Text to image AI models show strong gender bias in jobs
Newer Is Not Fairer: Gender Stereotyping in Text-to-Image AI Across Model Generations
Computer Vision and Pattern RecognitionArtificial IntelligenceComputers and SocietyMachine Learning
Summary
Text-to-image AI models often create images that wrongly show men more than women in many jobs, especially in roles typically done by women. The authors tested different versions of a popular AI model and found that newer versions did not get better at showing people fairly by gender. For some jobs, like scientists and cleaners, the images mostly showed men even when many women work in those roles in real life. This bias hasn’t gone away in these AI tools, which might affect how people see gender roles.
What this means in practice
- •For ai model developers: Evaluate and reduce gender bias in image generation across model versions and prompts using occupation-based benchmarks.
- •For diversity and inclusion teams: Assess AI tools used in creative workflows to identify and mitigate gender stereotypes in generated content.
Authors
Shesh Narayan Gupta, Nik Bear Brown
Abstract
Text-to-image generative models are widely used in professional and creative settings, yet how they represent gender across occupations -- and whether newer models are fairer -- remains poorly understood across multiple generations. We evaluate gender representation across 20 occupations, 5 prompt templates, and 4 Stable Diffusion model generations (SD 1.5, SD 2.1, SDXL, SD 3 Medium), generating 8,000 images with n = 100 per occupation-model cell (5 prompts x 20 images), and classifying all with DeepFace. Across the 8,000 open-source images, 76.4% show male subjects (95% CI [75.1%, 78.7%], p < 2.2 x 10^-16, Benjamini-Hochberg adjusted). More strikingly, 57.6% of images for historically female-coded occupations show male subjects (raw p = 3.43 x 10^-22, BH-adjusted p = 1.71 x 10^-21). All nine significant tests reported in this paper survive BH correction across 10 tests. When compared against U.S. Bureau of Labor Statistics workforce data, models underrepresent women by 20-46pp on average, with particularly large deviations for near gender-balanced occupations: scientist (48% female in BLS, 82-99% male in model outputs) and cleaner (46% female in BLS, 80-92% male in outputs). Model generations do not improve steadily: bias worsens from SD 1.5 to SDXL before partially recovering in SD 3 Medium. A preliminary comparison with GPT-image-1 on five occupations suggests lower bias than open-source models, though the practical effect is small (Cramer's V = 0.080) and the comparison is exploratory. No model achieves gender parity.