Aligned but Stereotypical? The Hidden Influence of System Prompts on Social Bias in LVLM-Based Text-to-Image Models

1KAIST 2HKUST
*Equal contribution, †Corresponding author

FairPro enables socially-fair image generation with automatic generation of fairness-aware system prompt!

Abstract

Large vision–language model (LVLM) based text-to-image (T2I) systems have become the dominant paradigm in image generation, yet whether they amplify social biases remains insufficiently understood. In this paper, we show that LVLM-based models produce markedly more socially biased images than non-LVLM-based models. We introduce a 1,024 prompt benchmark spanning four levels of linguistic complexity and evaluate demographic bias across multiple attributes in a systematic manner. Our analysis identifies system prompts, the predefined instructions guiding LVLMs, as a primary driver of biased behavior. Through decoded intermediate representations, token-probability diagnostics, and embedding-association analyses, we reveal how system prompts encode demographic priors that propagate into image synthesis. To this end, we propose FairPro, a training-free meta-prompting framework that enables LVLMs to self-audit and construct fairness-aware system prompts at test time. Experiments on two LVLM-based T2I models, SANA and Qwen-Image, show that FairPro substantially reduces demographic bias while preserving text–image alignment. We believe our findings provide deeper insight into the central role of system prompts in bias propagation and offer a practical, deployable approach for building more socially responsible T2I systems.

Overall pipeline of our paper.

We (1) evaluate social bias in LVLM-based T2I models with our carefully-constructed benchmark, (2) perform mechanistic analyses to identify system prompts as a key source of bias, and (3) propose FairPro, a training-free meta-prompting framework to mitigate social bias in LVLM-based T2I models.

LVLM-based T2I models offer better text alignment but at the cost of greater social bias.

LVLM-based T2I models demonstrate superior text alignment capabilities compared to traditional (e.g., CLIP) T2I models, but simultaneously exhibit significantly higher levels of social bias across demographic attributes.

FairPro facilitates socially-fair image generation.

With the same text prompt, FairPro generates a variety of images with reduced social bias. Even when explicitly prompted with specific attributes (e.g., gender), generated images with our FairPro show individuals with diverse demographic backgrounds.

BibTeX

@article{park2025fairpro,
  author    = {Park, NaHyeon and An, Namin and Kim, Kunhee and Yoon, Soyeon and Huo, Jiahao and Shim, Hyunjung},
  title     = {Aligned but Stereotypical? The Hidden Influence of System Prompts on Social Bias in LVLM-Based Text-to-Image Models},
  journal   = {arXiv preprint arXiv:2512.04981},
  year      = {2025},
}