How the Field of Fashion can use ChatGPT to Predict Fashion Trends

Abstract

Fashion trend prediction is based on intuition, judgment, and creativity and requires a sensual and intuitive approach, but can also be based on a highly scientific analysis method that uses a systematic method. This study aims to develop prompts that can verify and improve the accuracy of fashion trend prediction technology by analysing generative data from open AI models that are currently widely used. To this end, we propose ChatGPT as a means of fashion trend prediction and a method to derive efficient and objective data through human-AI interaction. We identified the characteristics of ChatGPT in the field of fashion trend prediction, set prompt guidelines, and developed a top-down prompt (TDP) using Lotus Blossom. Additionally, we applied TDP to ChatGPT to predict men's fashion trends for fall/winter 2024 and validated it using data from real fashion trend companies.