Creative combinational design through generative AI in different dimensional representations: An exploration
Boheng Wang, Ji Han, Xiaoyang Zhao, Yuan Yin, Liuqing Chen, Peter Childs
Abstract
Although generative AI models are increasingly integrated into creative work, there is limited evidence on how dimensional representations of generative AI—text, image, and 3D—impact the creative process. Identifying these effects aids researchers in selecting suitable models for developing generative design. Through two empirical studies, this research examines the capacity of generative models with varying dimensions to support combinational ideation and influence the creative design process. The results indicate that generative models in different dimensions demonstrate varying levels of creative combinational ideation. Generative AI supports combinational ideation by providing more ideas in divergent stages rather than helping decision-making in convergent stages. These results help inform AI design developers in choosing models and integrating models for combination.