Generative AI-Mediated Learning Trajectories: An Activity Theory Analysis of In-Service CS Educators’ Engagement with ChatGPT

• In Event: AERA Roundtable Session Wednesday 11:45 am - Gold 3
  In Roundtable Session: Educators Navigating AI: Identity, Practice, and Pedagogy

Abstract

This study examines how 17 in-service teachers in a graduate CS Education program used ChatGPT to deepen their understanding of foundational concepts and knowledge application in their teaching contexts. Grounded in Activity Theory, we analyzed 972 anonymized prompts from nine scaffolded assignments using K-means clustering and thematic coding. Four engagement patterns emerged, revealing that targeted prompt-engineering guidance may have potential to boost interaction among less confident users and help frequent users sustain focus. In-service teachers without CS backgrounds require both content support and AI-literacy scaffolds. These findings highlight the need to embed domain-aligned AI-literacy instruction that combines precise prompting techniques with concept-driven inquiry into online CS teacher education, enhancing both content mastery and effective use of generative AI tools.

Authors

• Rui Tammy Huang, University of Florida

• Ruohan Liu, Seattle University

• John Hampton, University of Florida