Designing an Assistant System for Digital Reading-Based Educational Environments : A Comparative Analysis of Graphic User Interface (GUI) and Human-Robot Interaction (HRI) Based on Multimodal Data

This study explores the potential of Human-Robot Interaction (HRI) to enhance “digital reading” experiences, one of the most common computer-based learning methods in higher education. To this end, we conducted a comparative analysis of interaction designs between a Graphical User Interface (GUI)-based reading system and an HRI system that provides reflective and empathic feedback through a humanoid robot. A comprehensive dataset, including eye-tracking and self-report data, was collected to analyze the additional value HRI can offer compared to traditional GUI-based systems through both objective and subjective indicators. Based on the hypothesis that ‘reflection’ and ‘empathy’ in robot feedback play a critical role in learning support within HRI design, we compared the two systems across four dimensions: 1) reading behavior analysis using eye-tracking data, 2) cognitive performance evaluation through pre- and post-tests, 3) perceived interaction quality measured by the AttrakDiff questionnaire, and 4) social perception of the interface assessed through the Social Presence questionnaire. The results showed that HRI-based digital reading positively influenced learners’ attentional focus, robot-looking behavior, and page re-trace actions, leading to sustained changes in learning behaviors beyond short-term effects. Furthermore, the HRI condition demonstrated higher knowledge gain, more positive hedonic experiences, and stronger social presence compared to the GUI condition. These findings suggest that reflective and empathic feedback design within HRI can effectively enhance digital reading learning experiences.