This project examines gender biases and stereotypes in ChatGPT-generated short stories and images, connecting its analysis to the Beijing Platform for Action’s Strategic Objective J2, which calls for balanced and non-stereotypical portrayals of women in the media. Motivated by concerns over how generative AI may perpetuate harmful representations—particularly in educational contexts—the study explores AI’s growing role in the creative media industry and its implications for media literacy. The project targets students aged 11–18, educators, AI users, researchers, and developers, aiming both to raise awareness of AI-driven stereotypes and to encourage improvements in the design of inclusive, bias-free creative tools.
The methodology involved designing neutral prompts for 300-character stories across three genres (sci-fi, romance, fairytale), each with male and female protagonists, generating corresponding images for five countries (US, Germany, Chile, Brazil, South Korea), and conducting qualitative analyses of physical and personality traits. The findings informed the creation of an educator’s guide, integrating activities and discussion prompts to help students critically engage with AI-generated content. While limited time and resources restricted the scope—preventing deeper cross-linguistic comparisons, broader intersectional analysis, and extended story or image generation—the project serves as a foundational step for advocacy in the education sector. It provides a starting point for broader conversations on how generative AI can reinforce or challenge societal stereotypes, with potential applications in both research and policy development.
