Newest papers

Act on Your Behalf, Beyond Your Authorization: Benchmarking LLM Agent Authorization Alignment

Zhiping Zhang, Franklin Mingzhe Li, Zihao Huang, Tianshi Li · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.27548 · cited by 0 works (OpenAlex) · open-access PDF

Chinese

当一个LLM代理人以较低的任务为代价完成操作时,它应该在用户授予的授权范围内进行操作。现有代理人评估主要衡量任务完成或有害后果,因此可以忽视违反该授权的行为,而不会产生常规安全或隐私损害。我们引入了授权Bench,这是一个基准的授权配置,基于记录的真实情况。一个可扩展的建筑管道产生了286个可执行的安全任务,包括局势触发因素,直接有害后果的存在或缺失,以及十个应用领域。我们在三个代理人中引入了授权配置的基准。我们在授权和非授权的结果和非授权的路径之间保持授权和非授权的安全路径可执行和可观察,从而产生了66

Machine translation (Meta M2M100); the English original follows.

Abstract

When an LLM agent acts on behalf of a user, it is expected to act within the authorization scope the user grants. Existing agent evaluations largely measure task completion or harmful consequences and can therefore overlook actions that violate this delegation without producing conventional safety or privacy harms. We introduce AuthorizationBench, a benchmark of authorization alignment grounded in documented real-world cases. A scalable construction pipeline produces 286 executable tasks spanning situational triggers, the presence or absence of direct harmful consequences, and ten application domains. We execute every task with four models at two reasoning-effort settings across three agent harnesses in an authorization-aware sandbox that keeps authorized and out-of-authorization paths executable and observable, yielding 6,864 trajectories. Agents act out of authorization in 19.2% of trajectories, with rates ranging from 10.5% to 26.6% across the 24 configurations. Task completion does not imply authorization. Increasing reasoning effort leaves the aggregate rate nearly unchanged, with effects varying across models. On controlled pairs, agents act out of authorization at similar rates with (17.9%) and without (20.8%) direct harmful consequences, and more often under benign pressures than malicious injections (23.5% versus 17.6%, p=0.002). Our results show that task capability, additional reasoning, and consequence-oriented safety evaluation do not by themselves ensure authorization alignment. Moreover, agents that consult the principal before completing a task exhibit lower rates of out-of-authorization behavior (14.0% versus 21.6%, p<0.001), pointing to a tension between agent autonomy and preserving users' decision authority that calls for further human-centered study. Dataset and code are available at https://github.com/PEACH-Research-Lab/AuthorizationBench .

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Computer science, Computer security, Task (project management), Benchmarking, Executable, Delegation

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange