The rapid adoption of generative AI (GAI) tools like ChatGPT has outpaced empirical research on their instructional value, particularly in computer science education (CSE), where learning relies heavily on debugging, problem-solving, and iterative reasoning. To clarify current knowledge and identify areas with limited evidence, this study conducted a rapid review of peer-reviewed empirical work published between 2022 and mid-2024, following the public release of ChatGPT. A systematic search of the EBSCO platform identified a small but growing body of studies where learners or instructors directly used GAI for programming or foundational mathematics.

Across the included studies, GAI tools served two main functions: providing stepwise hints for structured problem-solving and acting as coding companions that supported debugging. In programming contexts, students using ChatGPT engaged in more frequent cycles of diagnosis, revision, and feedback, though immediate performance gains were limited. In mathematics, AI-generated hints produced learning gains comparable to human-authored help when accuracy safeguards were applied. These results indicate that GAI is most effective when tasks are well-structured and feedback aligns closely with target skills.

Stakeholder perspectives reveal both enthusiasm and caution. Students appreciated the immediacy and accessibility of AI support, while instructors expressed concerns about over-reliance, reduced cognitive struggle, and academic integrity. Overall, the findings highlight the need for principled GAI integration that emphasizes transparency, process evidence, and reflective engagement.

This review provides an early synthesis of emerging evidence and identifies foundational questions for future research as educational systems begin to formalize the pedagogical role of GAI in CSE.

Authors:
Michael Pin-Chuan Lin, Mount Saint Vincent University, Canada
Gaganpreet Jhajj, Athabasca University, Canada
Fuhua Lin, Athabasca University, Canada
Eric Poitras, Dalhousie University, Canada
Daniel Chang, Simon Fraser University, Canada
Jeeho Ryoo, Fairleigh Dickinson University, Canada

About the Presenter(s)
Dr. Michael Lin is an assistant professor in the Faculty of Education at Mount Saint Vincent University (MSVU). He is a university educator and interdisciplinary researcher in educational technology and learning design. His research examines how emer

See this presentation on the full schedule – Saturday Schedule

Bookmark