Generative AI (genAI) has exploded in the 2020’s, well-known for text and image generation, but little known for its applications as a revolutionary technology for programming including software development and debugging, which is an essential professional skill outside of computer science. This systematic literature review investigates primary research on generative AI (genAI) tools for programming education in university settings. Following the PRISMA framework, we examined over 100,000 sources from major academic repositories, narrowing to peer-reviewed experimental studies conducted with tertiary students. The stringent exclusion criteria retained approximately 10% of sources, which were thematically analyzed and synthesised to form recommendations for teaching and learning at UQ. Results indicate that experimental research on genAI-assisted programming is predominantly concentrated in computer science education. While the majority of studies report positive impacts on learning outcomes and student engagement, notable limitations emerged in specific contexts. Particularly, students with limited metacognitive awareness of programming processes experienced reduced benefits from genAI tools. These findings highlight the importance of implementation strategies that account for varying student capabilities and learning approaches. This review addresses a critical gap in the literature by distinguishing empirical primary research from the abundance of opinion pieces and theoretical reviews on genAI in education. By focusing exclusively on primary data from student-centered experiments, we provide evidence-based insights into the practical applications and limitations of genAI for programming instruction across disciplines. The findings inform our current research at the University of Queensland investigating genAI-assisted programming for cross-disciplinary applications including in Engineering, Science, and Humanities curricula.

Authors:
Lachlan Miller, University of Queensland, Australia
Aneesha Bakharia, University of Queensland, Australia
Felix Eggers, University of Queensland, Australia

About the Presenter(s)
Lachlan Miller is currently a senior research assistant on a Teaching and Innovation Grant for cross-disciplinary genAI-assisted programming at the University of Queensland. He is interested in increasing accessibility of STEM education.

See this presentation on the full schedule – On Demand Schedule

Bookmark