Wenzhe Xu is a PhD Student in the School of Communication and Information at Rutgers–New Brunswick. Xu’s research focuses on human-AI interaction, human-centered AI, information behavior, and the psychological and social effects of emerging media technologies. Xu’s current work examines how users interact with large language models in information-seeking contexts, with particular attention to how users’ uncertain and evolving knowledge states shape interaction patterns that may contribute to hallucination-related failures in LLM responses. Drawing on quantitative, computational, and mixed-methods approaches, Xu aims to understand how emerging technologies, such as LLMs and digital media, shape people’s everyday information interactions, experiences, and behaviors.

Learning Democracy with AI: User Uncertainty and Civic Understanding in LLM-Mediated Information Seeking

This project examines how people use large language models, such as ChatGPT, to learn about political and democratic issues. Many citizens begin with uncertainty and may not know what questions to ask or what information they need. This study explores how people’s initial uncertainty shapes their conversations with AI systems, and how AI responses may redirect users’ questions, organize political information, and influence whether they develop a clearer, incomplete, or misleading understanding of democratic issues over time. The project connects to the fellowship area on the effects of AI on democracy by examining how AI-mediated information seeking may influence civic learning, informed citizenship, AI literacy, and responsible participation in U.S. democratic life.