From Deep Learning to Digital Pedagogy: A Bibliometric Study of Artificial Intelligence in Music Education

Abstract

Artificial intelligence is finding more and more applications in the field of music education in the form of automated performance feedback, adaptive practice systems, and online learning platforms. Since the sphere of this multidisciplinary approach that includes education, music, and computing is growing at a rapid pace, a bibliometric snapshot can be used to explain the trend of the literature development and reveal the areas of its focus. This paper presents a literature review on AI and music education and maps it with a Scopus-export dataset of 231 articles on that topic that were published between 2002 and 2025 in English. The analysis unites the increase in publications, leading journals, countries in which published papers have most of the authors, keywords primarily used, and the high-citation publications. The analysis indicates that there is a strong recent trend: 223 out of 231 publications have been released since 2021, and the trend is higher in 2024 (71 articles) and 2025 (52). Applied Mathematics and Nonlinear Sciences (23 articles), Computational Intelligence and Neuroscience (13), and Wireless Communications and Mobile Computing (11) are the most referred sources, meaning that there is a very strong representation of computational and engineering sources. The pattern of affiliation reveals that China is the leading contributor (183 papers) then South Korea (13) is followed by Malaysia (8). The topic of keyword frequency focuses on the music, music education, artificial intelligence, and deep learning, which indicates the field that is inclined towards AI-driven systems and modelling. The article with the highest number of citations in the data set is the article titled development and applications of artificial intelligence in music education (2023) which has 95 citations. In general, the literature is growing at a high rate and it is geographically and methodologically focused which means that it needs to be widely collaborated and well education-based assessment of learning outcomes in real classroom environments.