Skip to content
DraconiticResearch
Search Sources API
223,421 studies31 sourcesUpdated 43 min ago
Peer-reviewedCC BY 4.0NPJ Digital Medicine

Machine unlearning as a governance imperative for clinical AI

  • Anthony Porter1,2
  • Emily Kirkpatrick1
  • Arpit Garg1
  • Hemanth Saratchandran1
  • Simon Lucey1
  • Johan Verjans1,2,3,4
4 affiliations
  1. 1Australian Institute for Machine Learning (AIML), The University of Adelaide, Adelaide, SA Australia
  2. 2Central Adelaide Local Health Network (CALHN), Adelaide, SA Australia
  3. 3South Australian Health and Medical Research Institute (SAHMRI), Adelaide, SA Australia
  4. 4Jones Radiology, Adelaide, SA Australia
Read full textPDFDOICite

Abstract

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.

Related

  1. Clinical laboratory diagnostic stewardship as boundary infrastructure: a multi-stakeholder qualitative study of digital public health governance in two Chinese regional referral hospitals
    Long Yao, Zhuojun Deng, Fuhua Luo +1
    Frontiers in public healthPeer-reviewedCC BYFull textEpidemiology & public healthEurope PMC
  2. Shared Governance for Effective Clinical Education in Medical Radiation Science: A Narrative Review
    Jacob Leonard Ago, Bismark Ofori‐Manteaw, Minh Chau +4
    Journal of Medical Radiation SciencesPeer-reviewedCC BYFull textGeneral medicineEurope PMC
  3. From principles to practice: operationalizing responsible AI governance across predictive and generative clinical AI.
    Chung-Feng Liu, Chia-Jung Chen, Yen-Ling Ko
    Frontiers in digital healthPeer-reviewedCC BYFull textEpidemiology & public healthEurope PMCPMC
  4. Digital Health: Improve Patient Care while Preserving the Planet, an Ethical Imperative
    Brigitte Séroussi
    Yearbook of Medical InformaticsPeer-reviewedCC BY 4.0Full textEpidemiology & public healthPMC
  5. Moving Beyond the Plateau in Immune Checkpoint Inhibition-The Imperative of Early Referral and Precision Strategies.
    Razelle Kurzrock
    Journal of immunotherapy and precision oncologyPeer-reviewedCC BY-NC-NDFull textOncologyEurope PMC

Cite

BibTeXRIS
BibTeX
@article{porter2026machine,
  title = {{Machine unlearning as a governance imperative for clinical AI}},
  author = {Anthony Porter and Emily Kirkpatrick and Arpit Garg and Hemanth Saratchandran and Simon Lucey and Johan Verjans},
  journal = {NPJ Digital Medicine},
  year = {2026},
  month = jul,
  doi = {10.1038/s41746-026-03050-1},
  url = {https://doi.org/10.1038/s41746-026-03050-1},
  note = {CC BY 4.0},
}
RIS
TY  - JOUR
TI  - Machine unlearning as a governance imperative for clinical AI
AU  - Anthony Porter
AU  - Emily Kirkpatrick
AU  - Arpit Garg
AU  - Hemanth Saratchandran
AU  - Simon Lucey
AU  - Johan Verjans
PY  - 2026
DA  - 2026/07/18
JO  - NPJ Digital Medicine
DO  - 10.1038/s41746-026-03050-1
UR  - https://doi.org/10.1038/s41746-026-03050-1
AB  - 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.
SN  - 2398-6352
ER  -