{
  "id": 11373,
  "url": "https://arxiv.org/abs/2607.14285v1",
  "title": "ToolAlignBench: Investigating Alignment Conflicts in Tool-Calling Enabled LLMs",
  "summary": "Safety alignment in LLMs aims to align models with human values, but which values take precedence when they conflict? We investigate this question in the context of tool-calling LLM agents deployed in regulated industries, where agents processing confidential documents may encounter content that triggers safety-trained values (e.g., public welfare) that conflict with deployment-context instructions (e.g., internal logging). To empirically verify this phenomenon, we build a benchmark of 128 scena",
  "authors": "Aryan Keluskar, Amrita Bhattacharjee, Huan Liu",
  "category": "research",
  "topics": "regulation,safety-alignment,agents-autonomy,finance-investment",
  "orgs": null,
  "regions": null,
  "published_at": "2026-07-15T18:48:49.000Z",
  "fetched_at": "2026-07-18T05:10:55.931Z",
  "source_slug": "arxiv-ethics",
  "source_name": "arXiv",
  "source_homepage": "https://arxiv.org",
  "ethics_ai_record_url": "https://ethics.ai/record/11373",
  "original_url": "https://arxiv.org/abs/2607.14285v1",
  "evidence_status": "source-only",
  "attribution": "via ethics.ai"
}