed across the lifecycle, not only at launch
- Staged routes to market, sandboxes and stronger post-market learning
Not a False Trade-off
- Safe — protecting patients and managing risk across the lifecycle
- Fast — responsible innovation and timely access to benefit
- Trusted — confidence for patients, professionals, providers and developers
From Direction to Delivery
- The report sets direction; implementation is collective
- The MHRA and partners will show what proportionate regulation looks like in practice
- Then industry — and your questions