Generative AI for Fashion Innovation: A Systematic Literature Review of Models and Applications

Generative Artificial Intelligence (Generative AI) has become a transformative tool in the fashion industry, enabling new approaches to creativity, personalization, and innovation. Core generative models, such as Generative Adversarial Networks (GANs), Variational Auto Encoder (VAEs), and Diffusion Models are increasingly applied to support fashion-related tasks such as garment and textile design, outfit generation, and virtual try-on. Despite this progress, there has been limited systematic examination of how these models are utilized and what applications they enable within fashion innovation. In this work, we describe a Systematic Literature Review (SLR) focused on mapping generative models and their applications in the fashion industry. Following the PRISMA methodology, relevant publications were retrieved from IEEE Xplore, Scopus, and Science Direct databases. Our analysis highlights the dominance of GANs in early studies, the recent adoption of diffusion models for high-quality generation, and emerging uses of LLMs for multimodal fashion design. Applications are clustered into three domains: creative design, virtual try-on, recommendation systems and trend forecasting. This review identifies key challenges including dataset limitations, computational costs, and intellectual property issues, while also outlining opportunities for sustainable and personalized fashion innovation. The study provides a structured reference for researchers and practitioners seeking to leverage generative AI in advancing the future of fashion through scalable solutions and broader interdisciplinary innovation across emerging digital ecosystems that continuously evolve, adapt, and respond to dynamic consumer preferences and global industry trends.