Year: 2026
Type: review
Abstract: The current overoptimism surrounding AI threatens to disrupt clinical environments. We discuss the risks related to invasive automation and describe how AI performance exaggeration may lead to clinical errors. We also investigate cognitive damage to clinicians: automation bias and de-skilling, by exploring their relation to well-defined psychological principles. Finally, we propose solutions for automation that actually benefit medicine: generalise model bias assessments and quality control, and ensure that AI integration does not erode clinicians' clinical expertise.
Source: Danish Medical Journal
DOI: 10.61409/a10250852
Authors: Joris Fournel, Aasa Feragen, Martin Tolsgaard
Language: English
Subfield: Health Informatics
Field: Medicine
Domain: Health Sciences
Open Access status: hybrid
OpenAlex ID: W7144021948