<![CDATA[Generative artificial intelligence is increasingly embedded in industrial design, yet its effects on divergent thinking remain contested because rapid idea production may coexist with fixation, reduced agency, and uncertain authorship. This study examined how generative tools reshape fluency, flexibility, originality, and elaboration during early design ideation while exploring designers’ lived experiences of human–AI co-creation. A mixed-methods design combined a randomized pretest–posttest experiment with phenomenological interviews. One hundred twenty industrial design students were assigned to AI-assisted or conventional ideation conditions, while twenty participants were purposively selected for in-depth interviews. Divergent-thinking performance, process logs, questionnaires, sketches, prompts, and interview transcripts were analyzed using ANCOVA, correlation analysis, and thematic interpretation. The AI-assisted group achieved significantly higher posttest scores across all divergent-thinking dimensions, with the strongest effects in fluency and flexibility. Conceptual reframing and prompt revision were positively associated with originality, whereas direct adoption of generated outputs was linked to fixation, lower autonomy, and authorship uncertainty. Participants described generative AI as a source of accelerated externalization, productive surprise, expanded association, and increased evaluative burden. Generative AI functions as a conditional creativity amplifier whose benefits depend on critical human engagement, active transformation, and reflective control rather than technological access alone.]]>
