Technical AI Safety Research

{
  "id": "technical-research",
  "wikiId": "E297",
  "path": "/knowledge-base/responses/technical-research/",
  "filePath": "knowledge-base/responses/technical-research.mdx",
  "title": "Technical AI Safety Research",
  "quality": 66,
  "readerImportance": 85.5,
  "researchImportance": 24,
  "tacticalValue": 65,
  "contentFormat": "article",
  "causalLevel": null,
  "lastUpdated": "2026-01-29",
  "dateCreated": "2026-02-15",
  "summary": "Technical AI safety research encompasses six major agendas (mechanistic interpretability, scalable oversight, AI control, evaluations, agent foundations, and robustness) with 500+ researchers and \\$110-130M annual funding. Key 2024-2025 findings include tens of millions of interpretable features identified in Claude 3, 5 of 6 frontier models showing scheming capabilities, and deliberative alignment reducing scheming by up to 30x, though experts estimate only 2-50% x-risk reduction depending on timeline assumptions and technical tractability.",
  "description": "Technical AI safety research aims to make AI systems reliably safe through scientific and engineering work.",
  "ratings": {
    "novelty": 4.2,
    "rigor": 6.8,
    "completeness": 7.5,
    "actionability": 7.1
  },
  "category": "responses",
  "subcategory": "alignment",
  "clusters": [
    "ai-safety",
    "governance"
  ],
  "metrics": {
    "wordCount": 3765,
    "tableCount": 11,
    "diagramCount": 1,
    "internalLinks": 69,
    "externalLinks": 29,
    "footnoteCount": 0,
    "bulletRatio": 0.39,
    "sectionCount": 45,
    "hasOverview": true,
    "structuralScore": 14
  },
  "suggestedQuality": 93,
  "updateFrequency": 21,
  "evergreen": true,
  "wordCount": 3765,
  "unconvertedLinks": [
    {
      "text": "UK AISI",
      "url": "https://www.aisi.gov.uk/",
      "resourceId": "fdf68a8f30f57dee",
      "resourceTitle": "UK AI Safety Institute (AISI)"
    },
    {
      "text": "METR",
      "url": "https://metr.org/",
      "resourceId": "45370a5153534152",
      "resourceTitle": "METR: Model Evaluation and Threat Research"
    },
    {
      "text": "frontier AI safety policies",
      "url": "https://metr.org/blog/2025-03-26-common-elements-of-frontier-ai-safety-policies/",
      "resourceId": "a37628e3a1e97778",
      "resourceTitle": "Common Elements of Frontier AI Safety Policies (METR Analysis)"
    },
    {
      "text": "Anthropic's May 2024 \"Scaling Monosemanticity\"",
      "url": "https://transformer-circuits.pub/2024/scaling-monosemanticity/",
      "resourceId": "e724db341d6e0065",
      "resourceTitle": "Scaling Monosemanticity: Extracting Interpretable Features from Claude 3 Sonnet"
    },
    {
      "text": "UK AI Security Institute",
      "url": "https://www.aisi.gov.uk/frontier-ai-trends-report",
      "resourceId": "7042c7f8de04ccb1",
      "resourceTitle": "AISI Frontier AI Trends"
    },
    {
      "text": "Anthropic Transformer Circuits",
      "url": "https://transformer-circuits.pub/2024/scaling-monosemanticity/",
      "resourceId": "e724db341d6e0065",
      "resourceTitle": "Scaling Monosemanticity: Extracting Interpretable Features from Claude 3 Sonnet"
    },
    {
      "text": "OpenAI-Apollo Collaboration",
      "url": "https://openai.com/index/detecting-and-reducing-scheming-in-ai-models/",
      "resourceId": "b3f335edccfc5333",
      "resourceTitle": "OpenAI Preparedness Framework"
    },
    {
      "text": "UK AISI Frontier Trends Report",
      "url": "https://www.aisi.gov.uk/frontier-ai-trends-report",
      "resourceId": "7042c7f8de04ccb1",
      "resourceTitle": "AISI Frontier AI Trends"
    },
    {
      "text": "UK AISI Evaluations",
      "url": "https://www.aisi.gov.uk/blog/early-lessons-from-evaluating-frontier-ai-systems",
      "resourceId": "0fd3b1f5c81a37d8",
      "resourceTitle": "UK AI Security Institute's evaluations"
    },
    {
      "text": "OpenAI o1 System Card",
      "url": "https://openai.com/",
      "resourceId": "04d39e8bd5d50dd5",
      "resourceTitle": "OpenAI Official Homepage"
    },
    {
      "text": "UK AI Security Institute",
      "url": "https://www.aisi.gov.uk/",
      "resourceId": "fdf68a8f30f57dee",
      "resourceTitle": "UK AI Safety Institute (AISI)"
    },
    {
      "text": "US AI Safety Institute",
      "url": "https://www.nist.gov/aisi",
      "resourceId": "84e0da6d5092e27d",
      "resourceTitle": "Center for AI Standards and Innovation (CAISI)"
    },
    {
      "text": "METR",
      "url": "https://metr.org/",
      "resourceId": "45370a5153534152",
      "resourceTitle": "METR: Model Evaluation and Threat Research"
    },
    {
      "text": "Apollo Research",
      "url": "https://www.apolloresearch.ai/",
      "resourceId": "329d8c2e2532be3d",
      "resourceTitle": "Apollo Research - AI Safety Evaluation Organization"
    },
    {
      "text": "Redwood Research",
      "url": "https://www.redwoodresearch.org/",
      "resourceId": "42e7247cbc33fc4c",
      "resourceTitle": "Redwood Research: AI Control"
    },
    {
      "text": "Jaan Tallinn",
      "url": "https://quickmarketpitch.com/blogs/news/ai-safety-investors",
      "resourceId": "58d4f33d9ae57621",
      "resourceTitle": "Who is funding AI safety research?"
    },
    {
      "text": "Eric Schmidt / Schmidt Sciences",
      "url": "https://quickmarketpitch.com/blogs/news/ai-safety-investors",
      "resourceId": "58d4f33d9ae57621",
      "resourceTitle": "Who is funding AI safety research?"
    },
    {
      "text": "UK Government (AISI)",
      "url": "https://www.aisi.gov.uk/",
      "resourceId": "fdf68a8f30f57dee",
      "resourceTitle": "UK AI Safety Institute (AISI)"
    },
    {
      "text": "representing under 2% of estimated capabilities spending",
      "url": "https://www.lesswrong.com/posts/WGpFFJo2uFe5ssgEb/an-overview-of-the-ai-safety-funding-situation",
      "resourceId": "b1ab921f9cbae109",
      "resourceTitle": "An Overview of the AI Safety Funding Situation"
    },
    {
      "text": "cost-prohibitive for full coverage",
      "url": "https://transformer-circuits.pub/2024/scaling-monosemanticity/",
      "resourceId": "e724db341d6e0065",
      "resourceTitle": "Scaling Monosemanticity: Extracting Interpretable Features from Claude 3 Sonnet"
    },
    {
      "text": "Redwood's protocols",
      "url": "https://www.redwoodresearch.org/research/ai-control",
      "resourceId": "eb2318c5e3fc0f88",
      "resourceTitle": "Redwood Research, 2024"
    },
    {
      "text": "UK AISI tested 30+ models",
      "url": "https://www.aisi.gov.uk/",
      "resourceId": "fdf68a8f30f57dee",
      "resourceTitle": "UK AI Safety Institute (AISI)"
    },
    {
      "text": "o1 process supervision deployed",
      "url": "https://openai.com/",
      "resourceId": "04d39e8bd5d50dd5",
      "resourceTitle": "OpenAI Official Homepage"
    },
    {
      "text": "doubling time ≈7 months for autonomy",
      "url": "https://metr.org/",
      "resourceId": "45370a5153534152",
      "resourceTitle": "METR: Model Evaluation and Threat Research"
    }
  ],
  "unconvertedLinkCount": 24,
  "convertedLinkCount": 52,
  "backlinkCount": 14,
  "hallucinationRisk": {
    "level": "medium",
    "score": 45,
    "factors": [
      "no-citations",
      "conceptual-content"
    ]
  },
  "entityType": "crux",
  "redundancy": {
    "maxSimilarity": 21,
    "similarPages": [
      {
        "id": "research-agendas",
        "title": "AI Alignment Research Agenda Comparison",
        "path": "/knowledge-base/responses/research-agendas/",
        "similarity": 21
      },
      {
        "id": "intervention-effectiveness-matrix",
        "title": "Intervention Effectiveness Matrix",
        "path": "/knowledge-base/models/intervention-effectiveness-matrix/",
        "similarity": 19
      },
      {
        "id": "anthropic-core-views",
        "title": "Anthropic Core Views",
        "path": "/knowledge-base/responses/anthropic-core-views/",
        "similarity": 19
      },
      {
        "id": "interpretability",
        "title": "Mechanistic Interpretability",
        "path": "/knowledge-base/responses/interpretability/",
        "similarity": 19
      },
      {
        "id": "scalable-oversight",
        "title": "Scalable Oversight",
        "path": "/knowledge-base/responses/scalable-oversight/",
        "similarity": 19
      }
    ]
  },
  "coverage": {
    "passing": 7,
    "total": 13,
    "targets": {
      "tables": 15,
      "diagrams": 2,
      "internalLinks": 30,
      "externalLinks": 19,
      "footnotes": 11,
      "references": 11
    },
    "actuals": {
      "tables": 11,
      "diagrams": 1,
      "internalLinks": 69,
      "externalLinks": 29,
      "footnotes": 0,
      "references": 40,
      "quotesWithQuotes": 0,
      "quotesTotal": 0,
      "accuracyChecked": 0,
      "accuracyTotal": 0
    },
    "items": {
      "summary": "green",
      "schedule": "green",
      "entity": "green",
      "editHistory": "red",
      "overview": "green",
      "tables": "amber",
      "diagrams": "amber",
      "internalLinks": "green",
      "externalLinks": "green",
      "footnotes": "red",
      "references": "green",
      "quotes": "red",
      "accuracy": "red"
    },
    "ratingsString": "N:4.2 R:6.8 A:7.1 C:7.5"
  },
  "readerRank": 46,
  "researchRank": 454,
  "recommendedScore": 178.91
}

{
  "eightyK": "https://80000hours.org/career-reviews/ai-safety-researcher/"
}