OH Changhoon

Professor, Graduate School of Information, Yonsei University

OH Changhoon
Times of the Remarks 2026. 11. 05. 14:45-15:15
Title People Who Don’t Verify: Waking the Sleeping Button in the Age of AI News

OH Changhoon is a professor at the Graduate School of Information at Yonsei University and a Human-Computer Interaction (HCI) researcher whose work focuses on the points of contact between people and AI. He received his bachelor’s and doctoral degrees from Seoul National University and conducted postdoctoral research at institutions including Carnegie Mellon University in the United States. His research on Human-AI Interaction has examined, through both experiments and field studies, how people perceive algorithms and AI, and how far they are willing to trust them.

At the center of his work in the age of AI is a fundamental question: “What should we delegate to AI, and what should we not hand over?” He studies how user experiences can be designed to encourage people to reflect on their own judgment, so that they neither over-rely on AI-generated answers nor distrust them excessively.

He has regularly presented his research at ACM CHI, one of the leading international conferences in the HCI field, and has received multiple Honorable Mention Awards at CHI and DIS. At Yonsei University, he leads the AX Lab, where he collaborates with companies including Samsung Electronics and Shinhan Bank on user experience in the age of AI, while also teaching courses such as Generative AI UX.

In an era when AI summarizes and rewrites the news, people rarely click on the source links placed beneath its answers. The option to verify is there on the screen, but it often remains unused. Paradoxically, simply knowing that verification is possible can itself make people feel reassured.

This talk presents experimental research using an environment modeled on an AI news agent to examine how different designs of “source verification” buttons and verification labels affect people’s checking behavior and their judgments about whether information is true or false.

In particular, the study asks whether traditional media verification labels, such as “Verified by the newsroom’s fact-checking team,” still function as the strongest signals of trust, and whether people can be encouraged both to trust more and to verify more when such labels are designed not as the final word, but as a gateway to the underlying evidence.

Based on these findings, the talk offers practical suggestions for how news organizations and platforms should rethink the design of news experiences in the age of AI.