Unpacking Human–GenAI Collaboration in Software Engineering Education
TL;DR
This research unpacks the problem-solving strategies, collaboration patterns, and agency shifts underlying human–GenAI collaboration in software engineering education using a mixed-methods approach.
Abstract
Generative AI tools such as ChatGPT, Claude, and Gemini are increasingly shaping how students engage with software engineering (SE) problem-solving. However, there is limited understanding of how learners collaborate with GenAI across different stages of the software development lifecycle (SDLC), and how learner agency shifts during this process. This research unpacks the problem-solving strategies, collaboration patterns, and agency shifts underlying human–GenAI collaboration in software engineering education. Using a mixed-methods approach, it draws on multimodal data including screen recordings, GenAI transcripts, task artefacts, and retrospective think-aloud interviews. Preliminary findings suggest that learners engage with GenAI differently across SDLC phases, with variation in reliance, evaluation, and decision-making. Future work will extend this analysis to larger and more diverse learner groups and inform pedagogical supports for effective, critical, and agentic GenAI-supported problem-solving.