{
  "id": 7139,
  "url": "https://arxiv.org/abs/2605.13851v1",
  "title": "Invisible Orchestrators Suppress Protective Behavior and Dissociate Power-Holders: Safety Risks in Multi-Agent LLM Systems",
  "summary": "Multi-agent orchestration -- in which a hidden coordinator manages specialized worker agents -- is becoming the default architecture for enterprise AI deployment, yet the safety implications of orchestrator invisibility have never been empirically tested. We conducted a preregistered 3x2 experiment (365 runs, 5 agents per run) crossing three organizational structures (visible leader, invisible orchestrator, flat) with two alignment conditions (base, heavy), using Claude Sonnet 4.5. Four confirma",
  "authors": "Hiroki Fukui",
  "category": "research",
  "topics": "safety-alignment,agents-autonomy",
  "orgs": "anthropic",
  "regions": null,
  "published_at": "2026-03-17T03:18:57.000Z",
  "fetched_at": "2026-07-14T16:32:59.166Z",
  "source_slug": "arxiv-ethics",
  "source_name": "arXiv",
  "source_homepage": "https://arxiv.org",
  "ethics_ai_record_url": "https://ethics.ai/record/7139",
  "original_url": "https://arxiv.org/abs/2605.13851v1",
  "evidence_status": "source-only",
  "attribution": "via ethics.ai"
}