Bridging and Reconstructing: Designing a Programming AI Agent to Optimize Learners’ Experience and Develop Programming Self-Concept

• In Event: AERA Poster Session Wednesday 3:45pm
  In Poster Session: Modeling, Measurement, and Design in Technology-Enhanced Learning Environments

Abstract

Generative artificial intelligence (GenAI) agents emerge as new programming learning tools, enhancing learning experiences and motivation. Our study designed and developed a Programming AI Agent with anthropomorphic characteristics and knowledge base support. We used a “situational experience & user feedback” approach and drew on 463 responses in structural equation modeling. We found: (1) Personal privacy risk negatively affected human-computer trust. (2) Anthropomorphism and system quality positively impacted perceived value. (3) Human-computer trust and perceived value positively predicted learning motivation and self-concept. (4) GenAI literacy predicted learning motivation and self-concept and significantly moderated the effects of human-computer trust and perceived value. The results theoretically inspire deploying AI agent-assisted programming education and practically offer insights for intelligent learning system optimization.

Authors

• Haoming Wang, East China Normal University

• Chengliang Wang, Australian Catholic University

• Jinhua Shao, Anhui University of Science & Technology

• Xianru Shang, Shaanxi University of Science and Technology

• Teng Yu, Guangdong Polytechnic Normal University

• Yupeng Lin, Tsinghua University