arXiv cs.AI· Zihao Zhu, Yuhan Liu, Siwei Lyu, Baoyuan Wu·· 14 小时前AI 评分49
FVH-Bench：评估图像生成器超越显性有害图像的安全风险
FVH-Bench: Evaluating Image Generator Safety Beyond Overtly Harmful Imagery
AI 导读
研究者提出 FVH-Bench，用 669 条提示词覆盖五种有害视觉功能与十个风险领域，评估图像生成器在“功能性视觉危害”（FVH）上的安全对齐。11 个受测图像生成器很少拒绝这类请求，且常生成能实现目标有害功能的图像，目标危害实现度与有害产物可用性得分均偏高。现有安全检测器对 FVH 的覆盖同样有限。
正文 · 原文
Abstract:Modern image generators can produce complex visual artifacts, including diagrams, interfaces, and records that combine text, graphics, and layout to convey detailed information. They may refuse requests for overtly harmful imagery yet still generate artifacts that guide dangerous actions or support deception. We call the risk arising from what such images enable, rather than only what they depict, functional visual harm (FVH). Existing safety benchmarks focus largely on overtly harmful content, such as pornography and graphic violence, leaving image generators' safety alignment against FVH insufficiently evaluated. To address this gap, we introduce FVH-Bench, comprising 669 prompts across five harmful visual functions and ten risk domains. The 11 evaluated image generators rarely refuse these requests and frequently generate artifacts that fulfill the targeted harmful functions. High scores for target harm realization and harmful artifact usability further indicate that these outputs can support the intended harmful purposes. A further analysis finds that existing safety detectors also provide limited coverage of FVH, indicating gaps in their ability to recognize harmful visual functions. Together, these findings show that evaluating image-generator safety alignment requires considering not only what images depict, but also what they may enable.
| Subjects: | Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.37722 [cs.CR] |
| (or arXiv:2610.37722v1 [cs.CR] for this version) | |
| https://doi.org/10.48550/arXiv.2610.37722 arXiv-issued DOI via DataCite |
Submission history
From: Zihao Zhu [view email]
[v1]
Tue, 27 Oct 2026 10:35:58 UTC (21,350 KB)
来源：arXiv cs.AI · arxiv.org