AI-Powered Investigation Risks

{
  "id": "ai-investigation-risks",
  "numericId": null,
  "path": "/knowledge-base/risks/ai-investigation-risks/",
  "filePath": "knowledge-base/risks/ai-investigation-risks.mdx",
  "title": "AI-Powered Investigation Risks",
  "quality": 40,
  "readerImportance": 6.5,
  "researchImportance": 6.5,
  "tacticalValue": null,
  "contentFormat": "article",
  "tractability": null,
  "neglectedness": null,
  "uncertainty": null,
  "causalLevel": "amplifier",
  "lastUpdated": "2026-03-13",
  "dateCreated": "2026-02-15",
  "llmSummary": "Analysis of AI-powered investigation as a dual-use capability. AI dramatically lowers the discoverability threshold for connecting public information, benefiting accountability (corruption detection, fraud investigation, investigative journalism) while threatening privacy through automated deanonymization and erosion of privacy through obscurity. Documents real-world examples including Bellingcat OSINT investigations, UK SFO analyzing 30M documents, and deanonymization of Netflix Prize data. GPT-4 achieves 80-94% face verification accuracy with zero training; Pew finds 57% of Americans say AI's societal risks outweigh benefits.",
  "description": "AI dramatically lowers the bar for investigative discovery, connecting publicly available information at scale to surface secrets that previously required serious detective work. Beneficial for exposing corruption and fraud (21 U.S. agencies use AI for anticorruption; UK SFO analyzed 30M documents via AI), but threatening for privacy as \"privacy through obscurity\" erodes. LLMs automate deanonymization, GPT-4 achieves 80-94% face verification accuracy with zero training, and the mosaic effect enables identification from individually innocuous data points. 57% of consumers say AI's societal risks outweigh benefits.",
  "ratings": {
    "novelty": 7,
    "rigor": 5.5,
    "actionability": 6,
    "completeness": 5.5
  },
  "category": "risks",
  "subcategory": "misuse",
  "clusters": [
    "ai-safety",
    "governance",
    "cyber"
  ],
  "metrics": {
    "wordCount": 2333,
    "tableCount": 5,
    "diagramCount": 1,
    "internalLinks": 11,
    "externalLinks": 32,
    "footnoteCount": 0,
    "bulletRatio": 0.29,
    "sectionCount": 26,
    "hasOverview": true,
    "structuralScore": 15
  },
  "suggestedQuality": 100,
  "updateFrequency": 30,
  "evergreen": true,
  "wordCount": 2333,
  "unconvertedLinks": [
    {
      "text": "Bellingcat",
      "url": "https://www.bellingcat.com/",
      "resourceId": "9c6f6a2ea461bc08",
      "resourceTitle": "Bellingcat: Open source investigation"
    },
    {
      "text": "Russia's involvement in the MH17 downing",
      "url": "https://www.bellingcat.com/",
      "resourceId": "9c6f6a2ea461bc08",
      "resourceTitle": "Bellingcat: Open source investigation"
    },
    {
      "text": "97 of 179 countries",
      "url": "https://carnegieendowment.org/features/ai-global-surveillance-technology",
      "resourceId": "c2d3c5e8ef0a4b0b",
      "resourceTitle": "Carnegie Endowment AI Global Surveillance Index"
    }
  ],
  "unconvertedLinkCount": 3,
  "convertedLinkCount": 0,
  "backlinkCount": 3,
  "hallucinationRisk": {
    "level": "medium",
    "score": 55,
    "factors": [
      "no-citations"
    ]
  },
  "entityType": "risk",
  "redundancy": {
    "maxSimilarity": 18,
    "similarPages": [
      {
        "id": "ai-powered-investigation",
        "title": "AI-Powered Investigation",
        "path": "/knowledge-base/capabilities/ai-powered-investigation/",
        "similarity": 18
      },
      {
        "id": "ai-accountability",
        "title": "AI for Accountability and Anti-Corruption",
        "path": "/knowledge-base/responses/ai-accountability/",
        "similarity": 16
      },
      {
        "id": "deanonymization",
        "title": "AI-Powered Deanonymization",
        "path": "/knowledge-base/risks/deanonymization/",
        "similarity": 16
      },
      {
        "id": "ai-enabled-untraceable-misuse",
        "title": "AI-Enabled Untraceable Misuse",
        "path": "/knowledge-base/risks/ai-enabled-untraceable-misuse/",
        "similarity": 15
      },
      {
        "id": "surveillance",
        "title": "Mass Surveillance",
        "path": "/knowledge-base/risks/surveillance/",
        "similarity": 15
      }
    ]
  },
  "changeHistory": [
    {
      "date": "2026-02-16",
      "branch": "claude/review-pr-followup-sv4RY",
      "title": "PR follow-up review and fixes",
      "summary": "Reviewed last 5 days of PRs (Feb 11-16) for remaining work. Fixed three issues: corrected quality metrics on ea-shareholder-diversification-anthropic (was 3/100, now 60/100), added cross-reference notes between four overlapping AI investigation pages (ai-investigation-risks, ai-powered-investigation, deanonymization, ai-accountability), and updated Anthropic Investors TODOs with research findings on matching program and Tallinn holdings plus refreshed secondary market prices to Feb 2026.",
      "pr": 169
    },
    {
      "date": "2026-02-15",
      "branch": "claude/add-ai-investigation-risks-DKQ6P",
      "title": "Add AI-Powered Investigation Risks page",
      "summary": "Created a new wiki page covering AI-powered investigation as a dual-use risk — how AI lowers the discoverability threshold for connecting public information, benefiting accountability (corruption detection, OSINT journalism) while threatening privacy through automated deanonymization and erosion of \"privacy through obscurity.\" Added entity definition E694.",
      "pr": 143
    }
  ],
  "coverage": {
    "passing": 7,
    "total": 13,
    "targets": {
      "tables": 9,
      "diagrams": 1,
      "internalLinks": 19,
      "externalLinks": 12,
      "footnotes": 7,
      "references": 7
    },
    "actuals": {
      "tables": 5,
      "diagrams": 1,
      "internalLinks": 11,
      "externalLinks": 32,
      "footnotes": 0,
      "references": 2,
      "quotesWithQuotes": 0,
      "quotesTotal": 0,
      "accuracyChecked": 0,
      "accuracyTotal": 0
    },
    "items": {
      "llmSummary": "green",
      "schedule": "green",
      "entity": "green",
      "editHistory": "green",
      "overview": "green",
      "tables": "amber",
      "diagrams": "green",
      "internalLinks": "amber",
      "externalLinks": "green",
      "footnotes": "red",
      "references": "amber",
      "quotes": "red",
      "accuracy": "red"
    },
    "editHistoryCount": 2,
    "ratingsString": "N:7 R:5.5 A:6 C:5.5"
  },
  "readerRank": 608,
  "researchRank": 578,
  "recommendedScore": 104.97
}

No external links