Examining the impact of large language models on design: Functions, strengths, limitations, and roles

Abstract

Large Language Models (LLMs) are advancing design activities with human-like generative capabilities and context awareness, holding more significant potential compared to the previous discriminative AI. However, existing research frameworks studying the interaction between non-LLM AI and designers have yet to fully address the evolving dynamics and generative capacities introduced by LLMs, leaving a critical gap in understanding LLMs’ impact on design. Through a systematic literature review of 118 publications, this study investigates how LLMs support the entire design process—Discover, Define, Develop, and Deliver—with various functions including generating survey questionnaires and synthesizing evaluations, while revealing both the strengths and limitations as design tools or materials. Besides, we examine how LLMs influence the competencies of designers and discuss the potential for LLMs to augment or even transform the roles of the designer. This research highlights opportunities for future studies to enable LLMs to complement designers rather than replace them.

Topics & Concepts