能力强但粗心：计算机使用智能体是否遵循情境完整性？

AI 导读

AgentCIBench评估计算机使用智能体（CUA）是否遵循情境完整性。它针对三种常见失败模式：视觉共置（智能体拉取任务目标旁边被禁止的项目）、任务模糊性过度分享（在提示不明确时泄露个人状态）以及收件人错配（向不适当的收件人发送内容）。对15个前沿CUA的评测显示平均泄漏率67.9%，其中11个在超过50%的场景中泄漏，这些失败在端到端任务中同样存在。AgentCIBench已发布，旨在推动开发更安全的计算机使用智能体。

推荐理由

计算机使用代理的隐私泄露问题被严重低估了。这篇论文用 AgentCIBench 实测 15 个前沿代理，发现平均泄漏率接近 70%，把这个隐患摆到了台面上，做 agent 产品的团队该把它加入上线前测试清单。

正文

View PDF HTML (experimental)

Abstract:Computer-use agents (CUAs) now act on a user's behalf across personal applications such as email, calendars, and to-do lists. This cross-application access is useful, but it also creates a privacy risk that has been largely overlooked: when an agent works in one context, it can pull in information from another that is inappropriate in that context. Hence, we introduce AgentCIBench, an evaluation harness that turns this risk into executable, deterministically scored scenarios. We target three common failure modes in CUAs: visual co-location, where the agent pulls in prohibited items that sit next to the task target in the UI; task-ambiguity overshare, where the agent dumps dense personal state in response to an under-specified prompt; and recipient misalignment, where the agent sends content to an addressee for whom it is inappropriate. We evaluate 15 frontier agents and find a surprisingly high failure rate: 11 of 15 leak on more than 50% of scenarios, with an average leakage of 67.9%, and the same failures persist when agents act end-to-end in the environment to complete the task. We release AgentCIBench to encourage the development of safer computer-use agents and position contextual disclosure testing as a pre-deployment safety check.

Comments:	EMNLP 2026 Main; Code: this https URL
Subjects:	Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
Cite as:	arXiv:2606.23189 [cs.AI]
	(or arXiv:2606.23189v2 [cs.AI] for this version)
	https://doi.org/10.48550/arXiv.2606.23189 arXiv-issued DOI via DataCite

Submission history

From: Anmol Goel [view email]
[v1] Mon, 22 Jun 2026 11:36:58 UTC (1,147 KB)
[v2] Thu, 10 Sep 2026 18:49:54 UTC (1,146 KB)