Can AI Instructional Agents Better Foster Student Engagement? A Multimodal Neurophysiological Comparison with Human Teachers

• In Event: Designing Engaging Learning Environments: Insights from AI, VR, and Games

Abstract

Learning analytics comparing AI instructional agents and human teachers provides valuable insights for designing AI-based learning. In a previous randomized controlled trial, our team found that AI agents enhanced students’ perceived control and academic performance compared to human teachers (Authors, 2025). Building on this work, the present study further analyzed students’ multimodal neurophysiological data across instructional contexts. Students taught by human teachers generally exhibited stronger electrodermal activity, indicating higher emotional arousal. Compared to the lecture phase, interacting with human teachers evoked arousal, while interacting with AI agents increased both skin conductance variability and the proportion of high-frequency EEG activity, suggesting heightened cognitive and emotional engagement. These findings highlight the complementary roles of AI agents and human teachers in promoting engagement.

Authors

• Xiaobo Liu, Tsinghua University

• Zhanxin Hao, Tsinghua University

• Jifan Yu, Tsinghua University

• Fei Qin, Tsinghua University

• Zhiyuan Liu, Tsinghua University

• Yu Zhang, Tsinghua University