IMPACT OF GITHUB COPILOT USAGE ON PROGRAMMING PRODUCTIVITY AMONG UNDERGRADUATE COMPUTER SCIENCE STUDENTS

DOI: https://doie.org/10.65985/APER.2026632886

Authors:

Mr. Ajay Kumar Gupta

Keywords:

AI-assisted programming, GitHub Copilot, Computer science education, Programming Conceptual Technology acceptance, Skill augmentation

Abstract:

The integration of AI-powered coding assistants into undergraduate computer science education has generated substantial debate regarding their pedagogical impact. While proponents emphasize productivity gains, concerns remain about potential reductions in conceptual understanding. This study empirically examines the effects of AI-assisted coding on programming performance among undergraduate computer science students. Using a controlled experimental design with 300 participants, students were randomly assigned to either an AI-assisted condition with access to GitHub Copilot or a control group without AI support. Performance outcomes included task completion time, code quality, debugging errors, programming confidence, and conceptual understanding measured through a post-task theoretical assessment. Results indicate that AI assistance significantly reduces task completion time and debugging errors while moderately improving code quality. Students using AI tools also report higher programming confidence. However, findings reveal a small but statistically significant decline in conceptual understanding among AI-assisted participants. Moderation analysis further shows that productivity gains are more pronounced among students with lower prior programming experience. The results suggest that AI coding assistants function primarily as productivity enhancers with modest cognitive trade offs. These findings inform curriculum design, assessment practices, and institutional AI governance strategies within computer science education.

PDF