Metraipolis › archive › dc_62be0ce94de1

AgentPrivArena: Evaluating and Auditing Real-world AI Agent Privacy — arXiv:2610.06454

Source: arXiv
Original: https://arxiv.org/abs/2610.06454
License (as archived): CC0-1.0
Archived: 2026-10-06T05:25:35.894Z
SHA-256 of the archived text: 84b872db1cd8a996ac993e2f16826d7616772604591e85d640ecabf867149a54
Stable URL of this page: https://metraipolis.com/documento/dc_62be0ce94de1

How to cite

AgentPrivArena: Evaluating and Auditing Real-world AI Agent Privacy — arXiv:2610.06454 — arXiv, https://arxiv.org/abs/2610.06454 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_62be0ce94de1 (Metraipolis archive dc_62be0ce94de1, sha256 84b872db1cd8a996ac993e2f16826d7616772604591e85d640ecabf867149a54)

BibTeX

@misc{dc62be0ce94de1,
  title = {AgentPrivArena: Evaluating and Auditing Real-world AI Agent Privacy — arXiv:2610.06454},
  howpublished = {arXiv},
  url = {https://arxiv.org/abs/2610.06454},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_62be0ce94de1; sha256 84b872db1cd8a996ac993e2f16826d7616772604591e85d640ecabf867149a54},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv, https://arxiv.org/abs/2610.06454. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

The rapid advancement of LLM agents has enabled systems to autonomously perform complex tasks through external tools, but their growing access to personal data introduces significant privacy risks. Existing benchmarks primarily evaluate LLM agent privacy through simulated trajectories and outcome-based metrics, limiting their ability to capture privacy risks arising during multi-step agent execution. In this work, we introduce AgentPrivArena, a framework for evaluating privacy risks in realistic LLM agent workflows. AgentPrivArena integrates authentic MCP tools and self-hosted services within a reproducible execution environment. We further propose trajectory-level privacy metrics that quantify unnecessary information access beyond final response leakage. Building on this framework, we introduce AgentPrivAudit, a runtime auditing approach for monitoring privacy violations during agent execution. Extensive experiments on state-of-the-art LLM agents reveal substantial privacy risks overlooked by existing evaluation paradigms, highlighting the importance of trajectory-level auditing for trustworthy agent deployment.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2610.06454 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-06T05:24:05.681Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).