{
  "id": 4079,
  "url": "https://arxiv.org/abs/2605.19127v1",
  "title": "POLAR-Bench: A Diagnostic Benchmark for Privacy-Utility Trade-offs in LLM Agents",
  "summary": "LLM agents increasingly have access to private user data and act on the user's behalf when interacting with third-party systems. The user defines what may and must not be shared, and the agent must robustly follow that intent even when third-party systems behave adversarially. We introduce POLAR-Bench (Policy-aware adversarial Benchmark), in which a trusted model with a privacy policy and a task converses with a third-party model that adversarially probes for both task-relevant and protected att",
  "authors": "Qiaoyuan Zheng, Yiqu Yang, Qi Gao, Imanol Schlag",
  "category": "research",
  "topics": "regulation,privacy-surveillance,healthcare,agents-autonomy",
  "orgs": null,
  "regions": null,
  "published_at": "2026-05-18T21:27:07.000Z",
  "fetched_at": "2026-07-14T16:30:45.937Z",
  "source_slug": "arxiv-ethics",
  "source_name": "arXiv",
  "source_homepage": "https://arxiv.org",
  "ethics_ai_record_url": "https://ethics.ai/record/4079",
  "original_url": "https://arxiv.org/abs/2605.19127v1",
  "evidence_status": "source-only",
  "attribution": "via ethics.ai"
}