Yejin Choi

Yejin Choi is a computer scientist at Stanford University and a senior fellow at the Stanford Institute for Human-Centered Artificial Intelligence. She earned her doctorate at Cornell University and previously worked at Stony Brook University, the University of Washington, and the Allen Institute for AI. Her research centers on language and common-sense reasoning.

Choi co-authored COMET, a system that learned to generate common-sense descriptions from knowledge-graph examples. This work explored whether language-generation models could supply plausible information that people often leave unstated. Her broader research investigates model capabilities, limitations, and alternative approaches to training systems that reason about everyday situations.

She received a MacArthur Fellowship in 2022 and has contributed to public discussions about the strengths and weaknesses of modern AI. Her work combines empirical research with questions about what model performance actually demonstrates. The recurring theme is that fluent language alone does not settle whether a system has the reasoning abilities needed for dependable use.