People turn to language models for information, advice, and emotional support. In this lecture, Stanford professor Dan Jurafsky shares research from his lab on how these conversations can distort users’ judgment, encourage overreliance, and diminish their sense of personal responsibility. The studies identify recurring patterns: models agree too readily with users, express more certainty than warranted, present themselves in overly human terms, and answer from their own perspective without adequately accounting for what users know. Jurafsky will discuss datasets and metrics used to study these behaviors and explore ways to reduce harm and protect users’ well-being.

About Dan Jurafsky

Dan Jurafsky is the Professor of Linguistics, Professor of Computer Science, and Reynolds Professor in Humanities at Stanford University. He is an award-winning teacher, a MacArthur Fellow, and a recipient of the Richard C. Atkinson Prize in Psychological and Cognitive Sciences from the National Academy of Sciences. He is a member of the National Academy of Sciences and the American Academy of Arts and Sciences and a fellow of the Association for Computational Linguistics, the American Association for the Advancement of Science, and the Linguistic Society of America.

Jurafsky and his students study natural language processing and large language models, including their applications in the cognitive and social sciences and their potential for social good. His books include the co-authored textbook Speech and Language Processing and The Language of Food, a 2014 international bestseller nominated for a James Beard Award.

Individuals with disabilities requiring accommodations may contact the Office of Campus Activities at (845) 437-5370.

Sponsored by the Computer Science Department.

This event is free and open to the public.