{
 "name": "What model providers' usage rules say about training on their output",
 "title": "What model providers' usage rules say about training on their output",
 "kind": "terms-board",
 "url": "https://sellagentlogstoai.com/datasets/agent-model-output-rules/",
 "csv": "https://sellagentlogstoai.com/datasets/agent-model-output-rules.csv",
 "markdown": "https://sellagentlogstoai.com/datasets/agent-model-output-rules/index.md",
 "computed_at": "2026-10-28",
 "license": "https://sellagentlogstoai.com/terms",
 "headline": "3 model providers' own usage documents restrict using what their models produce to train another AI model, each rule quoted from the provider's page, read 28 October 2026.",
 "quotes": [
  "3 model providers' own usage documents restrict using what their models produce to train another AI model, each rule quoted from the provider's page, read 28 October 2026.",
  "Using its outputs to train another AI model, by model provider: barred unless the provider authorises it first 1 of 3, barred where the model or product competes with the provider 1 of 3, varies 1 of 3 (3 model providers, read 28 October 2026).",
  "The rule, in its own words, by model provider: stated 3 of 3 (3 model providers, read 28 October 2026)."
 ],
 "citation": "\"What model providers' usage rules say about training on their output\", updated 2026-10-28, https://sellagentlogstoai.com/datasets/agent-model-output-rules/.",
 "national": {
  "n": 3,
  "unit": "model providers",
  "documents": 4,
  "published_sources": 3,
  "named_sources": 4
 },
 "board": {
  "type": "clauses",
  "noun": {
   "one": "model provider",
   "many": "model providers"
  },
  "document": {
   "one": "usage document",
   "many": "usage documents"
  },
  "fields": [
   {
    "id": "rule",
    "label": "Using its outputs to train another AI model",
    "counts": {
     "approval-needed": 1,
     "varies": 1,
     "competitors-barred": 1
    },
    "values": [
     {
      "value": "substitutes-barred",
      "text": "barred where the model trained is meant as an alternative to or substitute for the provider's own"
     },
     {
      "value": "approval-needed",
      "text": "barred unless the provider authorises it first"
     },
     {
      "value": "competitors-barred",
      "text": "barred where the model or product competes with the provider"
     }
    ]
   },
   {
    "id": "words",
    "label": "The rule, in its own words",
    "counts": {
     "stated": 3
    },
    "values": [
     {
      "value": "stated",
      "text": "stated"
     }
    ],
    "clause": true
   }
  ],
  "lead": {
   "type": "count",
   "field": null,
   "value": null,
   "n": 3,
   "of": 3,
   "label": "Model providers whose usage rules restrict training on their outputs"
  },
  "vendors": [
   {
    "name": "Anthropic",
    "answers": {
     "rule": {
      "value": "approval-needed",
      "text": "barred unless the provider authorises it first",
      "documents": [
       {
        "document": "usage policy",
        "url": "https://www.anthropic.com/legal/aup",
        "value": "approval-needed",
        "quote": "Use outputs to train an AI model (e.g., “model scraping” or “model distillation”) without prior authorization from Anthropic"
       }
      ]
     },
     "words": {
      "value": "stated",
      "text": "Use outputs to train an AI model (e.g., “model scraping” or “model distillation”) without…",
      "documents": [
       {
        "document": "usage policy",
        "url": "https://www.anthropic.com/legal/aup",
        "value": "stated",
        "quote": "Use outputs to train an AI model (e.g., “model scraping” or “model distillation”) without prior authorization from Anthropic"
       }
      ]
     }
    }
   },
   {
    "name": "OpenAI",
    "answers": {
     "rule": {
      "value": "varies",
      "text": "varies by document",
      "documents": [
       {
        "document": "usage policies",
        "url": "https://openai.com/policies/usage-policies/",
        "value": "substitutes-barred",
        "quote": "You must not use OpenAI models or their outputs to generate, prepare, select, or supply training material or signals for models that are intended as alternatives to or substitutes for OpenAI models or services."
       },
       {
        "document": "terms of use for the UK and Europe",
        "url": "https://openai.com/policies/eu-terms-of-use/",
        "value": "competitors-barred",
        "quote": "Using Output to develop models that compete with OpenAI."
       }
      ]
     },
     "words": {
      "value": "stated",
      "text": "These changes are effective November 20, 2026.",
      "documents": [
       {
        "document": "usage policies",
        "url": "https://openai.com/policies/usage-policies/",
        "value": "stated",
        "quote": "These changes are effective November 20, 2026."
       },
       {
        "document": "terms of use for the UK and Europe",
        "url": "https://openai.com/policies/eu-terms-of-use/",
        "value": "stated",
        "quote": "Using Output to develop models that compete with OpenAI."
       }
      ]
     }
    }
   },
   {
    "name": "xAI",
    "answers": {
     "rule": {
      "value": "competitors-barred",
      "text": "barred where the model or product competes with the provider",
      "documents": [
       {
        "document": "acceptable use policy",
        "url": "https://x.ai/legal/acceptable-use-policy",
        "value": "competitors-barred",
        "quote": "Using the Service or any Output to develop (or assist anyone in developing) machine learning models or any products or services that compete with SpaceXSI, whether directly or indirectly"
       }
      ]
     },
     "words": {
      "value": "stated",
      "text": "Scraping, harvesting or reselling any Input or Output, or distilling model data or Outputs",
      "documents": [
       {
        "document": "acceptable use policy",
        "url": "https://x.ai/legal/acceptable-use-policy",
        "value": "stated",
        "quote": "Scraping, harvesting or reselling any Input or Output, or distilling model data or Outputs"
       }
      ]
     }
    }
   }
  ]
 },
 "subject_kind": "model provider",
 "rows": [
  {
   "subject": "Anthropic",
   "url": "https://www.anthropic.com/legal/aup",
   "document": "usage policy",
   "question": "rule",
   "answer": "approval-needed",
   "answer_text": "barred unless the provider authorises it first",
   "status": "stated",
   "retrieved_at": "2026-10-28",
   "quote": "Use outputs to train an AI model (e.g., “model scraping” or “model distillation”) without prior authorization from Anthropic",
   "text_sha256": "9d340a0b9c85b43ad63fdf2395c10ab8f304b2e0663bbafa07213648976fac6f"
  },
  {
   "subject": "Anthropic",
   "url": "https://www.anthropic.com/legal/aup",
   "document": "usage policy",
   "question": "words",
   "answer": "stated",
   "answer_text": "Use outputs to train an AI model (e.g., “model scraping” or “model distillation”) without…",
   "status": "stated",
   "retrieved_at": "2026-10-28",
   "quote": "Use outputs to train an AI model (e.g., “model scraping” or “model distillation”) without prior authorization from Anthropic",
   "text_sha256": "9d340a0b9c85b43ad63fdf2395c10ab8f304b2e0663bbafa07213648976fac6f"
  },
  {
   "subject": "OpenAI",
   "url": "https://openai.com/policies/eu-terms-of-use/",
   "document": "terms of use for the UK and Europe",
   "question": "rule",
   "answer": "competitors-barred",
   "answer_text": "barred where the model or product competes with the provider",
   "status": "stated",
   "retrieved_at": "2026-10-28",
   "quote": "Using Output to develop models that compete with OpenAI.",
   "text_sha256": "e102b5f0611959ff6ebb38b4c4a809f65fa174419f6574596c02ccfe7b1a4490"
  },
  {
   "subject": "OpenAI",
   "url": "https://openai.com/policies/eu-terms-of-use/",
   "document": "terms of use for the UK and Europe",
   "question": "words",
   "answer": "stated",
   "answer_text": "Using Output to develop models that compete with OpenAI.",
   "status": "stated",
   "retrieved_at": "2026-10-28",
   "quote": "Using Output to develop models that compete with OpenAI.",
   "text_sha256": "e102b5f0611959ff6ebb38b4c4a809f65fa174419f6574596c02ccfe7b1a4490"
  },
  {
   "subject": "OpenAI",
   "url": "https://openai.com/policies/usage-policies/",
   "document": "usage policies",
   "question": "rule",
   "answer": "substitutes-barred",
   "answer_text": "barred where the model trained is meant as an alternative to or substitute for the provider's own",
   "status": "stated",
   "retrieved_at": "2026-10-28",
   "quote": "You must not use OpenAI models or their outputs to generate, prepare, select, or supply training material or signals for models that are intended as alternatives to or substitutes for OpenAI models or services.",
   "text_sha256": "86eacb8f6b4fab162ca8e043dce0983bb10df4d48c3b538aab8228e32325dd5d"
  },
  {
   "subject": "OpenAI",
   "url": "https://openai.com/policies/usage-policies/",
   "document": "usage policies",
   "question": "words",
   "answer": "stated",
   "answer_text": "These changes are effective November 20, 2026.",
   "status": "stated",
   "retrieved_at": "2026-10-28",
   "quote": "These changes are effective November 20, 2026.",
   "text_sha256": "86eacb8f6b4fab162ca8e043dce0983bb10df4d48c3b538aab8228e32325dd5d"
  },
  {
   "subject": "xAI",
   "url": "https://x.ai/legal/acceptable-use-policy",
   "document": "acceptable use policy",
   "question": "rule",
   "answer": "competitors-barred",
   "answer_text": "barred where the model or product competes with the provider",
   "status": "stated",
   "retrieved_at": "2026-10-28",
   "quote": "Using the Service or any Output to develop (or assist anyone in developing) machine learning models or any products or services that compete with SpaceXSI, whether directly or indirectly",
   "text_sha256": "5a95a50f683e0baad771e51e95462e5b978a31b6313908afcc9ee49c86f1389b"
  },
  {
   "subject": "xAI",
   "url": "https://x.ai/legal/acceptable-use-policy",
   "document": "acceptable use policy",
   "question": "words",
   "answer": "stated",
   "answer_text": "Scraping, harvesting or reselling any Input or Output, or distilling model data or Outputs",
   "status": "stated",
   "retrieved_at": "2026-10-28",
   "quote": "Scraping, harvesting or reselling any Input or Output, or distilling model data or Outputs",
   "text_sha256": "5a95a50f683e0baad771e51e95462e5b978a31b6313908afcc9ee49c86f1389b"
  }
 ],
 "sources": [
  {
   "url": "https://openai.com/policies/usage-policies/",
   "label": "OpenAI, usage policies",
   "license": "from the publisher's own model providers' own usage policies, acceptable use policies and regional terms of use; reproduced for comment",
   "class": "primary-provider-own-site",
   "status": "read"
  },
  {
   "url": "https://openai.com/policies/eu-terms-of-use/",
   "label": "OpenAI, terms of use for the UK and Europe",
   "license": "from the publisher's own model providers' own usage policies, acceptable use policies and regional terms of use; reproduced for comment",
   "class": "primary-provider-own-site",
   "status": "read"
  },
  {
   "url": "https://www.anthropic.com/legal/aup",
   "label": "Anthropic, usage policy",
   "license": "from the publisher's own model providers' own usage policies, acceptable use policies and regional terms of use; reproduced for comment",
   "class": "primary-provider-own-site",
   "status": "read"
  },
  {
   "url": "https://x.ai/legal/acceptable-use-policy",
   "label": "xAI, acceptable use policy",
   "license": "from the publisher's own model providers' own usage policies, acceptable use policies and regional terms of use; reproduced for comment",
   "class": "primary-provider-own-site",
   "status": "read"
  }
 ],
 "methodology": "Each model provider's own model providers' own usage policies, acceptable use policies and regional terms of use were read through estate-fetch and stored whole (sha256 on every row). For every usage document, two questions were answered from its own words: using its outputs to train another AI model, the rule, in its own words. Every stated answer carries the sentence it rests on, checked against the stored text; 'not stated' means the usage document read does not address the question. A model provider with several usage documents reads 'varies' where they disagree. 3 model providers, 4 usage documents.",
 "embed": "<figure style=\"margin:0;padding:16px;border:1px solid #d0d0d0;background:#ffffff\"><blockquote style=\"margin:0 0 8px;background:#ffffff\">3 model providers' own usage documents restrict using what their models produce to train another AI model, each rule quoted from the provider's page, read 28 October 2026.</blockquote><figcaption style=\"background:#ffffff\"><a href=\"https://sellagentlogstoai.com/datasets/agent-model-output-rules/\" style=\"background:#ffffff\">What model providers' usage rules say about training on their output, sellagentlogstoai.com</a></figcaption></figure>"
}