Human-AI Collaboration in the Fashion Design Process

Abstract

The recent development of Generative AI (GenAI) revolutionized the fashion industry, automating tasks like market analysis and trend forecasting, as well as innovating cloth design. However, its potential application in supporting designers in the fashion design process for creativity purposes has yet to be explored. Our research studied the complexity and pitfalls of such a process through interviews with domain experts. Starting from that assessment, this article compares generic and context-specific GenAI tools from several perspectives to explore their features and capabilities in enhancing the cooperation between fashion designers and AI tools and agents in the creative process. The most promising systems are tested on typical creative tasks, including generating images from textual descriptions (specified according to technical criteria), generating variations of existing designs and pictures, and generating photographic images from sketches. The experiments involved two data sets and five domain-specific tasks. The effectiveness and limitations of the tools are evaluated quantitatively to assess how they can properly support the designers' work.