Perceived usefulness of AI-generated music predicts learning engagement and flow through attention and AI trust

Abstract

The integration of artificial intelligence into educational environments has prompted increasing interest in how AI-generated content is associated with student learning. This study examines the psychological pathways linking perceived usefulness of AI-generated music with university students’ learning engagement and flow experience, with mindful attentional awareness and AI trust functioning as parallel mediators. Drawing on the Technology Acceptance Model (TAM) and cognitive load theory, the study tests a structural model in which perceived usefulness predicts attention and AI trust, which in turn predict learning engagement and flow. A post-exposure cross-sectional survey was administered to 407 university students after a four-week naturalistic AI-music listening protocol. Data were analysed using partial least squares structural equation modelling (PLS-SEM) via SmartPLS 4.0. Perceived usefulness significantly predicted both attention and AI trust, and both mediators were positively associated with learning engagement and flow. All six hypothesised direct paths and four indirect paths were supported. By integrating technology evaluation, attentional disposition, and learning experience in one model, the study extends TAM beyond adoption intention and identifies two complementary pathways through which students’ evaluations of AI-generated music are associated with learning-related outcomes.