English

Personalized Safety Alignment for Text-to-Image Diffusion Models

Computer Vision and Pattern Recognition 2026-02-06 v3 Artificial Intelligence

Abstract

Text-to-image diffusion models have revolutionized visual content generation, yet their deployment is hindered by a fundamental limitation: safety mechanisms enforce rigid, uniform standards that fail to reflect diverse user preferences shaped by age, culture, or personal beliefs. To address this, we propose Personalized Safety Alignment (PSA), a framework that transitions generative safety from static filtration to user-conditioned adaptation. We introduce Sage, a large-scale dataset capturing diverse safety boundaries across 1,000 simulated user profiles, covering complex risks often missed by traditional datasets. By integrating these profiles via a parameter-efficient cross-attention adapter, PSA dynamically modulates generation to align with individual sensitivities. Extensive experiments demonstrate that PSA achieves a calibrated safety-quality trade-off: under permissive profiles, it relaxes over-cautious constraints to enhance visual fidelity, while under restrictive profiles, it enforces state-of-the-art suppression, significantly outperforming static baselines. Furthermore, PSA exhibits superior instruction adherence compared to prompt-engineering methods, establishing personalization as a vital direction for creating adaptive, user-centered, and responsible generative AI. Our code, data, and models are publicly available at https://github.com/M-E-AGI-Lab/PSAlign.

Keywords

Cite

@article{arxiv.2508.01151,
  title  = {Personalized Safety Alignment for Text-to-Image Diffusion Models},
  author = {Yu Lei and Jinbin Bai and Qingyu Shi and Aosong Feng and Hongcheng Gao and Xiao Zhang and Rex Ying},
  journal= {arXiv preprint arXiv:2508.01151},
  year   = {2026}
}
R2 v1 2026-07-01T04:30:29.176Z