Hidden Biases in AI-Powered Pathology: How Data Center Signatures Undermine Cancer Detection Models
Editor2026-01-26T11:50:20+00:00by Christos Evangelou, MSc, PhD – Medical Writer and Editor The combination of digital pathology with artificial intelligence (AI) shows promise in improving cancer diagnosis by enhancing analysis of histopathology images. However, recent studies revealed concerning biases in these systems that could impact their reliability in clinical settings. In a new study, researchers from Ontario Tech University, Brock University, and Wilfrid Laurier University in Canada identified factors that contribute to the ability of deep learning models designed [...]










