A Theoretical Framework of Student Agency in AI- Assisted Learning: A Grounded Theory Approach

Abstract

Generative AI(GenAI) is a kind of AI model capable of producing human-like content in various modalities, including text, image, audio, video, and computer programming. Although GenAI offers great potential for education, its value often depends on students' ability to engage with it actively, responsibly, and critically - qualities central to student agency. Nevertheless, student agency has long been a complex and ambiguous concept in educational discourses, with few empirical studies clarifying its distinct nature and process in AI-assisted learning environments. To address this gap, the qualitative study presented in this article examines how higher education students exercise agency in AI-assisted learning and proposes a theoretical framework using a grounded theory approach. Guided by agentic engagement theory, this article analyzes the authentic experiences of 26 students using data from their GenAI conversation records and cognitive interviews that capture their thought processes and decision-making. The findings identify four key aspects of student agency: initiating and (re)directing, mindful adoption, external help-seeking, and reflective learning. Together, these aspects form an empirically developed framework that characterizes student agency in AI-assisted learning as a proactive, intentional, adaptive, reflective, and iterative process. Based on the empirical findings, theoretical and practical implications are discussed for researchers, educators, and policymakers.