#1: Measuring Student Engagement: Leveraging Multimodal Data to Compare Human and AI Tutoring

This study compares differences in student engagement across tutor types (human, human-like AI, and nonhuman-like AI) in asynchronous online learning. Data were collected from 20 university students using eye-tracking, electroencephalography, and facial expression analysis. Results show the human-like AI tutor attracted visual attention and elicited the highest emotional responses. The human tutor sustained deeper attention and moderate emotional states. The nonhuman-like AI tutor elicited the fewest emotional reactions. This study highlights a hybrid tutoring design.