- 제목
- Bridging AI and Society: Analyzing Systemic Bias in Multilingual Wikipedia (DR. Chan Young Park)
- 작성자
- 첨단컴퓨팅학부
- 작성일
- 2025.02.18
- 최종수정일
- 2025.02.18
- 분류
- 세미나
- 게시글 내용
-
일시: 2025. 2. 19.(수) 14:00
장소: 제4공학관 D504
Presentor: Chan Young Park, Postdoctoral Researcher at the University of Washington
Title: Bridging AI and Society: Analyzing Systemic Bias in Multilingual Wikipedia
Abstract: As large language models (LLMs) increasingly excel at capturing the semantics of human language, new challenges arise: AI systems must also understand the social dimensions of language. In this talk, I will briefly introduce my research on developing socially-aware AI before turning to one of its key applications: employing LLMs to identify and mitigate social biases in Wikipedia, a global knowledge repository with significant societal impact. I will highlight two studies that examine how Wikipedia’s language editions may subtly, but systematically, portray people differently. In the first study, we analyze portrayals of individuals across multiple language editions while controlling for potential confounders, revealing how seemingly neutral content can subtly encode social biases. Building on these findings, the second study develops a computational method to pinpoint exactly where narrative inconsistencies and information gaps occur, illustrating how diverse cultural norms shape differing versions of the same story. Together, these studies illustrate how advanced AI methods can enable large-scale analyses of social bias, offer actionable insights for mitigating its effects, and ultimately contribute to more equitable information ecosystems.
Bio: Chan Young Park is a postdoctoral researcher at the University of Washington, working with Professor Yejin Choi. She earned her PhD in Computer Science from Carnegie Mellon University under the supervision of Professor Yulia Tsvetkov. Her research sits at the intersection of NLP, AI ethics, and computational social science, with a focus on building socially-aware AI systems that are adaptable and equitable. Her work has received several notable awards, including the ACL Best Paper Award and the Research Award of the Year by Wikimedia Foundation, and was featured in MIT Tech Review and the Washington Post. She was also named a University of Chicago Rising Star in Data Science and is a recipient of the K&L Gates Presidential and Korea Foundation for Advanced Studies Fellowship.