AI-assisted feedback and reflection in vocal music training: effects on metacognition and singing performance

Metacognition plays a vital role in enhancing learning outcomes and has received increasing attention in recent years. Studies have shown that accomplished musicians typically demonstrate high levels of metacognition, and that reflection and feedback are effective strategies for promoting metacognitive development. This study explores the impact of integrating artificial intelligence (AI) and e-learning tools into vocal music training. It focuses on feedback and reflection interventions aimed at enhancing the metacognitive abilities and singing performance of pre-service teachers.

An experimental design was employed over a six-week training period. Participants were randomly divided into a control group (N = 42), which received conventional singing instruction, and an experimental group (N = 38), which received additional interventions comprising: (a) self-assessment through the use of an audio comparison tool, (b) dialogic feedback through interaction with a large language model (Yuanbao, Tencent's generative AI chatbot), and (c) engagement in self-reflective journal writing. A two-way repeated measures ANOVA was employed to examine the interaction effects between time (pre-test vs. post-test) and group (experimental vs. control). In addition, linear mixed models were used to analyse the relationship between metacognitive abilities and singing performance.

The results demonstrated that AI-assisted training significantly affects the development of metacognitive abilities. While both the experimental and control groups exhibited significant improvements in singing performance following the intervention, no significant interaction effect between the group and time was detected. No correlation was found between metacognition and singing performance.

The significance of this study is its provision of an effective implementation framework for integrating AI and e-learning tools into music instructional practice. These technologies offer high-quality personalized feedback and foster deep reflective engagement, thereby supporting the metacognitive development process in music education contexts.