{
  "schema": "vedokrok.public-item.v1",
  "release_id": "MHC-RPUB-20260920-75ad787a",
  "url": "/knowledge/judge-the-expensive-error-separately-from-the-common-one",
  "id": "MHC-D-RESEARCH-0991",
  "version": "0.1.0",
  "title": "Judge the expensive error separately from the common one",
  "summary": "A thousand correct commas do not cancel one wrong recipient.",
  "kind": "template",
  "body": "Separate error types before deciding whether the system is acceptable. For a classifier, begin with false positives and false negatives; for an assistant, define similarly concrete failure categories. Assign review or blocking rules to consequential failures rather than letting them disappear inside an average.",
  "limits": [
    "Rare failures can be absent from a small test by chance. Consequence weights are judgments and should not be disguised as measured facts."
  ],
  "topics": [
    "work-03-ai-assurance"
  ],
  "intents": [],
  "source_ids": [
    "RS-758FDEE956835687"
  ],
  "evidence": [
    {
      "claim": "A confusion matrix separates actual classes from predicted classes and exposes different error directions hidden by an aggregate accuracy figure.",
      "source_id": "RS-758FDEE956835687",
      "role": "supports",
      "note": "The matrix does not determine the harm or acceptability of each error; those require task-specific judgment.",
      "locator": "Confusion matrix"
    }
  ],
  "use_when": [
    "An AI tool has an attractive average score but some mistakes would be much more costly than others."
  ],
  "avoid_when": [
    "Rare failures can be absent from a small test by chance. Consequence weights are judgments and should not be disguised as measured facts."
  ],
  "example": "A routing assistant's harmless category mistakes and messages sent to an unauthorized team are reported separately.",
  "check": "The acceptance decision names the critical error counts and rules, not just the overall pass rate.",
  "template": "Failure type: [error]. Consequence: [harm]. Observed cases and denominator: [evidence]. Required handling: [gate]. Owner of acceptance: [owner].",
  "sources": [
    {
      "id": "RS-758FDEE956835687",
      "title": "Metrics and scoring: quantifying the quality of predictions",
      "url": "https://scikit-learn.org/stable/modules/model_evaluation.html"
    }
  ],
  "relations": [
    {
      "from": "MHC-D-RESEARCH-0991",
      "to": "MHC-D-RESEARCH-0990",
      "type": "useful_with",
      "url": "/knowledge/keep-acceptance-examples-out-of-the-prompt-workshop"
    }
  ],
  "collections": [
    {
      "id": "RC-620AACFCDEA000D1",
      "title": "Make AI-assisted work earn your trust",
      "url": "/collections/make-ai-assisted-work-earn-your-trust"
    }
  ]
}