The SRI Seminar Series welcomes Lydia T. Liu, an assistant professor of computer science at Princeton University. Liu’s research examines the theoretical foundations of machine learning and algorithmic decision-making, with a focus on the long-term societal consequences of automated systems. She is a faculty affiliate of Princeton’s Center for Information Technology Policy, Center for Statistics and Machine Learning, and Princeton Language and Intelligence. In this session, Liu will argue that evaluating AI requires looking beyond standalone measures of capability to the social and organizational settings in which people and algorithms work together. Drawing on research in higher education and other domains—including causal analysis of algorithm-assisted college advising and emerging qualitative evidence on how advisors use AI-enabled data platforms—she will identify forms of expertise that conventional evaluations tend to miss: knowledge of individual circumstances, relational work, institutional constraints, and judgments about when an algorithmic recommendation is worth acting on at all. Liu will invite participants to consider how organizations might deploy AI in ways that recognize and preserve human expertise. Talk title “Expertise in the age of human-AI collaboration” Abstract As AI systems become increasingly embedded in consequential decisions, a central question is not simply whether their predictions are accurate, but how they interact with human expertise and downstream interventions. What do human experts know that AI systems do not? When do decision-makers act on algorithmic recommendations, when do they exercise discretion, and what happens to expertise as institutions reorganize work around AI? In this talk, I argue that answering these questions requires moving beyond standalone measures of AI capability towards understanding human-AI systems and collaboration in their social and organizational contexts. Drawing on ongoing work in higher education and other domains, I will examine settings in which human decision-makers work alongside predictive and generative AI tools. These studies reveal forms of worker expertise and productivity that are difficult to capture in standard evaluations of AI: knowledge of individual circumstances, relational work, institutional constraints, judgments about when and how to intervene, and even judgments about when an algorithmic prediction is useful in the first place. I will discuss empirical and conceptual approaches for identifying these contributions, including causal methods applied to algorithm-assisted college advising as well as emerging qualitative evidence on how advisors interact with AI-enabled data platforms in practice. Together, these results complicate familiar narratives of automation and augmentation, raising the question of how to design organizational deployments of AI that recognize and preserve human expertise. Moderator: Shion Guha, Faculty of Information Registration To register for the event, visit the official event page. Please note the “Add to calendar” button on this page does not register you for events. Please follow relevant registration instructions. About the speaker Lydia T. Liu is an assistant professor of computer science at Princeton University. Her research examines the theoretical foundations of machine learning and algorithmic decision-making, with a focus on long-term societal impact. She obtained her PhD in electrical engineering and computer sciences from the University of California, Berkeley, and completed her postdoctoral research in the Artificial Intelligence, Policy, and Practice initiative at Cornell University. Liu is a faculty affiliate of the Center for Information Technology Policy, the Center for Statistics and Machine Learning, Princeton Human Computer Interaction, Princeton Language and Intelligence, and the Princeton AI Lab. She is the recipient of awards and fellowships from the Spencer Foundation, Amazon Research, a Microsoft Research, and Open Philanthropy, and an ICML Best Paper Award. About the SRI Seminar Series The SRI Seminar Series brings together the Schwartz Reisman Institute for Technology and Society’s research community and wider audiences for a robust exchange of ideas that advances scholarship at the intersection of technology and society. Seminars feature leading and emerging scholars and provide substantial opportunities for discussion. Each seminar includes a presentation and an open discussion, with the format varying according to the speaker and topic. Registered attendees will receive a Zoom link before the event. Select seminars may be recorded and made available online.
A curated event at University of Toronto. Recommended by The Mindful City.
Event details are aggregated from public sources and may be incomplete or out of date. Dates, times, and speakers can change. Always confirm with the organizer before attending.