#193: Exploring Shared Agency in an Intelligent Immersive Learning Environment for Collaborative Writing
An increasing number of studies on AI-student collaboration highlight the importance of learner agency in the Generative AI (GenAI)-powered learning environment. However, limited research has examined how student-AI shared agency can be assessed, and it remains unclear whether individual interaction with AI agents or peer collaboration with AI agents is more effective in developing such agency. This study explores second language (L2) students’ shared agency within an intelligent immersive learning environment for writing. Using WriteCube, a self-developed AR- and AI-enabled writing system, 75 Primary 4 students from two classes in Singapore completed picture-based composition tasks under individual-AI and pair-AI collaboration conditions. Findings revealed no significant performance difference between groups, suggesting a trade-off between interactional richness and task efficiency. Regression results indicated that interaction quality, rather than frequency or language proficiency, positively predicted shared agency, particularly in ensuring equity. The study offers implications for assessing and enhancing learner agency in GenAI-supported L2 classrooms.
Speakers
- Yun Wen — National Institute of Education, Nanyang Technological University, Singapore
- Zongran Wang — Department of Education Information Technology, East China Normal University
Authors
Yun Wen, Zongran Wang, Lang Xu