Ethical Considerations of Using Machine Learning for Decision Support in Occupational Health: An Example Involving Periodic Workers' Health Assessments.
Marianne W M C Six Dijkstra, Egbert Siebrand, Steven Dorrestijn and 6 others
PMID 32500471WHAT IT FOUND
Machine learning tools that recommend occupational health interventions may be hard to verify, may change after consent, and can expose workers to privacy or discrimination risks.
Key findings
01Workers may consent to use of their health data for scientific purposes without knowing how a machine learning tool will later be modified or used.
02Machine learning recommendations can be non-transparent and may be invalid without users realizing it.
03Machine learning profiles could be used to exclude workers from tasks or jobs.
STILL TO COME
How it was doneWhat they found
Read the rest of this summary
You get three full summaries a month, free, and we do not ask for a card. Search, the TL;DRs and your library stay unlimited either way.
What it does not show
The paper is an ethical deliberation and narrative review, not a clinical study with patients or outcomes. The occupational health scenario is hypothetical, so it illustrates possible concerns rather than observed effects. The authors note their ethical framework is based on Western biomedical principles and may not reflect all cultures. The paper does not report performance data, validation results or harms from an actual machine learning tool.
Declared interests
The paper names funding from Nederlandse Organisatie voor Wetenschappelijk Onderzoek (NL) and Saxion University of Applied Sciences. It does not report other conflicts of interest.
The easy way to misread this
Do not read this paper as evidence that machine learning decision support improves occupational health outcomes. It reports ethical concerns and recommendations, not tested effects in patients.