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

The Question We Heard

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

Safe, Fast and Trusted

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

Handover

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