This study explores the connection between generative AI tool usage and perceived self-regulated learning, focusing on the mediating roles of technological self-efficacy and cognitive offloading in blended learning environments, grounded on the adapted UTAUT2 model and self-regulated learning theory. The study also examined students’ perceptions about the opportunities and challenges associated with self-regulated learning and cognitive offloading in the context of using generative AI tool usage for academic purposes. We surveyed 861 students from STEM and social science programs using a 42-item questionnaire in blended learning courses in Pakistan, China and Finland. A mixed method approach was applied with partial least square structural equation modeling (PLS-SEM) with qualitative content analysis. Results indicated a positive and significant relationship between generative AI tool usage and self-regulated learning, mediated by technological self-efficacy through cognitive offloading. A finding of serial mediation demonstrates that enhanced technological self-efficacy and cognitive offloading significantly associated with self-regulated learning. The results also revealed that students perceived GenAITU positively and significantly connected with SRL and cognitive offloading but highlighted some challenges like over-reliance, cognitive overload, metacognitive laziness and potential erosion of essential academic skills. The study offers practical implications for practitioners like universities, faculty, and curriculum developers to integrate generative AI tool usage to enhance technological self-efficacy and cognitive offloading, ultimately fostering improved self-regulated learning in blended learning environments.
Hashmi et al. (Thu,) studied this question.