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 2502.08586cs.LGcs.AI
收藏

商用LLM智能体已易受简单却危险的攻击

Commercial LLM Agents Are Already Vulnerable to Simple Yet Dangerous Attacks

Columbia University(哥伦比亚大学);University of Maryland(马里兰大学)

Ang Li, Yin Zhou, Vethavikashini Chithrra Raghuram, Tom Goldstein, Micah Goldblum

AI总结:

本文针对LLM智能体特有的安全与隐私漏洞展开研究,提出攻击分类体系并对开源及商用智能体实施简单攻击,证明其比孤立LLMs更易受攻击,为智能体安全研究提供方向。

摘要:

近期大量机器学习安全文献聚焦于针对对齐大型语言模型(LLMs)的攻击,这些攻击可能提取隐私信息或迫使模型生成有害输出。在实际部署中,LLMs常作为更大智能体管道的一部分,包含记忆系统、检索、网络访问和API调用等组件。这些额外组件引入了漏洞,使LLM驱动的智能体比孤立LLMs更易被攻击,但针对LLM智能体安全性的研究相对较少。本文分析LLM智能体特有的安全与隐私漏洞,首先提出按威胁行为者、目标、入口点、攻击者可观测性、攻击策略及智能体管道固有漏洞分类的攻击 taxonomy;随后对流行开源及商用智能体开展一系列示例攻击,证明其漏洞的直接实际影响,且这些攻击实现简单,无需机器学习知识。

英文摘要:

A high volume of recent ML security literature focuses on attacks against aligned large language models (LLMs). These attacks may extract private information or coerce the model into producing harmful outputs. In real-world deployments, LLMs are often part of a larger agentic pipeline including memory systems, retrieval, web access, and API calling. Such additional components introduce vulnerabilities that make these LLM-powered agents much easier to attack than isolated LLMs, yet relatively little work focuses on the security of LLM agents. In this paper, we analyze security and privacy vulnerabilities that are unique to LLM agents. We first provide a taxonomy of attacks categorized by threat actors, objectives, entry points, attacker observability, attack strategies, and inherent vulnerabilities of agent pipelines. We then conduct a series of illustrative attacks on popular open-source and commercial agents, demonstrating the immediate practical implications of their vulnerabilities. Notably, our attacks are trivial to implement and require no understanding of machine learning.

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。

↑