Using ChatGPT to extract design concepts from stories 

In this paper, I devised and embedded textual Generative Artificial Intelligence (GenAI) into a design approach to digitally translate stories into design concepts. This approach allows design students to brainstorm ideas variedly and quickly based on a story and scope out design opportunities. Currently, design students manually conceptualise ideas from stories by reinterpreting the story, abstracting key ideas from the story, and using metaphors and allegories to extrapolate design narratives from the given story. While visual GenAI software has sparked great design and research interest in its ability to visualise ideas rapidly, textual GenAI software has received less attention in the design discipline. This design-led research presents my textual GenAI-embedded design approach for students to digitally generate design ideas from stories that are then developed into installation designs. The outcomes show that integrating textual GenAI into the design workflow affords a range of allegories from the same story for design students to translate into different design concepts. Nonetheless, design students must still realise these digital interpretations as visual designs and then into installation proposals. Based on the outcomes, design students can use the proposed textual GenAI embedded design workflow to generate different allegories and then choose one for design development.

keywords: Generative AI; ChatGPT; concept design; allegorical architecture

• year: September 2023 – January 2024
• for: LearnxDesign 2023 Conference, Design Research Society
• download: Article