As ambient AI scribes rapidly enter routine healthcare, they generate rich, highly sensitive data streams, including audio, transcripts, and clinical notes, that raise questions regarding downstream reuse. This talk presents findings from a cross-sectional study (n=309) evaluating patient preferences for consent and comfort and acceptability regarding the secondary use of scribe-generated data. While patients accept clinical and quality improvement uses, comfort drops significantly for developing new AI tools, and acceptability falls sharply for identifiable data and commercial uses. We also find that patients consistently want clear notification, which lies in tension with the increasing ubiquity and invisibility of ambient AI use. In this presentation, we will discuss how these empirical patient-defined boundaries could shape governance and consider how emerging AI-enabled workflows can support research and ethical values.

Bio: Jodyn Platt  is an Associate Professor of Learning Health Sciences at the University of Michigan. Her work explores how data and technology can be used responsibly to improve health while earning and sustaining public trust. She studies the ethical, social, and policy questions that arise as health systems become more data-driven and powered by artificial intelligence. She works closely with patients, communities, and healthcare organizations to understand how people think about and experience data use in health and medicine.