Challenges and opportunities in using ChatGPT as a team member to promote code review education and self-regulated learning

Abstract

In this study, we examine the use of ChatGPT as a team member to enhance code review education and promote self-regulated learning among programmers of varying experience levels. A case-study approach was used to answer three research questions about code quality, perceived benefits, and challenges. Participants reported diverse perspectives on the quality and utility of ChatGPT-generated code, with novice programmers gaining a deeper understanding of code interactions and experienced programmers refining their code through iterative prompt optimisation. Despite the benefits, challenges emerged, such as concerns about the potential negative impact of generative AI on learning experiences and the difficulties posed by foundational knowledge gaps in effectively interacting with ChatGPT. The findings suggest that ChatGPT is a promising tool for supporting code review education, especially for novice programmers, and for fostering self-regulated learning strategies like elaboration and metacognition. Future research should validate these results with larger sample sizes and address the identified challenges to better integrate generative AI tools in educational contexts.