arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2503.09780cs.AI
收藏

AgentDAM:面向自主Web智能体的隐私泄露评估

AgentDAM: Privacy Leakage Evaluation for Autonomous Web Agents

FAIR at Meta(Meta 的公平性研究部门)

Arman Zharmagambetov, Chuan Guo, Ivan Evtimov, Maya Pavlova, Ruslan Salakhutdinov, Kamalika Chaudhuri

AI总结:

本研究提出评估Web智能体数据最小化隐私遵循情况的基准AgentDAM,发现GPT-4等主流模型构建的智能体易误用敏感信息,同时提出基于提示的防御方法,证实端到端基准比直接探测LLM更具现实评估价值。

摘要:

能够遵循指令并执行复杂多步骤任务的自主AI智能体在提升人类生产力方面具有巨大潜力。然而,为完成其中许多任务,智能体需要获取用户的个人信息,这引发了它们是否能够恰当使用这些信息的疑问。在本研究中,我们提出了一个全新的基准AgentDAM,用于衡量AI网页导航智能体是否遵循“数据最小化”隐私原则。在本基准的定义中,数据最小化指智能体仅在某条潜在敏感信息对完成特定任务“必要”时才使用它。该基准端到端模拟真实的网页交互场景,且适配所有现有的网页导航智能体。我们使用AgentDAM评估了基于GPT-4、Llama-3和Claude构建的AI智能体在限制潜在隐私信息处理方面的表现,结果显示它们容易在无意中使用不必要的敏感信息。我们还提出了一种基于提示工程的防御方法以减少信息泄露,并证明我们的端到端基准测试比直接探测大语言模型(LLM)的隐私性更具现实衡量价值。我们的研究结果表明,亟需进一步研究以开发出能在推理阶段优先遵循数据最小化原则的AI智能体。

英文摘要:

Autonomous AI agents that can follow instructions and perform complex multi-step tasks have tremendous potential to boost human productivity. However, to perform many of these tasks, the agents need access to personal information from their users, raising the question of whether they are capable of using it appropriately. In this work, we introduce a new benchmark AgentDAM that measures if AI web-navigation agents follow the privacy principle of ``data minimization''. For the purposes of our benchmark, data minimization means that the agent uses a piece of potentially sensitive information only if it is ``necessary'' to complete a particular task. Our benchmark simulates realistic web interaction scenarios end-to-end and is adaptable to all existing web navigation agents. We use AgentDAM to evaluate how well AI agents built on top of GPT-4, Llama-3 and Claude can limit processing of potentially private information, and show that they are prone to inadvertent use of unnecessary sensitive information. We also propose a prompting-based defense that reduces information leakage, and demonstrate that our end-to-end benchmarking provides a more realistic measure than probing LLMs about privacy. Our results highlight that further research is needed to develop AI agents that can prioritize data minimization at inference time.

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。

↑