Responsible Scaling Policies

{
  "id": "responsible-scaling-policies",
  "numericId": null,
  "path": "/knowledge-base/responses/responsible-scaling-policies/",
  "filePath": "knowledge-base/responses/responsible-scaling-policies.mdx",
  "title": "Responsible Scaling Policies",
  "quality": 64,
  "readerImportance": 63,
  "researchImportance": 28,
  "tacticalValue": null,
  "contentFormat": "article",
  "tractability": null,
  "neglectedness": null,
  "uncertainty": null,
  "causalLevel": null,
  "lastUpdated": "2026-03-13",
  "dateCreated": "2026-02-15",
  "llmSummary": "RSPs are voluntary industry frameworks that trigger safety evaluations at capability thresholds, currently covering 60-70% of frontier development across 3-4 major labs. Estimated 10-25% risk reduction is limited by 30-50% evaluation gaps, 0% external enforcement, and 20-60% abandonment risk under competitive pressure. ASL-3 activated for first time with Claude Opus 4 (30%+ bioweapon time reduction threshold); High threshold targets thousands of deaths or \\$100B+ damages.",
  "description": "Industry self-regulation frameworks establishing capability thresholds that trigger safety evaluations. Anthropic's ASL-3 requires 30%+ bioweapon development time reduction threshold; OpenAI's High threshold targets thousands of deaths or \\$100B+ damages. Current RSPs provide 10-25% estimated risk reduction across 60-70% of frontier development, limited by 0% external enforcement and 20-60% abandonment risk under competitive pressure.",
  "ratings": {
    "novelty": 4.5,
    "rigor": 6.8,
    "actionability": 7.2,
    "completeness": 7.5
  },
  "category": "responses",
  "subcategory": "industry",
  "clusters": [
    "ai-safety",
    "governance"
  ],
  "metrics": {
    "wordCount": 4506,
    "tableCount": 5,
    "diagramCount": 1,
    "internalLinks": 31,
    "externalLinks": 21,
    "footnoteCount": 0,
    "bulletRatio": 0.09,
    "sectionCount": 39,
    "hasOverview": true,
    "structuralScore": 15
  },
  "suggestedQuality": 100,
  "updateFrequency": 21,
  "evergreen": true,
  "wordCount": 4506,
  "unconvertedLinks": [
    {
      "text": "Introduced by Anthropic in September 2023",
      "url": "https://metr.org/blog/2023-09-26-rsp/",
      "resourceId": "73bedb360b0de6ae",
      "resourceTitle": "METR: Responsible Scaling Policies"
    },
    {
      "text": "UK AI Security Institute",
      "url": "https://www.aisi.gov.uk/blog/early-lessons-from-evaluating-frontier-ai-systems",
      "resourceId": "0fd3b1f5c81a37d8",
      "resourceTitle": "UK AI Security Institute's evaluations"
    },
    {
      "text": "activated ASL-3 protections for the first time",
      "url": "https://www.anthropic.com/news/activating-asl3-protections",
      "resourceId": "7512ddb574f82249"
    },
    {
      "text": "Preparedness Framework underwent significant revision in April 2025",
      "url": "https://openai.com/index/updating-our-preparedness-framework/",
      "resourceId": "ded0b05862511312",
      "resourceTitle": "Preparedness Framework"
    },
    {
      "text": "v1.0 (May 2024)",
      "url": "https://deepmind.google/blog/introducing-the-frontier-safety-framework/",
      "resourceId": "8c8edfbc52769d52",
      "resourceTitle": "Google DeepMind: Introducing the Frontier Safety Framework"
    },
    {
      "text": "v3.0 (September 2025)",
      "url": "https://deepmind.google/blog/strengthening-our-frontier-safety-framework/",
      "resourceId": "a5154ccbf034e273",
      "resourceTitle": "Google DeepMind: Strengthening our Frontier Safety Framework"
    },
    {
      "text": "UK AI Security Institute has evaluated over 30 frontier AI models",
      "url": "https://www.aisi.gov.uk/frontier-ai-trends-report",
      "resourceId": "7042c7f8de04ccb1",
      "resourceTitle": "AISI Frontier AI Trends"
    },
    {
      "text": "METR notes",
      "url": "https://metr.org/faisc",
      "resourceId": "7e3b7146e1266c71",
      "resourceTitle": "METR's analysis"
    },
    {
      "text": "v2.2 (May 2025)",
      "url": "https://www.anthropic.com/responsible-scaling-policy",
      "resourceId": "afe1e125f3ba3f14"
    },
    {
      "text": "v2 (April 2025)",
      "url": "https://cdn.openai.com/pdf/18a02b5d-6b67-4cec-ab64-68cdfbddebcd/preparedness-framework-v2.pdf",
      "resourceId": "ec5d8e7d6a1b2c7c",
      "resourceTitle": "OpenAI: Preparedness Framework Version 2"
    },
    {
      "text": "v3.0 (September 2025)",
      "url": "https://storage.googleapis.com/deepmind-media/DeepMind.com/Blog/strengthening-our-frontier-safety-framework/frontier-safety-framework_3.pdf",
      "resourceId": "3c56c8c2a799e4ef",
      "resourceTitle": "Google DeepMind: Frontier Safety Framework Version 3.0"
    },
    {
      "text": "AI Seoul Summit",
      "url": "https://www.gov.uk/government/publications/frontier-ai-safety-commitments-ai-seoul-summit-2024",
      "resourceId": "944fc2ac301f8980",
      "resourceTitle": "Seoul Frontier AI Commitments"
    },
    {
      "text": "METR data",
      "url": "https://metr.org/faisc",
      "resourceId": "7e3b7146e1266c71",
      "resourceTitle": "METR's analysis"
    }
  ],
  "unconvertedLinkCount": 13,
  "convertedLinkCount": 18,
  "backlinkCount": 49,
  "hallucinationRisk": {
    "level": "medium",
    "score": 55,
    "factors": [
      "no-citations"
    ]
  },
  "entityType": "policy",
  "redundancy": {
    "maxSimilarity": 24,
    "similarPages": [
      {
        "id": "metr",
        "title": "METR",
        "path": "/knowledge-base/organizations/metr/",
        "similarity": 24
      },
      {
        "id": "voluntary-commitments",
        "title": "Voluntary Industry Commitments",
        "path": "/knowledge-base/responses/voluntary-commitments/",
        "similarity": 24
      },
      {
        "id": "us-aisi",
        "title": "US AI Safety Institute",
        "path": "/knowledge-base/organizations/us-aisi/",
        "similarity": 22
      },
      {
        "id": "ai-safety-institutes",
        "title": "AI Safety Institutes",
        "path": "/knowledge-base/responses/ai-safety-institutes/",
        "similarity": 22
      },
      {
        "id": "thresholds",
        "title": "Compute Thresholds",
        "path": "/knowledge-base/responses/thresholds/",
        "similarity": 22
      }
    ]
  },
  "coverage": {
    "passing": 5,
    "total": 13,
    "targets": {
      "tables": 18,
      "diagrams": 2,
      "internalLinks": 36,
      "externalLinks": 23,
      "footnotes": 14,
      "references": 14
    },
    "actuals": {
      "tables": 5,
      "diagrams": 1,
      "internalLinks": 31,
      "externalLinks": 21,
      "footnotes": 0,
      "references": 23,
      "quotesWithQuotes": 0,
      "quotesTotal": 0,
      "accuracyChecked": 0,
      "accuracyTotal": 0
    },
    "items": {
      "llmSummary": "green",
      "schedule": "green",
      "entity": "green",
      "editHistory": "red",
      "overview": "green",
      "tables": "amber",
      "diagrams": "amber",
      "internalLinks": "amber",
      "externalLinks": "amber",
      "footnotes": "red",
      "references": "green",
      "quotes": "red",
      "accuracy": "red"
    },
    "ratingsString": "N:4.5 R:6.8 A:7.2 C:7.5"
  },
  "readerRank": 211,
  "researchRank": 431,
  "recommendedScore": 181.36
}

{
  "lesswrong": "https://www.lesswrong.com/tag/responsible-scaling-policies"
}