{
  "id": 4481,
  "url": "https://arxiv.org/abs/2605.11882v1",
  "title": "On-Policy Self-Evolution via Failure Trajectories for Agentic Safety Alignment",
  "summary": "Tool-using LLM agents fail through trajectories rather than only final responses, as they may execute unsafe tool calls, follow injected instructions, comply with harmful requests, or over-refuse benign tasks despite producing a seemingly safe answer. Existing safety-alignment signals are largely response-level or off-policy, and often incur a safety-utility trade-off: improving agent safety comes at the cost of degraded task performance. Such sparse and single-objective rewards severely limit r",
  "authors": "Bo Yin, Qi Li, Xinchao Wang",
  "category": "research",
  "topics": "regulation,safety-alignment,agents-autonomy",
  "orgs": null,
  "regions": null,
  "published_at": "2026-05-12T09:56:28.000Z",
  "fetched_at": "2026-07-14T16:31:03.577Z",
  "source_slug": "arxiv-ethics",
  "source_name": "arXiv",
  "source_homepage": "https://arxiv.org",
  "ethics_ai_record_url": "https://ethics.ai/record/4481",
  "original_url": "https://arxiv.org/abs/2605.11882v1",
  "evidence_status": "source-only",
  "attribution": "via ethics.ai"
}