When
Where
Information Disclosure and Acceptance of Algorithmic Advice: A Field Study on Human-AI Interaction in Healthcare Gatekeeping
Abstract: This study explores the impact of technical and social information on human acceptance of algorithmic recommendations in a healthcare context. Specifically, we examine interactions between patients and an artificial intelligence (AI) gatekeeping app that recommends specialty care based on symptoms provided by patients. In a field experiment conducted at a large hospital, patients were randomly assigned to four groups, each receiving AI recommendations in one of four information disclosure scenarios: (1) recommendation only, (2) recommendation with the technical information of algorithm accuracy, (3) recommendation with the social information of peer acceptance, and (4) recommendation with both algorithm accuracy and peer acceptance. Our analysis shows that patient acceptance of AI recommendations increases in algorithm accuracy or peer acceptance when each information is presented independently. When presented together, algorithm accuracy and peer acceptance are complementary in influencing patient acceptance of AI recommendations, demonstrating a positive interaction effect. However, their marginal effects are asymmetric, with peer acceptance dominating algorithm accuracy. Furthermore, the impact of information disclosure depends on patient expertise level: patients with less capability of articulating symptoms or new patients without previous hospital experience are more likely to be influenced by the disclosed technical and social information.
Bio: Susan Feng Lu is a Professor of Operations Management and Statistics and holds the Alan Hudson Chair in Health Policy at the Rotman School of Management, University of Toronto. She earned her Ph.D. from the Kellogg School of Management, Northwestern University and has established herself as a leading researcher at the intersection of health economics, operations, and analytics. Professor Lu’s research explores the operational drivers of healthcare delivery, focusing on how public policies and technological innovations influence the efficiency, quality, and equity of healthcare systems. With expertise spanning health technology, nursing home operations and cardiac care delivery, she employs empirical methods and machine learning to address pressing challenges in healthcare operations. Her research has been published in prestigious journals, including Management Science, MSOM, POMS, Review of Economics and Statistics, Journal of Health Economics and Information Systems Research, as well as influential interdisciplinary outlets such as Science and Nature Scientific Reports. Professor Lu’s work has earned numerous awards, significant research grants, and wide-reaching media attention, with features in outlets like Barron’s, Nature, Freakonomics, and Vox. Beyond her research, Professor Lu contributes to the academic community as an Associate Editor for Management Science, MSOM, and POMS. She is currently the President of the Chinese Economists Society. Besides, she has also provided expert testimony before the United States Senate (2023) and served as an Economist Mentor for the Creative Destruction Lab, fostering innovation and entrepreneurship.