The Great Debaters Meet the Great Facilitators: How Debate Students Evaluate Human Moderators and Their AI Counterparts

ABSTRACT Background Debate cultivates reasoning, logic and persuasion, but its impact depends on skilled moderation, a role traditionally reserved for humans. Advances in large language models (LLMs) make Artificial Intelligence (AI) moderators a plausible alternative. Yet, little is known about how students perceive AI moderators compared to their human counterparts in educational debates. Objectives This study examines how university students evaluate human versus AI moderation in a synchronous, text‐based debate platform, focusing on moderation effectiveness, enjoyment, satisfaction, perceived utility and intention to use. Methods A quantitative experimental study was conducted using a synchronous text‐based debate platform. The study involved 32 university students from various universities in the United States who participated in 32 debate sessions. Each participant engaged in two debates moderated by human moderators, AI moderators or a combination of both across four moderator conditions: Human–Human, Bot–Bot, Human–Bot and Bot–Human. Following each debate, participants completed a questionnaire assessing moderation effectiveness, hedonic motivation, satisfaction, perceived utility and usage intention, resulting in a total of 64 observations. Data were analysed using descriptive statistics, independent‐sample t ‐tests, multiple regression analyses and mediation analyses to examine differences between moderator types and the relationships among the study constructs. Ethical approval was obtained from the university's Research Ethics Committee, and informed consent was secured from all participants. Results and Conclusions Our analysis reveals no significant differences between human‐ and AI‐moderated debates across all measured dimensions. Across conditions, perceived effectiveness of moderation and hedonic motivation to use the platform predicted satisfaction and perceived utility. Perceived utility significantly drove usage intention, especially under AI moderation, whereas satisfaction more strongly predicted usage intention with human moderators. These findings suggest that, in well‐scaffolded, text‐based contexts with clear rules, thoughtfully designed AI moderators can match human facilitation while offering scalable support for equitable participation and structured discourse.