Seminar Abstract:
Artificial intelligence (AI) is increasingly integrated into safety-critical domains, offering benefits but also introducing challenges in human-AI interaction. Jinyong’s research examines human trust and performance in AI-assisted decision-making under uncertainty, with healthcare as a primary application. In this seminar, Jinyong will present collaborative work with pharmacists, examining how communicating AI uncertainty information influences trust and performance in medication verification. Expanding this applied research, Jinyong developed a theoretical human-AI error categorization framework that moves beyond binary decision frameworks to capture complex, multi-class AI error patterns. Together, these studies advance both the theoretical understanding and practical guidance for designing trustworthy, human-centered AI systems.

Presenter Bio:
Jinyong Kim is a PhD Candidate in Industrial and Operations Engineering (IOE) at the University of Michigan, advised by Dr. Jessie Yang. His research is centered on human trust and performance in AI-assisted decision-making, with a primary application in healthcare. He aims to extend this agenda to guide the design of AI that supports safe decision-making. Jinyong contributes to the IOE community through service on the Graduate Application Mentoring Program (GAMP) planning committee, the IOE Community Engagement committee, and the IOE Student Leadership Board.