Social Dimensions in Content Creation for Games: A Systematic Review of Fairness, Social Bias, and Diversity in Computational Character and Environment Generation

Computational content generation increasingly shapes digital games, visual computing, and interactive environments, automating the creation of characters and environments. However, these systems also raise complex challenges related to fairness, social bias, and representational diversity. We present a systematic literature review of 37 recent studies addressing fairness-aware computational content generation across character creation, face modeling, and procedural environments. We analyze both algorithmic strategies-including deep learning, reinforcement learning, generative adversarial networks, and optimizationas well as design-oriented frameworks such as behavioral, participatory, and cultural methods. Our review reveals substantial conceptual fragmentation in how fairness, social bias, and diversity are defined, measured, and operationalized across domains. While several metrics (e.g., statistical parity, balanced accuracy, generative diversity scores) are applied, standardized computational frameworks and large-scale auditing tools remain largely absent. We conclude by outlining key research challenges, including metric standardization, scalable fairness, social bias, and diversity auditing, cross-cultural dataset development, and interdisciplinary collaboration, to advance aware content generation in games.