paperMay 2026Unreviewed

Towards trustworthy agentic AI: a comprehensive survey of safety, robustness, privacy, and system security

Jinhu Qi, Muzhi Li, Jiahong Liu, Yuqin Shu, Dianzhi Yu, Shicheng Ma, Wenqian Cui, Yiyang Zhao, Yiyi Chen, Ruoxi Jiang, Irwin King, Zenglin Xu

View resource PDF

Abstract

Agentic AI systems -- Large Language Models (LLMs) augmented with planning, tool use, memory, and long-horizon interactions -- can execute complex tasks autonomously, but their multi-step trajectories introduce new failure modes that challenge trustworthiness. This survey provides a focused examination of trustworthy agentic AI through two core dimensions that are critical for high-risk deployments: Safety and Robustness, and Privacy and System Security. For each dimension, we clarify key concep

Framework mappings

OWASP Top 10 for Agentic Applications

ASI02Tool Misuse & Exploitation

MITRE ATLAS

AML.T0053AI Agent Tool Invocation

Suggested from the entry's categories.

Cite

@misc{qi2026trustworthy,
  title = {{Towards trustworthy agentic AI: a comprehensive survey of safety, robustness, privacy, and system security}},
  author = {Jinhu Qi and Muzhi Li and Jiahong Liu and Yuqin Shu and Dianzhi Yu and Shicheng Ma and Wenqian Cui and Yiyang Zhao and Yiyi Chen and Ruoxi Jiang and Irwin King and Zenglin Xu},
  year = {2026},
  month = may,
  eprint = {2605.23989},
  archivePrefix = {arXiv},
  doi = {10.20935/AcadAI8260},
  url = {https://arxiv.org/abs/2605.23989}
}