Fostering adolescent engagement in generative AI art therapy: a dual SEM-ANN analysis of emotional

Abstract

Introduction: This study explores the application of generative artificial intelligence (AI) art in digital art therapy, focusing on how it influences adolescents’ interest-driven participation. With mental health concerns rising among youth, understanding motivational mechanisms in AI-assisted therapeutic tools is both timely and essential. Methods: A cross-sectional survey was conducted with 444 junior and senior high school students in Hubei Province, China. The study integrated Emotional Design Theory and the Technology Acceptance Model (TAM) to construct a predictive model. Structural equation modeling (SEM) and artificial neural network (ANN) analyses were employed to validate the model and identify key predictors of engagement. Results: SEM results indicated that perceived usefulness (PU), perceived ease of use (PEOU), perceived fun (PF), and perceived trust (PT) significantly influenced users’ attitudes toward use (ATT) (p < 0.001). ATT, PF, and PT were strong predictors of interest-driven participation, while the behavioral level had no direct effect on perceived enjoyment (PE). ANN analysis further highlighted ATT as the most influential predictor (100% normalized importance), notably exceeding PE (19.3%). Discussion: These findings emphasize the importance of intuitive design, seamless interaction, and trust-building in sustaining adolescents’ engagement with AI-based art therapy. The study provides a theoretical foundation for understanding interest formation in youth and offers practical implications for improving emotional design, digital therapeutic tools, and mental health interventions.