AI Co-Tutors and Instructor Adaptation in Online Classrooms: Evolving Roles and Teaching Practices

• In Event: Artificial Intelligence in Online Teaching and Learning

Abstract

Teacher knowledge and experience with generative artificial intelligence (AI) are critical when evaluating the instructional impact of large language models (LLMs) in online learning. This study investigated how a GPT-4-based AI co-tutor influenced instructional practice and student engagement in a three-week asynchronous science course for academically advanced learners. The instructor facilitated two iterations of the course: one that included AI integration and one without such support. Data sources included student work, chat transcripts, and a teacher interview. The presence of the AI shifted the teacher’s feedback practices and appeared to support student reasoning by offering an additional channel for explanation and reflection. Findings highlight teacher adaptation as key to leveraging AI’s instructional value and inform future AI-integrated course design.

Authors

• Kimberley L. Chandler, Johns Hopkins University

• Kathryn N. Thompson, Johns Hopkins Center for Talented Youth