Explaining student engagement and satisfaction in AI-supported learning: a study demands–resources and cognitive load perspective

Purpose This study integrated the SD-R and CLT to explain how resources, demands, and cognitive load shape engagement and satisfaction in AI-mediated higher education. Design/methodology/approach A quantitative, non-experimental survey was conducted with 432 Indonesian university students. Data were analysed using PLS-SEM to test the effects of study demands, study resources, personal resources, and cognitive load dimensions on engagement and satisfaction. Findings Lecturer support, self-efficacy, self-compassion, and AI use increased engagement and indirectly supported satisfaction, with personal resources emerging as the strongest predictors. The burnout classroom climate reduced satisfaction through cognitive load. Intrinsic and germane loads were positively related to satisfaction, whereas extraneous loads were negatively related. Practical implications Lecturers should provide clear task structures, guidance on responsible AI use, and feedback that supports productive cognitive efforts. Reducing burnout classroom climate is also recommended. Originality/value This study extends SD-R beyond motivational pathways by embedding differentiated cognitive load, while extending CLT beyond task-level explanations by situating cognitive effort within study demands and resources in AI-supported learning.