{
  "id": 7414,
  "url": "https://arxiv.org/abs/2603.09947v1",
  "title": "The Confidence Gate Theorem: When Should Ranked Decision Systems Abstain?",
  "summary": "Ranked decision systems -- recommenders, ad auctions, clinical triage queues -- must decide when to intervene in ranked outputs and when to abstain. We study when confidence-based abstention monotonically improves decision quality, and when it fails. The formal conditions are simple: rank-alignment and no inversion zones. The substantive contribution is identifying why these conditions hold or fail: the distinction between structural uncertainty (missing data, e.g., cold-start) and contextual un",
  "authors": "Ronald Doku",
  "category": "research",
  "topics": "safety-alignment,healthcare",
  "orgs": null,
  "regions": null,
  "published_at": "2026-03-10T17:44:10.000Z",
  "fetched_at": "2026-07-14T16:33:12.391Z",
  "source_slug": "arxiv-ethics",
  "source_name": "arXiv",
  "source_homepage": "https://arxiv.org",
  "ethics_ai_record_url": "https://ethics.ai/record/7414",
  "original_url": "https://arxiv.org/abs/2603.09947v1",
  "evidence_status": "source-only",
  "attribution": "via ethics.ai"
}