{
  "docId": "019dd923-5e88-73ef-bd5d-06b04d219fea",
  "docSlug": "dd91c78f6570bf29",
  "documentTitle": "2023 Air Street Capital The State of AI Report 2023",
  "authorId": "AirStreetCapital",
  "authorName": "Air Street Capital",
  "documentKindSlug": "consulting-deck",
  "documentKindLabel": "Consulting deck",
  "sourceTypeSlug": "vc_research",
  "sourceTypeLabel": "VC research",
  "presentationDate": null,
  "orientation": "landscape",
  "aspectRatio": 1.777,
  "pageNumber": 145,
  "pageCount": 163,
  "prevPage": 144,
  "nextPage": 146,
  "slideType": "industry_trends",
  "function": "summarize",
  "density": "dense",
  "nDataPoints": 0,
  "notes": "The slide highlights specific industry responses to AI safety concerns, featuring diagrams of Anthropic's ASL framework and a general model evaluation lifecycle.",
  "elementsJson": [
    "headline_text",
    "action_title",
    "bullet_list",
    "process_diagram",
    "other"
  ],
  "metadataConfidence": 1,
  "imagePath": null,
  "slideHref": "/slides/019dd923-5e88-73ef-bd5d-06b04d219fea/145",
  "deckHref": "/decks/019dd923-5e88-73ef-bd5d-06b04d219fea",
  "deckJsonHref": "/decks/019dd923-5e88-73ef-bd5d-06b04d219fea.json",
  "deckAnchorHref": "/decks/019dd923-5e88-73ef-bd5d-06b04d219fea#slide-145",
  "components": [
    {
      "bbox": null,
      "kind": "callout",
      "text": "While every reputable lab already has responsible development principles and evaluates risks around bias, toxicity, copyright infringement, and other common challenges - there are concerns these processes don't routinely address extreme risk.",
      "attrs": null,
      "subkind": null,
      "toolName": "Visual emphasis",
      "toolSlug": "visual-emphasis",
      "confidence": null,
      "componentId": "019dd952-47c4-7719-a22e-84f2f0a8ef28",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.35,
        "w": 0.48,
        "x": 0.46,
        "y": 0.6
      },
      "kind": "image",
      "text": "Model evaluation lifecycle diagram",
      "attrs": null,
      "subkind": "infographic",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "cfd246cf-84bc-4b0c-8187-bb73d601fa32",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.35,
        "w": 0.38,
        "x": 0.05,
        "y": 0.6
      },
      "kind": "image",
      "text": "Anthropic ASL safety buffer diagram",
      "attrs": null,
      "subkind": "infographic",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "d3d42fe0-e87f-4f7c-a44e-5456bb4cff07",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.08,
        "w": 0.95,
        "x": 0.02,
        "y": 0.32
      },
      "kind": "list",
      "text": "DeepMind have proposed a toolkit and associated workflow for extending standard model evaluations to assess for potentially dangerous capabilities (e.g. cyber-offense, self-proliferation) and propensity to cause harm.",
      "attrs": null,
      "subkind": "bullet",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "471dc779-e249-4a72-9cb1-574a5bfe6192",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.1,
        "w": 0.95,
        "x": 0.02,
        "y": 0.2
      },
      "kind": "list",
      "text": "While every reputable lab already has responsible development principles and evaluates risks around bias, toxicity, copyright infringement, and other common challenges - there are concerns these processes don't routinely address extreme risk.",
      "attrs": null,
      "subkind": "bullet",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "9bb93420-2db6-4eaf-9032-397bf4add3ef",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.12,
        "w": 0.95,
        "x": 0.02,
        "y": 0.42
      },
      "kind": "list",
      "text": "Anthropic released a new Responsible Scaling Policy, with a risk-based list of safety commitments, building in development breaks if safety measures fail to keep up with capabilities. The commitments cover internal access controls, red-teaming, third-party evaluations, and tiered access for different AI Safety Levels (ASLs).",
      "attrs": null,
      "subkind": "bullet",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "edbe8925-b4a4-4f34-94fb-90347eb0a567",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.04,
        "w": 0.7,
        "x": 0.02,
        "y": 0.14
      },
      "kind": "title",
      "text": "Amid the theoretical debate, labs are building in their own mitigations",
      "attrs": null,
      "subkind": "headline",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "83388dda-0f47-4d6f-96d4-bc921b1fbd7a",
      "frameworkName": null,
      "frameworkSlug": null
    }
  ],
  "metrics": [],
  "tools": [
    {
      "name": "Logical chain",
      "slug": "logical-chain",
      "agent": null,
      "layer": "slide",
      "matchId": "af30171c-fcf8-4547-974d-780c98cda27f",
      "evidence": "Anthropic released a new Responsible Scaling Policy, with a risk-based list of safety commitments, building in development breaks if safety measures fail to keep up with capabilities.",
      "confidence": 0.7
    }
  ],
  "frameworks": [
    {
      "name": "Responsible Scaling Policy",
      "slug": null,
      "matchId": "98710ef7-7c2b-47bb-b74d-e1bc4b5078e6",
      "evidence": "Anthropic's risk-based safety framework",
      "confidence": 0.9
    }
  ],
  "arcBeats": [
    {
      "to": 155,
      "from": 121,
      "beatId": "019dd95a-0682-776c-8e35-478c8c58938c",
      "arcName": "The Triple Take",
      "arcSlug": "triple-take",
      "beatName": "The Implications (So What)",
      "beatSlug": "triple-take-the-implications-so-what",
      "evidence": "Politics + Safety sections explore regulatory, geopolitical and risk implications.",
      "position": 2,
      "confidence": 78,
      "parentBeatName": "Reflection",
      "parentBeatSlug": "reflection"
    },
    {
      "to": 155,
      "from": 121,
      "beatId": "019dd95a-0682-776c-8e35-554ec1797b3a",
      "arcName": "The Mountain",
      "arcSlug": "mountain",
      "beatName": "Climax",
      "beatSlug": "mountain-climax",
      "evidence": "Regulatory divergence and the x-risk debate erupt into the mainstream.",
      "position": 3,
      "confidence": 45,
      "parentBeatName": "Turn",
      "parentBeatSlug": "turn"
    }
  ],
  "loops": [
    {
      "to": 148,
      "from": 139,
      "name": "Pattern Hunter",
      "slug": "02-pattern-hunter",
      "bestFor": "Time-pressed audiences, building consensus, when data is strong",
      "matchId": "019dd95a-07fe-70ce-8d3e-17c0a044d464",
      "evidence": "Hendrycks taxonomy, x-risk explosion, sceptics, govt attention, jailbreak evidence accumulate.",
      "position": 14,
      "objective": "Frame the safety debate, both x-risk and demonstrated unsafe behaviour",
      "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion",
      "confidence": 78,
      "description": "Group multiple pieces of evidence that together point to a pattern or conclusion"
    }
  ],
  "imagePathAlt": null,
  "thumbSrc": null,
  "thumbSrcAlt": null,
  "locked": true
}