Generative AI as Learning Catalyst: Interest Development in Social Science Education

Abstract

This mixed-methods study investigates the role of Generative Artificial Intelligence (GAI), specifically ChatGPT, in stimulating learning interest among social science undergraduates at a major Indonesian public university. Moving beyond the prevalent discourse on AI's efficiency, this research employs a sequential explanatory design, integrating a quasi-experiment (N=240) with in-depth phenomenological interviews (N=25). It examines not only the statistical relationship but, more critically, the mediating psychological and pedagogical mechanisms through which AI tools influence the developmental process of academic interest. Descriptive analysis confirms high adoption rates, with 74.5% of students reporting regular integration of ChatGPT into their learning routines. Crucially, simple linear regression reveals a significant positive effect (β = 0.42, p < 0.001, t = 6.354), accounting for 17.6% of the variance in self-reported interest scores. Qualitative findings unpack this relationship, identifying three key mediators: (1) cognitive offloading, which frees mental resources for engagement with complex concepts; (2) dialogic scaffolding, where iterative AI conversations foster personalized relevance; and (3) epistemological negotiation, as students critically reconcile AI-generated content with social science paradigms. The study contributes to the Generative Interest Development Theory, arguing that GAI, when pedagogically framed, can act as a catalyst for situational and emerging individual interest. It concludes with a framework for critical-AI pedagogy that leverages these tools to enhance, rather than undermine, deep disciplinary engagement in higher education.