Mapping Research on Generative Artificial Intelligence for Programming Education: A Systematic Review

Abstract

ABSTRACT Background Generative artificial intelligence (AI) has the potential to transform programming education, yet existing research on its integration and impact remains fragmented, underscoring the need for a comprehensive synthesis. Objectives This systematic review aims to consolidate empirical evidence on the use of generative AI in programming education within higher education, examining research objectives, AI tools, instructional strategies, programming languages and publication trends. Methods Following PRISMA guidelines, 46 peer‐reviewed empirical studies were analysed to identify patterns in tool usage, pedagogical approaches and research dissemination. Results and Conclusion ChatGPT dominated the literature, featuring in 76% of studies, while Java and Python were the most frequently targeted programming languages. Research spanned 28 countries, with nearly 70% of studies published in conference proceedings, indicating the exploratory nature of the field. Most studies investigated student‐level outcomes, including learning performance and engagement, and examined generative AI as a pedagogical aid. However, more than half lacked explicit pedagogical frameworks. Reported benefits included enhanced engagement, feedback, and learning support, but concerns emerged regarding over‐reliance, reduced conceptual understanding, and threats to academic integrity. Gaps were identified in research on instructors' roles, institutional practices, and comparative evaluation of AI tools. This review highlights both the promise and challenges of generative AI in programming education. It emphasises the need for thoughtful instructional design, ethical guidelines and rigorous evaluation. By synthesising current evidence, the study provides a foundation to guide future research, pedagogical innovation and policy development in the evolving landscape of AI‐driven programming education.