Designing Context-Aware Artificial Intelligence Agents for Computer Science Education

• In Event: AERA Roundtable Session Thursday 2:15 pm - Gold 4
  In Roundtable Session: AI-Enabled Curriculum and Pathways in Computing and Beyond

Abstract

This study investigates how to design culturally relevant and context-aware AI agents for CS education. We conducted a mixed-methods study with 12 students and teaching assistants from a culturally diverse region, employing computational grounded theory and qualitative inquiry on collected rich data. Our findings identified eight key super-themes, including the need for adaptive feedback, culturally relevant content (e.g., place-based learning), and a careful balance to prevent over-reliance on AI. Participants emphasized that useful AI agents must emulate qualities of human mentors, like patience and adaptive support. This study offers informative design heuristics for building AI systems that are not just technically proficient but also equitable, context-aware, and emotionally attuned to the diverse needs of learners for personally meaningful learning experience.

Authors

• Chih-Pu Dai, Texas A&M University

• Azibun Nuder, University of Hawaii - Manoa