AI-assisted divergent pedagogy: Understanding human-AI co-creativity in design thinking education

This research examines how generative artificial intelligence (AI) participates in students' judgement practices within a design thinking class. Drawing on 35 design portfolios from a third-year interaction design course in New Zealand, the study examines students’ engagement with generative tools across the stages of design thinking. Through thematic and cross-case analysis, the research identifies a recurrent feedback-driven generative loop in which generation, comparison, and revision form the basic unit of activity. Rather than functioning as an external aid, AI is documented in the portfolios as a form of generative mediation that participates in how generation, comparison, and revision unfold across the design process. Empirical analysis indicates multiple modes through which generative feedback is associated with problem framing, divergent trajectories, strategic adjustment, and boundary negotiation. Students’ practices demonstrate how iterative prompting, cross-platform orchestration, and reflective correction situate AI within evolving judgement cycles. Building on theories of design cognition and distributed cognition, the paper proposes the framework of AI-assisted Divergent Pedagogy as a process-oriented interpretive approach for examining how generative feedback repeatedly enters design processes under specific pedagogical conditions. The contribution lies not in classifying AI functions, but in identifying a recurrent feedback pattern through which aspects of judgement become visible in iterative processes of generation, comparison, and revision. These findings contribute a context-bound interpretive account of how generative systems may inform reflective processes in classroom-embedded design learning. • Examines processual dynamics of human-AI co-creativity in design thinking. • Analyses 35 student portfolios across five stages of design thinking. • Identifies four analytical dimensions of generative feedback and recurrent strategic orientations. • Proposes an interpretive framework of AI-assisted Divergent Pedagogy. • How generative feedback becomes involved in reflective judgement in design learning.