Join us for a webinar with Riccardo Fogliato :
Who goes first? Influences of Human-AI Workflows on Decision Making in Clinical Imaging
Algorithmic risk assessment instruments are increasingly used to support human decisions in high-stakes domains, making it crucial to understand how their design and use influence outcomes. In this talk, I present a user study investigating interactions between RAIs and human decision-makers, specifically veterinary radiologists diagnosing medical conditions from X-rays. The study compares two distinct workflow configurations: a one-step workflow, where radiologists receive AI recommendations simultaneously with the images, and a two-step workflow, where radiologists make provisional diagnoses first and view AI recommendations afterward. We examine how these workflows influence anchoring effects, diagnostic accuracy, agreement between radiologists and AI, and other aspects of decision-making.
Link to the paper associated with the talk: https://dl.acm.org/doi/abs/10.1145/3531146.3533193
About the presenter:
Riccardo Fogliato is an applied scientist in Amazon Web Service's Responsible AI team. He works on evaluating vision and language models, developing statistical methods for scalable assessments. He holds a PhD in Statistics from Carnegie Mellon University.
About the AVI Webinar Series: The purpose of the AVI Webinar Series is to provide a vendor-agnostic forum for education and networking. Join us to learn from academic, industry, and government organizations who are using information technology and data science to improve the lives of animals and veterinary healthcare teams.
Contact Pam Okerholm, AVI Secretary with any questions at secretary@avinformatics.org
Members: Please login at the top of our website prior to registering.
