{
  "identifier": "tag:aim-pro.eu,2026:oer/ac97f0396dda/record",
  "profile": "aimpro-oer-profile/1",
  "normaliser": "oerlib.metadata/2",
  "status": "ready",
  "normalised_at": "2026-10-09T09:48:08+00:00",
  "resource": {
    "id": "ac97f0396dda",
    "identifier": "tag:aim-pro.eu,2026:oer/ac97f0396dda",
    "uri": "https://arxiv.org/abs/1811.10154",
    "title": "Stop Explaining Black Box Machine Learning Models for High Stakes Decisions and Use Interpretable Models Instead",
    "description": "Black box machine learning models are currently being used for high stakes decision-making throughout society, causing problems throughout healthcare, criminal justice, and in other domains. People have hoped that creating methods for explaining these black box models will alleviate some of these problems, but trying to \\textit{explain} black box models, rather than creating models that are \\textit{interpretable} in the first place, is likely to perpetuate bad practices and can potentially cause catastrophic harm to society. There is a way forward -- it is to design models that are inherently interpretable. This manuscript clarifies the chasm between explaining black boxes and using inherently interpretable models, outlines several key reasons why explainable black boxes should be avoided in high-stakes decisions, identifies challenges to interpretable machine learning, and provides several example applications where interpretable models could potentially replace black box models in criminal justice, healthcare, and computer vision.",
    "creators": [
      "Cynthia Rudin"
    ],
    "keywords": [
      "stat.ML",
      "cs.LG"
    ],
    "kind": "html",
    "content_language": "en",
    "declared_language": "en",
    "word_count": 13437,
    "content_hash": "00138a0e564ebec0",
    "metadata_only": false
  },
  "licence": {
    "raw": "CC-BY-SA-4.0",
    "spdx": "CC-BY-SA-4.0",
    "uri": "http://creativecommons.org/licenses/by-sa/4.0/",
    "verdict": "SHARE_ALIKE",
    "note": "open but ShareAlike/copyleft — derivatives keep the license",
    "ingestable": true,
    "commercial_ok": true,
    "derivable": true,
    "share_alike": true,
    "evidence": {
      "scope": "this submitted version of the paper",
      "method": "arXiv OAI-PMH GetRecord (metadataPrefix=arXiv)",
      "schema": 1,
      "status": "captured",
      "location": "arXiv:arXiv/arXiv:license",
      "source_url": "https://oaipmh.arxiv.org/oai?verb=GetRecord&identifier=oai:arXiv.org:1811.10154&metadataPrefix=arXiv",
      "asserted_by": "arXiv submitter",
      "captured_at": "2026-10-09T09:48:05+00:00",
      "source_system": "arxiv",
      "observed_value": "http://creativecommons.org/licenses/by-sa/4.0/",
      "documentation_url": "https://info.arxiv.org/help/oa/index.html",
      "responsibility_url": "https://info.arxiv.org/help/license/index.html",
      "documentation_label": "arXiv OAI-PMH metadata documentation",
      "responsibility_note": "The submitter chooses the article-version licence and certifies the right to grant it."
    }
  },
  "descriptive": {
    "contributions": [
      {
        "name": "Cynthia Rudin",
        "role": "author",
        "affiliation": "",
        "identifier": "",
        "scope": "resource"
      }
    ],
    "identifiers": [
      {
        "scheme": "arxiv",
        "value": "1811.10154",
        "scope": "resource"
      }
    ],
    "dates": [
      {
        "meaning": "issued",
        "value": "2018-11-26",
        "scope": "resource"
      },
      {
        "meaning": "modified",
        "value": "2019-09-22",
        "scope": "resource"
      }
    ],
    "subjects": [
      {
        "value": "stat.ML",
        "scheme": "arxiv",
        "scope": "resource"
      },
      {
        "value": "cs.LG",
        "scheme": "arxiv",
        "scope": "resource"
      }
    ],
    "relations": [],
    "educational": {
      "learning_resource_type": "",
      "intended_end_user_role": "",
      "context": "",
      "difficulty": "",
      "typical_learning_time": "",
      "description": ""
    },
    "container": {
      "kind": "",
      "uri": "",
      "title": "",
      "description": "",
      "identifier": ""
    },
    "declared_language": "en",
    "declared_type": "preprint",
    "rights_holder": "",
    "access_rights": "open",
    "publisher": "",
    "version": "v3",
    "citation": "Nature Machine Intelligence, Vol 1, May 2019, 206-215",
    "notes": "",
    "lifecycle_status": "revised",
    "structure_role": "",
    "snapshot": {
      "resource": {
        "title": "Stop Explaining Black Box Machine Learning Models for High Stakes Decisions and Use Interpretable Models Instead",
        "authors": [
          "Cynthia Rudin"
        ],
        "summary": "Black box machine learning models are currently being used for high stakes decision-making throughout society, causing problems throughout healthcare, criminal justice, and in other domains. People have hoped that creating methods for explaining these black box models will alleviate some of these problems, but trying to \\textit{explain} black box models, rather than creating models that are \\textit{interpretable} in the first place, is likely to perpetuate bad practices and can potentially cause catastrophic harm to society. There is a way forward -- it is to design models that are inherently interpretable. This manuscript clarifies the chasm between explaining black boxes and using inherently interpretable models, outlines several key reasons why explainable black boxes should be avoided in high-stakes decisions, identifies challenges to interpretable machine learning, and provides several example applications where interpretable models could potentially replace black box models in criminal justice, healthcare, and computer vision.",
        "updated": "2019-09-22T03:05:09Z",
        "version": 3,
        "arxiv_id": "1811.10154",
        "published": "2018-11-26T03:00:25Z",
        "categories": [
          "stat.ML",
          "cs.LG"
        ],
        "journal_ref": "Nature Machine Intelligence, Vol 1, May 2019, 206-215",
        "affiliations": [
          ""
        ],
        "primary_category": "stat.ML"
      }
    }
  },
  "representations": [
    {
      "role": "extracted",
      "media_type": "text/markdown",
      "size_bytes": 84815,
      "filename": "",
      "location": "",
      "source_format": "latexml-html",
      "converter": "arxiv-html"
    }
  ],
  "observed": {
    "language": {
      "method": "declared",
      "reason": "declared English",
      "accepted": true,
      "declared": "en",
      "detected": "en",
      "confidence": null,
      "review_required": false
    }
  },
  "collection": {
    "source": "arxiv",
    "source_label": "arXiv",
    "collected_at": "2026-10-09T09:48:08.346206+00:00",
    "query": "(all:\"artificial intelligence\" OR all:\"machine learning\" OR all:\"generative AI\" OR all:\"deep learning\" OR all:\"reinforcement learning\" OR all:\"large language model\") AND (all:\"AI concepts\" OR all:\"types of AI\" OR all:\"AI fundamentals\" OR all:\"recognizing AI\" OR all:\"recognising AI\" OR all:\"general versus narrow AI\" OR all:\"narrow AI\" OR all:\"general AI\" OR all:\"machine intelligence\" OR all:\"AI strengths and weaknesses\" OR all:\"traditional software\" OR all:\"rule-based systems\" OR all:\"introduction to AI\" OR all:\"introduction to artificial intelligence\" OR all:\"artificial intelligence introduction\" OR all:\"AI primer\" OR all:\"foundations of artificial intelligence\" OR all:\"overview of AI\" OR all:\"understanding AI\" OR all:\"history of AI\" OR all:\"AI essentials\" OR all:\"AI terminology\" OR all:\"metaphors for AI\" OR all:\"AI fundamental concepts\" OR all:\"AI key concepts\" OR all:\"philosophy of AI\" OR all:\"critical AI literacy\")",
    "note": "where this was collected from. Not a claim that the provider published or authored it"
  },
  "provenance": [
    {
      "field": "resource_id",
      "scope": "resource",
      "method": "engine",
      "evidence_ref": "engine#/resource_id"
    },
    {
      "field": "query",
      "scope": "resource",
      "method": "engine",
      "evidence_ref": "engine#/query"
    },
    {
      "field": "collected_from",
      "scope": "resource",
      "method": "engine",
      "evidence_ref": "engine#/source"
    },
    {
      "field": "title",
      "scope": "resource",
      "method": "source_metadata",
      "evidence_ref": "snapshot#/resource/title"
    },
    {
      "field": "description",
      "scope": "resource",
      "method": "source_metadata",
      "evidence_ref": "snapshot#/resource/summary"
    },
    {
      "field": "citation",
      "scope": "resource",
      "method": "source_metadata",
      "evidence_ref": "snapshot#/resource/journal_ref"
    },
    {
      "field": "version",
      "scope": "resource",
      "method": "source_metadata",
      "evidence_ref": "snapshot#/resource"
    },
    {
      "field": "lifecycle_status",
      "scope": "resource",
      "method": "source_metadata",
      "evidence_ref": "snapshot#/resource"
    },
    {
      "field": "subjects",
      "scope": "resource",
      "method": "source_metadata",
      "evidence_ref": "snapshot#/resource/categories/0"
    },
    {
      "field": "contributions",
      "scope": "resource",
      "method": "source_metadata",
      "evidence_ref": "snapshot#/resource/authors/0"
    },
    {
      "field": "dates",
      "scope": "resource",
      "method": "source_metadata",
      "evidence_ref": "snapshot#/resource"
    },
    {
      "field": "identifiers",
      "scope": "resource",
      "method": "source_metadata",
      "evidence_ref": "snapshot#/resource/arxiv_id"
    },
    {
      "field": "content_language",
      "scope": "resource",
      "method": "gate_observation",
      "evidence_ref": "gate#/language/declared"
    },
    {
      "field": "declared_language",
      "scope": "resource",
      "method": "source_metadata",
      "evidence_ref": "snapshot#/resource"
    },
    {
      "field": "licence",
      "scope": "resource",
      "method": "gate_observation",
      "evidence_ref": "gate#/licence/evidence"
    },
    {
      "field": "access_rights",
      "scope": "resource",
      "method": "source_metadata",
      "evidence_ref": "snapshot#/resource"
    },
    {
      "field": "representations",
      "scope": "resource",
      "method": "conversion",
      "evidence_ref": "conversion#/latexml-html"
    },
    {
      "field": "declared_type",
      "scope": "resource",
      "method": "source_metadata",
      "evidence_ref": "snapshot#/resource"
    },
    {
      "field": "educational",
      "scope": "resource",
      "method": "declared_type_mapping",
      "evidence_ref": "profile#/learning_resource_type/preprint"
    }
  ],
  "incidents": [
    {
      "code": "absent",
      "field": "rights_holder",
      "detail": "no rights holder is named at source; the licence is recorded without one rather than attributed to the platform that served it"
    },
    {
      "code": "absent",
      "field": "publisher",
      "detail": "the source named no publisher of the work; where it was collected from is recorded as collection provenance instead, which is a different claim"
    },
    {
      "code": "absent",
      "field": "educational",
      "detail": "the source declared no educational metadata — no resource type, audience, context, difficulty or learning time. Nothing here estimates them"
    }
  ],
  "inferred": [
    {
      "rule": "purpose.declared_type",
      "field": "purpose",
      "value": "reference",
      "method": "inferred",
      "evidence": {
        "declared_type": "preprint"
      },
      "rule_version": "purpose/1"
    },
    {
      "field": "references",
      "value": "https://doi.org/10.1093/biostatistics/kxy002",
      "label": "doi:10.1093/biostatistics/kxy002",
      "resource_id": "",
      "method": "inferred",
      "rule": "citation",
      "rule_version": "links/1",
      "evidence": {
        "links": [
          {
            "how": "citation",
            "line": 315,
            "basis": "reference",
            "literal": "doi:10.1093/biostatistics/kxy002"
          }
        ],
        "kind": "citation"
      }
    }
  ]
}