Beyond Prompting: Empowering Student Self-Regulation in Generative AI Writing Through a Learning Analytics Dashboard

• In Event: AERA Roundtable Session Friday 11:45 am - Gold 4
  In Roundtable Session: Self-Regulation and Emotional Development in the age of AI

Abstract

This study investigates if a theory-driven learning analytics dashboard (LAD) can mitigate risks of student over-reliance on Generative AI (GenAI) in academic writing. In a quasi-experiment with 52 postgraduates, an experimental group used a LAD based on Zimmerman's self-regulated learning (SRL) framework, while a control group did not. The LAD group showed significantly higher writing knowledge gains and improved SRL skills, such as self-efficacy and cognitive strategy use. However, they also reported increased test anxiety and cognitive load. Dialogue analysis revealed the dashboard group engaged in more reflective and evaluative interactions with GenAI, unlike the control group’s transactional exchanges. Findings suggest that structured interventions can enhance learning in human-AI collaboration.

Author

• Angxuan Chen, Peking University