Supporting Designers’ Cultural Cognition and Creative Transformation with LLMs: A Comparative Study in Cultural Product Design

Abstract

Abstract

In the early stage of cultural product design, designers need to understand relevant cultural knowledge in depth before transforming it into product concepts. However, search engines (SEs)-based cultural surveys often make it difficult for designers to efficiently identify, filter, and translate cultural information into design-relevant ideas. Large language models (LLMs), with conversational interaction and structured information generation, provide new possibilities for supporting designers’ cultural cognition and design ideation. This study took Xunxian Mud Goo as the survey object and introduced LLMs as an intelligent collaborative tool to compare their application differences with SEs in cultural surveys. A comparative experiment involving design students was conducted to evaluate the two tools in terms of cultural cognition, design inspiration, survey efficiency, and tool preference. The results indicated that LLM-supported interaction improved designers’ efficiency in cultural cognition acquisition, reduced search-related behavioral costs, and provided design guidance that helped transform cultural information into initial design ideas. These findings suggest that LLMs can optimize designers’ cultural survey processes through structured information and intelligent interaction, providing new perspectives and methodological support for the survey phase of cultural product design.