Generative AI in Clothes Design: A Scoping Review of Workflows, Challenges, and Future Pathways

Abstract

The clothes design and development sector is under growing pressure to accelerate workflows as product lifecycles shrink. Generative AI (GenAI), driven by diffusion models, Generative Adversarial Networks (GANs), and Large Language Models (LLMs), is reshaping this process, yet existing studies often examine individual tools in isolation, overemphasise 2D visual outputs, and largely overlook real-world production. Following PRISMA-ScR, this scoping review examines 57 peer-reviewed articles on AI-assisted clothes design from Web of Science and Scopus (2021–2026), spanning computer science, HCI, textile engineering, and design. Studies cluster in early stages: ideation (29.8%) and sketch rendering (35.1%) make up nearly two-thirds, while physical-engineering stages combined (pattern-making and fabric visualization) account for only 24.5%, with pattern-making and structural design alone at 10.5%. This imbalance reveals a visual-engineering disconnect: AI still struggles to produce production-ready structures. To synthesize these insights, the review proposes a comprehensive diagram identifying four systemic challenges, future pathways, and stage-specific design implications.

Keywords: Generative AI, Clothes Design, Scoping Review, Human-AI Collaboration, Design Workflow

DOI: 10.54941/ahfe1007814