{"ID":2887475,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2508.01151","arxiv_id":"2508.01151","title":"Personalized Safety Alignment for Text-to-Image Diffusion Models","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.","short_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 Personaliz...","url_abs":"https://arxiv.org/abs/2508.01151","url_pdf":"https://arxiv.org/pdf/2508.01151v3","authors":"[\"Yu Lei\",\"Jinbin Bai\",\"Qingyu Shi\",\"Aosong Feng\",\"Hongcheng Gao\",\"Xiao Zhang\",\"Rex Ying\"]","published":"2025-08-02T02:23:20Z","proceeding":"cs.CV","tasks":"[\"cs.CV\",\"cs.AI\"]","methods":"[\"Diffusion Model\"]","has_code":false,"code_links":[{"ID":611439,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_id":2887475,"paper_url":"https://arxiv.org/abs/2508.01151","paper_title":"Personalized Safety Alignment for Text-to-Image Diffusion Models","repo_url":"https://github.com/M-E-AGI-Lab/PSAlign","is_official":false,"mentioned_in_paper":false,"mentioned_in_github":true,"github_stars":0}]}