{
  "id": 967,
  "url": "https://arxiv.org/abs/2606.17114v1",
  "title": "An Evaluation of Data Leakage Risks in Tool-Using LLM Agents in Realistic Scenarios",
  "summary": "AI agents are increasingly being adopted in enterprise and personal settings with access to emails, databases, documents, and other tools where they can read, update, and disseminate sensitive information. Much of prior research on data leakage risks in agents has focused on adversarial data exfiltration through prompt injections and jailbreaks. However, sensitive information may also be exposed during non-adversarial use, creating leakage risks even when users issue benign requests. We report a",
  "authors": "Hankyul Baek, Jaewon Noh, Sang Seo, Yongsu Kim, Gabriel Waikin Loh Matienzo, Young Il Kim et al.",
  "category": "research",
  "topics": "safety-alignment,agents-autonomy",
  "orgs": null,
  "regions": null,
  "published_at": "2026-06-15T09:16:38.000Z",
  "fetched_at": "2026-07-14T14:14:54.534Z",
  "source_slug": "arxiv-ethics",
  "source_name": "arXiv",
  "source_homepage": "https://arxiv.org",
  "ethics_ai_record_url": "https://ethics.ai/record/967",
  "original_url": "https://arxiv.org/abs/2606.17114v1",
  "evidence_status": "source-only",
  "attribution": "via ethics.ai"
}