arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2601.04043cs.CL
收藏

当助手变成危险:一个多模态大语言模型在日常生活中的安全分析基准

When Helpers Become Hazards: A Benchmark for Analyzing Multimodal LLM-Powered Safety in Daily Life

Key Laboratory of Big Data & Artificial Intelligence in Transportation (Beijing Jiaotong University), Ministry of Education(大数据与人工智能在交通运输中的关键实验室(北京交通大学),教育部);School of Computer Science and Technology, Beijing Jiaotong University(计算机科学与技术学院,北京交通大学);Tsinghua University(清华大学);Peking University(北京大学);University of Montreal(蒙特利尔大学)

Xinyue Lou, Jinan Xu, Jingyi Yin, Xiaolong Wang, Zhaolu Kang, Youwei Liao, Yixuan Wang, Xiangyu Shi, Fengran Mo, Su Yao, Kaiyu Huang

AI总结:

本文提出SaLAD基准,通过2013个真实世界图像-文本样本评估多模态大语言模型在日常生活中的安全影响,揭示模型在识别危险行为方面的局限性。

摘要:

随着多模态大语言模型(MLLMs)成为人类生活中不可或缺的助手,由MLLMs生成的不安全内容对人类行为构成威胁,如同达摩克利斯之剑悬于人类社会之上。为研究和评估MLLMs响应对人类日常行为的安全影响,我们引入SaLAD,一个多模态安全基准,包含2013个真实世界图像-文本样本,覆盖10种常见类别,平衡设计涵盖不安全场景和过度敏感情况。它强调现实风险暴露、真实视觉输入和细粒度跨模态推理,确保安全风险不能仅凭文本推断。我们进一步提出基于安全警告的评估框架,鼓励模型提供清晰且信息丰富的安全警告,而非通用拒绝。在18个MLLMs上的结果表明,表现最佳的模型在不安全查询上的安全响应率仅为57.2%。此外,即使流行的安全对齐方法在我们的场景中也限制了模型的有效性,揭示了当前MLLMs在识别日常生活中危险行为方面的脆弱性。我们的数据集可在https://github.com/xinyuelou/SaLAD获取。

英文摘要:

As Multimodal Large Language Models (MLLMs) become an indispensable assistant in human life, the unsafe content generated by MLLMs poses a danger to human behavior, perpetually overhanging human society like a sword of Damocles. To investigate and evaluate the safety impact of MLLMs responses on human behavior in daily life, we introduce SaLAD, a multimodal safety benchmark which contains 2,013 real-world image-text samples across 10 common categories, with a balanced design covering both unsafe scenarios and cases of oversensitivity. It emphasizes realistic risk exposure, authentic visual inputs, and fine-grained cross-modal reasoning, ensuring that safety risks cannot be inferred from text alone. We further propose a safety-warning-based evaluation framework that encourages models to provide clear and informative safety warnings, rather than generic refusals. Results on 18 MLLMs demonstrate that the top-performing models achieve a safe response rate of only 57.2% on unsafe queries. Moreover, even popular safety alignment methods limit effectiveness of the models in our scenario, revealing the vulnerabilities of current MLLMs in identifying dangerous behaviors in daily life. Our dataset is available at https://github.com/xinyuelou/SaLAD.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑