#169: Supporting Students’ Shared Epistemic Agency Discourse with GAI via Learning Analytics Augmented Meta-discourse

This study investigates the effectiveness of a learning analytics-augmented meta- discourse environment in fostering students shared epistemic agency discourse in a Generative AI (GAI) learning context. Two types of learning analytical tools and meta-discourse were utilized to support students ’ epistemic understanding of critical discourse, thus, to support students’ epistemic agency discourse. A total of 48 project management students participated in the 18-week course, engaging in four online discussions and two meta-discourse sessions. Findings show that the intervention significantly improved students' epistemic understanding of critical AI discourse. Furthermore, students demonstrated progress in their shared epistemic agency discourse, moving from information sharing to more projective and regulative activities across the three phases of discussion. Students with high epistemic understanding showed a stronger tendency to build on, evaluate AI information, and create joint action plans, in contrast to low-understanding students who primarily used AI information for idea negotiation. These results underscore the critical role of pedagogical design in utilizing GAI as an epistemic tool to support collaborative knowledge advancement.