TLDR: Researchers at Pusan National University have successfully leveraged generative AI, specifically ChatGPT and DALL-E 3, to predict trends and create a 2024 Fall/Winter men’s fashion collection. The study highlights AI’s potential to enhance efficiency and creativity in fashion design, while also underscoring the critical role of human expertise in prompt engineering.
BUSAN, South Korea – A groundbreaking study from Pusan National University, led by Professor Yoon Kyung Lee and a Master’s student, has demonstrated the significant potential of generative artificial intelligence (AI) in streamlining and enhancing the fashion design process. The research focused on creating a 2024 Fall/Winter men’s fashion collection, showcasing how AI models can predict trends and generate realistic design images.
The study utilized a two-pronged AI approach: ChatGPT (versions 3.5 and 4) for trend analysis and prediction, and DALL-E 3 for image generation. Researchers first analyzed historical men’s fashion data up to September 2021 using ChatGPT to identify past styles and then predict future trends for Fall/Winter 2024. These predicted trends formed ‘initial codes’ for the design process.
Further enriching the data, design elements from Vogue’s 2024 Fall/Winter Men’s Fashion Trend data were incorporated as ‘modified codes,’ alongside ‘codes from literature’ on fashion design concepts. These diverse elements were then meticulously regrouped into six core categories: trends, silhouette elements, materials, key items, garment details, and embellishments.
Using these refined codes, the team crafted 35 unique prompts for DALL-E 3, each describing a distinct outfit. The prompts adhered to a consistent template, envisioning a male model on a runway at a 2024 Fall/Winter fashion show, with customizable details such as aspect ratios, camera angles, model appearance, runway design, background, and mood. Each prompt was executed three times, resulting in a total of 105 generated images.
The results were promising, with DALL-E 3 successfully implementing the prompts 67.6% of the time. Notably, prompts containing adjectives showed a higher rate of accurate implementation. Some of the generated images bore striking resemblances to actual 2024 Fall/Winter Men’s fashion collections, underscoring AI’s capability to produce relevant and aesthetically pleasing designs.
However, the study also identified limitations. Many generated images leaned towards ready-to-wear fashion, and DALL-E struggled to incorporate complex trend elements like gender fluidity. The researchers found that simple trend keywords were insufficient for accurate results, highlighting the necessity for ‘expertly worded prompts’ from seasoned fashion designers. Professor Lee remarked, “Our results show that expertly worded prompts are necessary for accurate fashion design implementation of generative AI, highlighting the important role of fashion experts.”
Also Read:
- Pusan National University Researchers Advance Fashion Design with Generative AI
- Generative AI Poised for Explosive Growth in Consumer Packaged Goods Sector, Projected to Hit $18.9 Billion by 2033
This research, published online in the Clothing and Textiles Research Journal on June 22, 2025, suggests that generative AI has the potential to revolutionize the fashion industry. By recognizing patterns and generating new content, AI models powered by deep learning algorithms can significantly expand designers’ creativity and accelerate product development, bringing collections to market faster. The study reinforces that while AI offers powerful tools, human expertise remains crucial for guiding and refining its creative output.
