GenAI-supported project-based product design learning: differential changes in multidimensional learning outcomes

Abstract

Generative artificial intelligence (GenAI) is increasingly used in design education, yet evidence remains limited on whether its educational value derives from tool access itself or from how human–AI interaction is structured within project work. This quasi-experimental pre-test–post-test study compared a 4-week GenAI-Integrated Project-Based Learning Module (GIPLM) with traditional project-based learning (PjBL) in an undergraduate product design course (N = 60). Rather than treating GenAI as a stand-alone content-generation tool, GIPLM embedded GenAI as process scaffolding across four stages: product and requirement analysis, user research and knowledge structuring, ideation with multimodal visualization and 3D prototyping, and feedback, validation, and reflective presentation. Learning outcomes were assessed across cognitive, affective, and psychomotor-related domains, together with course performance. According to a mixed-design ANOVA, the GenAI-supported group was associated with significantly greater gains than the control group in cognitive understanding, problem-solving and creativity, fundamental tool operation, advanced design application, and final academic performance. Because advanced design application differed at baseline, ANCOVA was additionally conducted; the post-test group effect remained significant after controlling for pre-test scores, F(1, 57) = 20.519, p < 0.001, partial η2 = 0.265. Engagement and participation and ethical and user-centered responsibility improved over time but showed no significant Time × Group interactions. These findings suggest that the observed advantages are more consistent with structured, workflow-embedded GenAI scaffolding than with AI access alone, with stronger short-term effects on cognitive, procedural, and performance outcomes than on affective development.