Teachers-Generative AI Collaboration in Developing Learning Tasks: Mapping the Challenges

• In Event: AERA Poster Session Sunday 11:45 am
  In Poster Session: Technology, Instruction, Cognition, and Learning SIG Poster Session

Abstract

There is a lack of understanding of how teachers engage with generative artificial intelligence (GenAI) for learning task design and the nuanced challenges inherent in such collaboration. Therefore, this mixed-methods study involved 28 in-service teachers collaborating with GenAI tools to design learning tasks over three weeks. Results identified teacher-side challenges (insufficient input of context information, insufficient pedagogical guidance, and insufficient skills with GenAI), GenAI-side challenges (accuracy issues, limited prompt understanding, limited instructional arcs, insufficient knowledge of context, incomplete instructional solutions, and usability issues), as well as four overall interaction challenges (inefficient interaction, limited quality, lack of distributed cognition, negative human-GenAI loops). This study provides empirical evidence of challenges and a framework for enhancing effective teacher-GenAI collaboration.

Authors

• Jie Cao, University of North Carolina - Chapel Hill

• Shuman Wang, Stanford University

• Christian D. Schunn, University of Pittsburgh