Vando Gusti Al Hakim, Su-Hang Yang, Jen-Hang Wang, Chih-Yang Peng, Gwo-Dong Chen
Robots are increasingly used in education, but challenges limit their effectiveness: limited numbers, interactions restricted to class time, uniform designs that prevent ownership, and difficulty maintaining engagement over a semester. To address these issues, we propose a novel AI-powered personalized robot framework. Each robot is paired with a digital companion, accessible via students' mobile devices and enhanced by large language models (LLMs) for dynamic, personalized interactions. Students nurture their digital companions by studying, earning virtual currency, and customizing the robot's appearance and behaviors. The nurtured companion's 'spirit' is later embodied in a shared physical robot during class, integrating digital and physical learning experiences while using only one robot per classroom. A quasi-experimental study in Taiwan involved 90 Hospitality Management students training for Japanese restaurant service. Three groups were compared: AI-Powered Personalized Robots (with LLM-enhanced companions), Standardized Robots (non-personalized but accessible inside and outside class), and a Conventional Classroom (no robots). Results showed that students in the AI-Powered Personalized Robots group achieved higher learning outcomes, demonstrated greater ownership toward their studies, and engaged more frequently than other groups. This framework lays the groundwork for scalable, AI-integrated education by reimagining how learners interact with intelligent toys, fostering engagement across an entire semester. © 2025 IEEE.
Universitas Negeri, Department of Electrical Engineering Education, Yogyakarta, Indonesia; National Central University, Department of Computer Science and Information Engineering, Taiwan; Chien Hsin University of Science and Technology, Department of Hospitality Management, Taiwan; National Central University, Research Center for Science and Technology for Learning, Taiwan
