From deep learning to generative AI: research status and prospects of product appearance design in the artificial intelligence era

In the experience economy, product appearance design serves as a vital medium for conveying needs and expressing aesthetic intent. As user demands become increasingly diverse, product appearance design (PAD) has emerged as a primary factor influencing consumer purchasing behaviour. Artificial Intelligence (AI) has catalysed substantial transformations within the manufacturing industry, progressively establishing AI as an essential enabling technology for PAD. Unlike conventional PAD, AI-enabled PAD introduces distinctive capabilities, including dynamic information sharing, interactive design solutions and gradual design goals of aesthetics, style and visual innovation. As a relatively new design paradigm, many critical issues within AI-enabled PAD remain underexplored. This study systematically reviewed the literature on AI-enabled PAD to understand its evolution and state-of-the-art. It offers a comprehensive overview of the key aspects of AI-enabled PAD research status, encompassing design modes, defining characteristics, key elements and implementation frameworks. Additionally, it critically evaluates the key technology of PAD across three dimensions: application mechanisms in the design phase, intelligence capability and intelligent tools. Building on this analysis, the study discusses the limitations of current AI-enabled PAD across three design characteristics and proposes potential research directions at last. It is hoped that this work can provide useful insights to both academics and industry.