Prompting somatic practice performance with AI-facilitated peer-assisted learning: A self-determination theory perspective

Abstract

Incorporating peer-assisted learning (PAL) into curricula to foster students’ feedback skills, reduce their reliance on teachers, and enhance their confidence in learning as well as strengthening connections among peers has been recommended by scholars; however, the ability of students, who are inexperienced peer tutors, to guide others by providing feedback based on individual peers’ needs is frequently questioned. To address this issue, this study proposed an artificial intelligence(AI)-supported PAL mode based on the perspective of self-determination theory (SDT). Additionally, an AI-facilitated PAL system was developed to serve as scaffolding to support students while executing PAL tasks. To explore the effectiveness of this learning mode, this study adopted a quasi-experimental design involving two classes of students enrolled in a physical education course at a technological university. One class of 44 students served as the experimental group adopting the AI-PAL mode, while the other class of 46 students served as the control group adopting the conventional (C-PAL) mode. The collected data were analyzed using SPSS for ANCOVA. The results indicated that the AI-PAL mode significantly enhanced students’ skills performance, learning motivation, autonomy, competence, and relatedness, compared to the C-PAL mode. Furthermore, the interview results revealed the contributions of this mode to enhancing learning effectiveness, promoting the quality of PAL, and reducing teachers’ workload. This study demonstrates the system development, application, and learning benefits of the AI-PAL mode in the physical education context. The findings could be a reference for those who intend to improve students’ skills performances using the PAL approach; moreover, it could also be a reference for developing policies to guide teachers to adopt AI in educational settings.