AI Fairness: Reducing Biases towards Minoritized Populations in Medical curriculum - September 2024

In this Responsible AI in Medicine seminar (September 16, 2024), Dr. Shiri Dori-Hacohen, Assistant Professor at the School of Computing at the University of Connecticut and head of the Reducing Information Ecosystem Threats (RIET) Lab, will present. Dr. Dori-Hacohen will discuss BRICC, a pioneering initiative that addresses "bisinformation" in medical curricula by using machine learning to identify and flag biased content. The project involved creating a gold-standard dataset of over 12,000 pages, annotated for biases like gender and race. Various classifier models were tested, with binary classifiers showing the most promise in detecting general biases. The lecture will explore how this work lays the foundation for more effective debiasing of medical education.