{
  "api_version": "1",
  "updated": "2026-11-05",
  "latest_event": "2026-11-04",
  "rights": "© Frontier Log, all rights reserved.",
  "story": {
    "slug": "ai-text-watermarking",
    "title": "How AI Labs Watermark Text",
    "description": "Google, Anthropic, OpenAI and xAI each draw the line differently on marking AI-written text: by app, by API, by region and by who may check for the mark.",
    "url": "https://frontierlog.ai/stories/ai-text-watermarking",
    "date": "2026-11-01",
    "image": "https://frontierlog.ai/og/story-ai-text-watermarking.png",
    "event": null,
    "tags": [
      "policy",
      "safety"
    ],
    "steps": [],
    "tables": [],
    "records": [
      {
        "date": "2024-05-14",
        "party": "Google",
        "what": " marks Gemini app and web text with SynthID",
        "by": "Google DeepMind",
        "source": "https://deepmind.google/blog/watermarking-ai-generated-text-and-video-with-synthid/",
        "key": true
      },
      {
        "date": "2025-05-20",
        "party": "Google",
        "what": " opens the SynthID Detector to early testers and a waitlist",
        "by": "Google",
        "source": "https://blog.google/innovation-and-ai/products/google-synthid-ai-content-detector/"
      },
      {
        "date": "2026-07-31",
        "party": "The EU",
        "what": " says about 190 organizations signed the Code of Practice by the end of July",
        "by": "European Commission",
        "source": "https://digital-strategy.ec.europa.eu/en/news/strong-backing-code-practice-transparency-ai-generated-content"
      },
      {
        "date": "2026-08-02",
        "party": "The EU",
        "what": "'s Article 50 marking obligations apply",
        "by": "European Commission",
        "source": "https://digital-strategy.ec.europa.eu/en/policies/code-practice-ai-generated-content",
        "quote": "These transparency obligations, applicable from 2 August 2026",
        "key": true
      },
      {
        "date": "2026-08-14",
        "party": "Anthropic",
        "what": " says Claude models will watermark text, globally",
        "by": "Anthropic",
        "source": "https://www.anthropic.com/news/claude-text-watermark"
      },
      {
        "date": "2026-08-19",
        "party": "Google",
        "what": " staff say Gemini API text is SynthID-watermarked, correcting an Aug 5 reply",
        "by": "Google",
        "source": "https://discuss.ai.google.dev/t/does-gemini-api-text-output-carry-synthid-watermarking-gemini-2-5-flash-lite-gemini-3-1-flash-lite-eu-ai-act-art-50-2/177241/4"
      },
      {
        "date": "2026-09-01",
        "party": "Anthropic",
        "what": " launches Claude Fable 5.1 and Mythos 5.1, saying models released after Aug 2 carry the text watermark",
        "by": "Anthropic",
        "source": "https://www.anthropic.com/claude-fable-and-mythos-5-1"
      },
      {
        "date": "2026-10-05",
        "party": "OpenAI",
        "what": " opens opt-in API watermarking and plans EU-only marking in ChatGPT and Codex",
        "by": "OpenAI",
        "source": "https://openai.com/index/eu-text-provenance"
      },
      {
        "date": "2026-11-07",
        "party": "xAI",
        "what": " turns on watermarking by default for seven Grok API models",
        "by": "xAI",
        "source": "https://docs.x.ai/developers/advanced-api-usage/watermarking",
        "key": true
      },
      {
        "date": "2026-12-02",
        "party": "The EU",
        "what": "'s transitional period ends for AI systems placed on the market before Aug 2",
        "by": "European Commission",
        "source": "https://digital-strategy.ec.europa.eu/en/faqs/code-practice-transparency-ai-generated-content"
      }
    ],
    "statements": [
      {
        "party": "Google",
        "by": "Google DeepMind",
        "date": "2024-05-14",
        "source": "https://deepmind.google/blog/watermarking-ai-generated-text-and-video-with-synthid/",
        "quote": "Today, we’re expanding SynthID’s capabilities to watermarking AI-generated text in the Gemini app and web experience, and video in Veo, our most capable generative video model."
      },
      {
        "party": "Google",
        "by": "Google",
        "date": "2025-05-20",
        "source": "https://blog.google/innovation-and-ai/products/google-synthid-ai-content-detector/",
        "quote": "We are starting to roll this out to early testers today, before making it available more broadly. Journalists, media professionals and researchers can join our waitlist to gain access to the SynthID Detector."
      },
      {
        "party": "Google",
        "by": "Google",
        "date": "2026-08-19",
        "source": "https://discuss.ai.google.dev/t/does-gemini-api-text-output-carry-synthid-watermarking-gemini-2-5-flash-lite-gemini-3-1-flash-lite-eu-ai-act-art-50-2/177241/4",
        "quote": "After checking further with the team, it turns out that text generated via the Gemini API IS actually SynthID-watermarked! I apologize for the confusion earlier — the previous statement was incorrect. This also applies to text generated through Google AI Studio and/or antigravity!"
      },
      {
        "party": "Google",
        "by": "Google DeepMind",
        "source": "https://deepmind.google/models/synthid/",
        "quote": "With the new SynthID Detector, anyone can easily check if an image, video, or audio file was made with AI from Google or our partners, including OpenAI, NVIDIA, Kakao, and soon, Apple."
      },
      {
        "party": "Anthropic",
        "by": "Anthropic",
        "source": "https://support.claude.com/en/articles/16266773",
        "quote": "Marks will apply to output from supported Claude models across Claude Platform (API), Claude, Claude Code, Claude Cowork, and Claude Tag, and wherever Claude is offered, worldwide."
      },
      {
        "party": "Anthropic",
        "by": "Anthropic",
        "source": "https://support.claude.com/en/articles/16266773",
        "quote": "Claude models launched in the EU on or after August 2, 2026 will support machine-readable marking at launch."
      },
      {
        "party": "Anthropic",
        "by": "Anthropic",
        "date": "2026-08-14",
        "source": "https://www.anthropic.com/news/claude-text-watermark",
        "quote": "We’re applying watermarking globally at launch because we don't yet have a durable way to scope it by region."
      },
      {
        "party": "Anthropic",
        "by": "Anthropic",
        "source": "https://support.claude.com/en/articles/16266773",
        "quote": "Watermark detection is currently in private preview, available to eligible organizations as required under EU law (such as regulators, law enforcement, media, fact-checkers, independent researchers, educational organizations, and EU civil society groups)."
      },
      {
        "party": "Anthropic",
        "by": "Anthropic",
        "date": "2026-09-01",
        "source": "https://www.anthropic.com/claude-fable-and-mythos-5-1",
        "quote": "This required us to add a watermark—a numerical way of determining the likelihood that Claude was involved in writing a piece of text—to the outputs of models released after August 2, 2026."
      },
      {
        "party": "Anthropic",
        "by": "Anthropic",
        "date": "2026-09-01",
        "source": "https://platform.claude.com/docs/en/release-notes/overview",
        "quote": "We’ve launched Claude Fable 5.1"
      },
      {
        "party": "OpenAI",
        "by": "OpenAI",
        "date": "2026-10-05",
        "source": "https://openai.com/index/eu-text-provenance",
        "quote": "Starting today, API customers globally will be able to opt in to text watermarking for select models. Text watermarking will remain off by default in the API."
      },
      {
        "party": "OpenAI",
        "by": "OpenAI",
        "date": "2026-10-05",
        "source": "https://openai.com/index/eu-text-provenance",
        "quote": "Over the coming weeks, we will introduce text watermarking to eligible ChatGPT and Codex users across all plans in the EU only. We are not making text watermarking a global default at launch."
      },
      {
        "party": "OpenAI",
        "by": "OpenAI",
        "date": "2026-10-05",
        "source": "https://openai.com/index/eu-text-provenance",
        "quote": "We’re opening applications to access our text watermark detector. Access will initially be limited to approved researchers and expert organizations that can help us evaluate and improve the technology."
      },
      {
        "party": "xAI",
        "by": "xAI",
        "source": "https://docs.x.ai/developers/advanced-api-usage/watermarking",
        "quote": "Starting November 7, 2026, responses from these models include a watermark. It's on by default, and you can opt out per model."
      },
      {
        "party": "xAI",
        "by": "xAI",
        "source": "https://docs.x.ai/developers/advanced-api-usage/watermarking",
        "quote": "The watermark is enabled on every SpaceXSI API endpoint, including the global endpoint"
      },
      {
        "party": "xAI",
        "by": "xAI",
        "source": "https://docs.x.ai/developers/advanced-api-usage/watermarking",
        "quote": "Opting out asks you to confirm that an exception under applicable law applies to your use, and that you are solely responsible for any marking, disclosure, or other legal requirement."
      },
      {
        "party": "xAI",
        "by": "xAI",
        "source": "https://docs.x.ai/developers/advanced-api-usage/watermarking",
        "quote": "Access to watermark detection is granted by application."
      },
      {
        "party": "xAI",
        "by": "xAI",
        "source": "https://docs.x.ai/grok/faq",
        "quote": "Generated images and videos include a Grok watermark to indicate that the content was created with AI. There is no setting to remove the watermark."
      },
      {
        "party": "European Commission",
        "by": "European Commission",
        "source": "https://digital-strategy.ec.europa.eu/en/faqs/code-practice-transparency-ai-generated-content",
        "quote": "There is an exception for AI systems placed on the market before that date. These AI systems benefit from a transitional period for compliance until 2 December 2026."
      }
    ],
    "cites": [
      {
        "url": "https://docs.x.ai/developers/advanced-api-usage/watermarking",
        "by": "xAI",
        "medium": "API documentation",
        "date": "2026-10-31"
      },
      {
        "url": "https://help.openai.com/en/articles/8912793",
        "by": "OpenAI",
        "medium": "help center article"
      },
      {
        "url": "https://support.claude.com/en/articles/16266773",
        "by": "Anthropic",
        "medium": "help center article"
      },
      {
        "url": "https://digital-strategy.ec.europa.eu/en/policies/code-practice-ai-generated-content",
        "by": "European Commission",
        "medium": "policy page",
        "date": "2026-10-19"
      },
      {
        "url": "https://digital-strategy.ec.europa.eu/en/news/strong-backing-code-practice-transparency-ai-generated-content",
        "by": "European Commission",
        "medium": "news article",
        "date": "2026-07-31"
      },
      {
        "url": "https://digital-strategy.ec.europa.eu/en/faqs/code-practice-transparency-ai-generated-content",
        "by": "European Commission",
        "medium": "FAQ",
        "date": "2026-06-10"
      },
      {
        "url": "https://openai.com/index/eu-text-provenance",
        "by": "OpenAI",
        "medium": "blog post",
        "date": "2026-10-05"
      },
      {
        "url": "https://www.anthropic.com/news/claude-text-watermark",
        "by": "Anthropic",
        "medium": "blog post",
        "date": "2026-08-14"
      },
      {
        "url": "https://deepmind.google/blog/watermarking-ai-generated-text-and-video-with-synthid/",
        "by": "Google DeepMind",
        "medium": "blog post",
        "date": "2024-05-14"
      },
      {
        "url": "https://blog.google/innovation-and-ai/products/google-synthid-ai-content-detector/",
        "by": "Google",
        "medium": "blog post",
        "date": "2025-05-20"
      },
      {
        "url": "https://discuss.ai.google.dev/t/does-gemini-api-text-output-carry-synthid-watermarking-gemini-2-5-flash-lite-gemini-3-1-flash-lite-eu-ai-act-art-50-2/177241/4",
        "by": "Google",
        "medium": "staff reply on its developer forum",
        "date": "2026-08-19"
      },
      {
        "url": "https://discuss.ai.google.dev/t/does-gemini-api-text-output-carry-synthid-watermarking-gemini-2-5-flash-lite-gemini-3-1-flash-lite-eu-ai-act-art-50-2/177241/2",
        "by": "Google",
        "medium": "staff reply on its developer forum",
        "date": "2026-08-05"
      },
      {
        "url": "https://deepmind.google/models/synthid/",
        "by": "Google DeepMind",
        "medium": "product page"
      },
      {
        "url": "https://docs.x.ai/grok/faq",
        "by": "xAI",
        "medium": "Grok FAQ"
      },
      {
        "url": "https://www.anthropic.com/claude-fable-and-mythos-5-1",
        "by": "Anthropic",
        "medium": "blog post",
        "date": "2026-09-01"
      },
      {
        "url": "https://platform.claude.com/docs/en/release-notes/overview",
        "by": "Anthropic",
        "medium": "API release notes"
      },
      {
        "url": "https://discuss.ai.google.dev/t/does-gemini-api-text-output-carry-synthid-watermarking-gemini-2-5-flash-lite-gemini-3-1-flash-lite-eu-ai-act-art-50-2/177241/5",
        "by": "a developer",
        "medium": "question on Google's developer forum",
        "date": "2026-08-19"
      }
    ],
    "sources": [
      "https://docs.x.ai/developers/advanced-api-usage/watermarking",
      "https://openai.com/index/eu-text-provenance",
      "https://help.openai.com/en/articles/8912793",
      "https://www.anthropic.com/news/claude-text-watermark",
      "https://support.claude.com/en/articles/16266773",
      "https://deepmind.google/blog/watermarking-ai-generated-text-and-video-with-synthid/",
      "https://blog.google/innovation-and-ai/products/google-synthid-ai-content-detector/",
      "https://digital-strategy.ec.europa.eu/en/policies/code-practice-ai-generated-content",
      "https://digital-strategy.ec.europa.eu/en/news/strong-backing-code-practice-transparency-ai-generated-content",
      "https://digital-strategy.ec.europa.eu/en/faqs/code-practice-transparency-ai-generated-content",
      "https://discuss.ai.google.dev/t/does-gemini-api-text-output-carry-synthid-watermarking-gemini-2-5-flash-lite-gemini-3-1-flash-lite-eu-ai-act-art-50-2/177241/4",
      "https://discuss.ai.google.dev/t/does-gemini-api-text-output-carry-synthid-watermarking-gemini-2-5-flash-lite-gemini-3-1-flash-lite-eu-ai-act-art-50-2/177241/2",
      "https://deepmind.google/models/synthid/",
      "https://docs.x.ai/grok/faq",
      "https://www.anthropic.com/claude-fable-and-mythos-5-1",
      "https://platform.claude.com/docs/en/release-notes/overview",
      "https://discuss.ai.google.dev/t/does-gemini-api-text-output-carry-synthid-watermarking-gemini-2-5-flash-lite-gemini-3-1-flash-lite-eu-ai-act-art-50-2/177241/5"
    ],
    "body": [
      {
        "figure": "main"
      },
      {
        "p": "Google, Anthropic and OpenAI each watermark AI-written text, and xAI is about to start, each drawing the line in a different place. xAI's [API documentation](https://docs.x.ai/developers/advanced-api-usage/watermarking), published under the name SpaceXSI, says that from Nov 7, 2026, seven Grok models will mark their text by default on every API endpoint. OpenAI [said on Oct 5](https://openai.com/index/eu-text-provenance) that its API stays off by default, and that it would mark ChatGPT and Codex text in the EU only."
      },
      {
        "table": {
          "cols": [
            "Lab",
            "Watermark",
            "As the lab's pages state it"
          ],
          "rows": [
            [
              "Google",
              "Apps",
              "[On (Gemini app and web). No off switch mentioned](https://deepmind.google/blog/watermarking-ai-generated-text-and-video-with-synthid/)"
            ],
            [
              "Google",
              "API",
              "[On (Gemini API, AI Studio and Antigravity). Per a Google staff reply. No off switch mentioned](https://discuss.ai.google.dev/t/does-gemini-api-text-output-carry-synthid-watermarking-gemini-2-5-flash-lite-gemini-3-1-flash-lite-eu-ai-act-art-50-2/177241/4)"
            ],
            [
              "Google",
              "Detector",
              "[Waitlist (May 2025)](https://blog.google/innovation-and-ai/products/google-synthid-ai-content-detector/)"
            ],
            [
              "Google",
              "Announced",
              "[May 14, 2024](https://deepmind.google/blog/watermarking-ai-generated-text-and-video-with-synthid/)"
            ],
            [
              "Google",
              "Started",
              "[May 14, 2024 (apps)](https://deepmind.google/blog/watermarking-ai-generated-text-and-video-with-synthid/)"
            ],
            [
              "Google",
              "EU Code signed",
              "[Yes](https://digital-strategy.ec.europa.eu/en/news/strong-backing-code-practice-transparency-ai-generated-content)"
            ],
            [
              "Anthropic",
              "Apps",
              "[On (the Claude apps, worldwide). No off switch mentioned](https://support.claude.com/en/articles/16266773)"
            ],
            [
              "Anthropic",
              "API",
              "[On (supported models). No off switch mentioned](https://support.claude.com/en/articles/16266773)"
            ],
            [
              "Anthropic",
              "Detector",
              "[Eligible organizations](https://support.claude.com/en/articles/16266773)"
            ],
            [
              "Anthropic",
              "Announced",
              "[Aug 14, 2026](https://www.anthropic.com/news/claude-text-watermark)"
            ],
            [
              "Anthropic",
              "Started",
              "[Sep 1, 2026 (new models)](https://platform.claude.com/docs/en/release-notes/overview)"
            ],
            [
              "Anthropic",
              "EU Code signed",
              "[Yes](https://digital-strategy.ec.europa.eu/en/news/strong-backing-code-practice-transparency-ai-generated-content)"
            ],
            [
              "OpenAI",
              "Apps",
              "[On (ChatGPT, with Codex announced). EU only. No off switch mentioned](https://openai.com/index/eu-text-provenance)"
            ],
            [
              "OpenAI",
              "API",
              "[Off by default. Customers can turn it on (for select models)](https://openai.com/index/eu-text-provenance)"
            ],
            [
              "OpenAI",
              "Detector",
              "[Researchers and expert organizations](https://openai.com/index/eu-text-provenance)"
            ],
            [
              "OpenAI",
              "Announced",
              "[Oct 5, 2026](https://openai.com/index/eu-text-provenance)"
            ],
            [
              "OpenAI",
              "Started",
              "[Oct 5, 2026 (API)](https://openai.com/index/eu-text-provenance)"
            ],
            [
              "OpenAI",
              "EU Code signed",
              "[Yes](https://digital-strategy.ec.europa.eu/en/news/strong-backing-code-practice-transparency-ai-generated-content)"
            ],
            [
              "xAI",
              "Apps",
              "No text watermark mentioned"
            ],
            [
              "xAI",
              "API",
              "[On by default (seven Grok models). Team admins can turn it off (per model, some models locked on)](https://docs.x.ai/developers/advanced-api-usage/watermarking)"
            ],
            [
              "xAI",
              "Detector",
              "[By application](https://docs.x.ai/developers/advanced-api-usage/watermarking)"
            ],
            [
              "xAI",
              "Announced",
              "[Docs updated Oct 31, 2026](https://docs.x.ai/developers/advanced-api-usage/watermarking)"
            ],
            [
              "xAI",
              "Started",
              "[Nov 7, 2026 (planned)](https://docs.x.ai/developers/advanced-api-usage/watermarking)"
            ],
            [
              "xAI",
              "EU Code signed",
              "[Not on the list](https://digital-strategy.ec.europa.eu/en/news/strong-backing-code-practice-transparency-ai-generated-content)"
            ]
          ]
        }
      },
      {
        "h": "The EU deadline behind it"
      },
      {
        "p": "The EU AI Act's Article 50 requires providers of generative AI to mark its output in a machine-readable way. The obligations [apply from Aug 2, 2026](https://digital-strategy.ec.europa.eu/en/policies/code-practice-ai-generated-content), and AI systems placed on the market before that date have [until Dec 2, 2026](https://digital-strategy.ec.europa.eu/en/faqs/code-practice-transparency-ai-generated-content). About 190 organizations had signed the EU's Code of Practice on AI-generated content by the end of July, and the [EU's list of signatories](https://digital-strategy.ec.europa.eu/en/news/strong-backing-code-practice-transparency-ai-generated-content) includes Anthropic, Google and OpenAI. xAI is not on it. Anthropic, OpenAI and xAI each say they watermark text to comply with the Act."
      },
      {
        "h": "Anthropic marks everywhere, OpenAI in the EU"
      },
      {
        "p": "Anthropic [said on Aug 14](https://www.anthropic.com/news/claude-text-watermark) that it applies watermarking globally \"because we don't yet have a durable way to scope it by region.\" Its [help center](https://support.claude.com/en/articles/16266773) lists the surfaces: the API, the Claude apps, Claude Code, Claude Cowork, Claude Tag and the AWS, Google Cloud and Microsoft Foundry clouds. Models launched from Aug 2, 2026 are marked at launch: its post launching [Claude Fable 5.1](https://frontierlog.ai/model/claude-fable-5-1) and Mythos 5.1, which it [released on Sep 1](https://platform.claude.com/docs/en/release-notes/overview), [says](https://www.anthropic.com/claude-fable-and-mythos-5-1) the Code \"required us to add a watermark\" to \"the outputs of models released after August 2, 2026\". Anthropic says it is adding marks to older models. On Nov 1 its table showed text marks on [Claude Opus 4.8](https://frontierlog.ai/model/claude-opus-4-8), Fable 5, Sonnet 5 and Opus 5, which launched before Aug 2, and none yet on models such as [Claude Opus 4.7](https://frontierlog.ai/model/claude-opus-4-7) and Sonnet 4.6."
      },
      {
        "p": "OpenAI took the opposite approach. Its [Oct 5 post](https://openai.com/index/eu-text-provenance) says it is \"not making text watermarking a global default at launch\": it would add the mark to ChatGPT and Codex text in the EU \"over the coming weeks\", and API customers anywhere can turn it on for select models. Its [help center](https://help.openai.com/en/articles/8912793) lists text marking as \"textGrain watermarks (EU only)\" and says ChatGPT text in the EU now carries it."
      },
      {
        "h": "xAI marks its API, with an opt-out"
      },
      {
        "p": "xAI's page names grok-4.7, grok-4.6, grok-4.5, grok-4.3 and three grok-4.20 models, and covers plain text, structured outputs and tool-call arguments. Team admins can turn the mark off per model after confirming \"that an exception under applicable law applies to your use\", and some models may be locked on. Neither this page nor xAI's consumer terms and FAQs mention a text watermark in the Grok apps. Its [Grok FAQ](https://docs.x.ai/grok/faq) says generated images and videos carry a Grok watermark."
      },
      {
        "h": "Google's API, per its forum"
      },
      {
        "p": "Google's pages name the Gemini app and web for text watermarking, and its [SynthID Detector post](https://blog.google/innovation-and-ai/products/google-synthid-ai-content-detector/) says SynthID covers text from its Gemini models \"across Google\". None of its documentation names the Gemini API. On Google's developer forum, a Google staff member [wrote on Aug 19](https://discuss.ai.google.dev/t/does-gemini-api-text-output-carry-synthid-watermarking-gemini-2-5-flash-lite-gemini-3-1-flash-lite-eu-ai-act-art-50-2/177241/4) that \"text generated via the Gemini API IS actually SynthID-watermarked\", and that the same applies to Google AI Studio and Antigravity. The reply corrected [an Aug 5 reply](https://discuss.ai.google.dev/t/does-gemini-api-text-output-carry-synthid-watermarking-gemini-2-5-flash-lite-gemini-3-1-flash-lite-eu-ai-act-art-50-2/177241/2) that said API text was not watermarked. Later questions in the thread, on Vertex AI, on [when API marking began](https://discuss.ai.google.dev/t/does-gemini-api-text-output-carry-synthid-watermarking-gemini-2-5-flash-lite-gemini-3-1-flash-lite-eu-ai-act-art-50-2/177241/5) and on how others can check API text, have no reply from Google."
      },
      {
        "h": "Who can check"
      },
      {
        "p": "Each lab limits who can check text for its mark. xAI grants detection by application, and OpenAI to approved researchers and expert organizations. Anthropic opens its detection API to eligible organizations such as regulators, media and researchers, and to enterprises that must verify watermarks for their own compliance. Google opened its [SynthID Detector](https://blog.google/innovation-and-ai/products/google-synthid-ai-content-detector/), which takes text as well as images, audio and video, to early testers and a waitlist on May 20, 2025. Its [SynthID page](https://deepmind.google/models/synthid/) now says anyone can check an image, video or audio file with it, and names no wider access for text."
      }
    ],
    "events": [
      "2024-05-14-synthid-text-and-video-watermarking",
      "2025-05-20-synthid-detector",
      "2026-08-14-claude-text-watermark",
      "2026-09-01-claude-fable-5-1-mythos-5-1"
    ],
    "stories": [],
    "models": [
      "claude-fable-5-1",
      "claude-opus-4-8",
      "claude-opus-4-7"
    ]
  }
}