{
  "$schema": "https://themachinepress.com/schemas/story-v1.schema.json",
  "schema_version": "1.0.0",
  "document_type": "machine_press_story",
  "story": {
    "story_id": "mp-2026-08-20-018",
    "source_story_id": "tmp-sidebar-high-risk-document-extraction",
    "edition_id": "mp-2026-08-20-morning-0042",
    "edition_url": "https://themachinepress.com/edition/2026-08-20",
    "position": 20,
    "story_type": "ticker",
    "section": "safety-security",
    "editorial_classification": "editorial",
    "headline": "Thirty-One of Thirty-Five Pipelines Missed the Bar",
    "slug": "thirty-one-of-thirty-five-pipelines-missed-the-bar",
    "dek": "Only four open document-extraction configurations cleared 0.5 F1 on a high-risk student-application task.",
    "summary": "Only four open document-extraction configurations cleared 0.5 F1 on a high-risk student-application task.",
    "body_text": "Vision-language models generally led OCR-plus-LLM systems, yet about three quarters of all configurations scored below 0.25. Preserving document structure mattered independently of model size, warning against zero-shot deployment in consequential workflows.",
    "why_it_matters": "Only four open document-extraction configurations cleared 0.5 F1 on a high-risk student-application task.",
    "limitations": [],
    "importance": 9,
    "canonical_url": "https://themachinepress.com/story/mp-2026-08-20-018/thirty-one-of-thirty-five-pipelines-missed-the-bar",
    "json_url": "https://themachinepress.com/story/mp-2026-08-20-018.json",
    "first_published_at": "2026-08-20T09:00:00.000-04:00",
    "modified_at": "2026-08-20T09:00:00.000-04:00",
    "content_status": "new",
    "is_carryover": false,
    "carryover_reason": null,
    "key_claims": [
      {
        "claim_id": "claim-mp-2026-08-20-018-001",
        "text": "Only four open document-extraction configurations cleared 0.5 F1 on a high-risk student-application task.",
        "source_ids": [
          "source-2026-08-20-020"
        ],
        "qualification": null
      }
    ],
    "source_ids": [
      "source-2026-08-20-020"
    ],
    "tags": [
      "document AI",
      "public sector",
      "evaluation"
    ],
    "image_url": null,
    "corrections": []
  },
  "sources": [
    {
      "source_id": "source-2026-08-20-020",
      "title": "arXiv preprint 2608.18289",
      "publisher": "arXiv",
      "url": "https://arxiv.org/abs/2608.18289",
      "canonical_url": "https://arxiv.org/abs/2608.18289",
      "source_type": "primary_research",
      "is_primary_source": true,
      "published_at": "2026-08-17T20:00:00.000-04:00",
      "accessed_at": "2026-08-20T08:20:00.000-04:00",
      "supports_claim_ids": [
        "claim-mp-2026-08-20-018-001"
      ]
    }
  ],
  "corrections": [],
  "publisher": {
    "name": "The Machine Press",
    "url": "https://themachinepress.com",
    "description": "A daily newspaper for the age of artificial intelligence."
  },
  "cite_this_report": {
    "title": "Thirty-One of Thirty-Five Pipelines Missed the Bar",
    "publisher": "The Machine Press",
    "published_at": "2026-08-20T09:00:00.000-04:00",
    "canonical_url": "https://themachinepress.com/story/mp-2026-08-20-018/thirty-one-of-thirty-five-pipelines-missed-the-bar"
  }
}