Machine unlearning as a governance imperative for clinical AI
Clinical AI assumes that the influence of training data can persist indefinitely. This premise fails when patients withdraw consent, evidence evolves, or bias is identified. Machine unlearning aims to remove specific data influence without full retraining. We argue that unlearning readiness should be built into the infrastructure of high-risk healthcare AI across patient autonomy, clinical validity, and system governance, and we outline a governance pathway to keep updates auditable and clinically safe.
Authors
- Simon Lucey (ORCID: https://orcid.org/0000-0002-6326-042X)
- Hemanth Saratchandran (ORCID: https://orcid.org/0009-0007-7995-2731)
- Arpit Garg
- Johan Verjans
- Anthony Porter
- Emily Kirkpatrick
Institutions
- South Australia Pathology (AU)
- SA Health (AU)
- Australian Centre for Robotic Vision (AU)
- South Australian Health and Medical Research Institute (AU)
- Central Adelaide Local Health Network
- The University of Adelaide (AU)
Publication Details
- Journal
- npj Digital Medicine
- Published
- 2026-07-18
- DOI
- https://doi.org/10.1038/s41746-026-03050-1
- Primary Topic
- Artificial Intelligence in Healthcare and Education
- Type
- article
- Field-Weighted Citation Impact
- 0.00