Visualizing gamer identity through text-to-image AI: algorithmic stigma in Chinese and U.S. generative systems

Introduction This article examines AI-generated images as symbolic products, and explores how these images visually represent social identities. This study takes the case of gamer identity as the research subject. Methods The research analyzed 1,200 encoded images from text-to-image generative AI tools in China and the United States. Based on the theory of visual social semiotics, the study examined how image arrangements such as physical features and spatial layout jointly shape the image of “gamers”. Results The research found that the AI-generated images of gamers mainly present in a narrow way, focusing on young groups, masculine traits, private space, screen immersion, and technological devices. Three recurring visual templates are particularly prominent: socially withdrawn gamers, professional e-sports champions, and casual gamers. Female gamers were more often represented through softer aesthetic styles, decorative environments, and social forms of gaming, and were less frequently depicted in technically intensive or competitive roles. Comparative analysis further reveals different representational tendencies: China-based tools more often associate games with fatigue and chaos, while U.S.-based tools emphasize skills, style, and competitive value. Discussion In this study, these patterns are conceptualized as “algorithmic stigma”, illustrating how text-to-image AI can organize the visual representation and recognisability of social identities through recurring stereotypical configurations.