Generative AI–social media coordinated learning and university students' psychological well-being: dual pathways and the buffering role of perceived support

Abstract

Introduction: Generative artificial intelligence (GAI) and social media are increasingly integrated in university learning, reshaping collaboration and psychological outcomes. This study proposes Intelligent Media–Coordinated Learning Experience (IMCLE) to capture perceived coordination quality in GAI–social media learning, including empowerment effectiveness, collaboration facilitation, feedback visibility, and boundary safety. Methods: Using two-wave survey data from Chinese university students, we applied an explanation–prediction–necessity strategy integrating PLS-SEM, artificial neural networks (ANN), and necessary condition analysis (NCA). We examined how IMCLE cues relate to collaborative learning (CL), psychological distress (PD), and psychological well-being (PWB), with perceived support (PSup) as a moderator. Results: All IMCLE cues positively predicted CL, which in turn enhanced PWB. IMCLE cues also positively predicted PD, and PD formed a significant indirect pathway to PWB, indicating that gains and strain may co-occur in highly coordinated learning. PSup weakened the IMCLE.PD relationships and attenuated the PD.PWB association. ANN confirmed the predictive salience of key IMCLE cues and showed nonlinear importance patterns, while NCA identified threshold conditions for high PWB. Discussion: IMCLE has a dual effect, producing both collaborative gains and psychological strain. The findings inform feedback governance and support infrastructure design to improve students' digital well-being.