AI Insight
This study analyzed 1,200 AI-generated images of gamers from Chinese and U.S. text-to-image systems, revealing systematic stereotypical patterns in how these algorithms visually represent gamer identity. The research identified three dominant visual templates: socially withdrawn gamers, professional e-sports champions, and casual gamers, with representations skewing heavily toward young, masculine individuals in private spaces. Comparative analysis showed that Chinese AI tools more frequently associated gaming with fatigue and disorder, while U.S. tools emphasized skill, style, and competitive aspects.
Why it matters
The findings demonstrate that generative AI systems can encode and perpetuate social stereotypes through what the researchers term "algorithmic stigma," which has implications for how emerging technologies may reinforce or amplify existing biases about social identities. This is particularly relevant as text-to-image AI tools become increasingly integrated into media production, education, and communication platforms.
Understand the Science
IntroductionThis 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.MethodsThe 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”.ResultsThe 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 Three recurring visual templates are particularly prominent: socially withdrawn gamers, professional e-sports champions, and casual gamers.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.DiscussionIn 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.