Abstract
Artificial intelligence (AI)-generated creative exemplars are increasingly used as scalable digital instructional resources in creativity education. However, it remains unclear whether labeling identical exemplars as human- or AI-generated shapes how learners engage in group creative activities. We tested this possibility in 113 dyads using an fNIRS hyperscanning group creative ideation paradigm. Dyads completed creative ideation tasks before, during, and after exemplar-stimulation. During stimulation, the human- and AI-attribution groups received identical exemplars labeled as produced by human or AI, whereas the control group received no exemplar. Human attribution increased idea uniqueness during stimulation relative to both the other phases and the AI-attribution and control groups. AI attribution temporarily increased switching among ideation pathways. Neither behavioral effect persisted after exemplar withdrawal. The two attribution groups did not differ in exemplar-category overlap or nearest-exemplar semantic distance, but human-attribution dyads rated the exemplars as more helpful. Human attribution also increased interbrain synchronization (IBS) in bilateral premotor and supplementary motor cortices during stimulation, whereas AI attribution was associated with decreased IBS in the left supramarginal gyrus and right frontopolar cortex from pre- to post-stimulation. Thus, exemplar-source attribution shapes the process and outcome of group creative ideation, revealing an important social-contextual factor in AI-supported creativity education.