{"slug": "watermarks-without-verification-ai-text-watermarking-after-the-eu-ai-act", "title": "Watermarks Without Verification: AI Text Watermarking After the EU AI Act", "summary": "A paper submitted to arXiv on 9 Sep 2026 argues that the AI text watermarks now required under Article 50 of the EU AI Act, which took effect on August 2, 2026, cannot currently be verified by users or regulators. Anthropic disclosed that every Claude model released after that date embeds a SynthID-Text watermark in all generated text by default with no user opt-out, while Google has deployed SynthID-Text in Gemini since 2024. Testing the open-source SynthID-Text implementation on two open-weight models, the authors found the watermark's effect on prose does not exceed that of changing the sampling seed, while on code it costs three points of correctness on one model and remains below measurement on the other, with detection near chance.", "body_md": "# Computer Science > Computers and Society\n\n  [Submitted on 9 Sep 2026]\n\n# Title:Watermarks Without Verification: AI Text Watermarking After the EU AI Act\n\n[View PDF](/pdf/2609.09604)\n\n[HTML (experimental)](https://arxiv.org/html/2609.09604v1)\n\nAbstract:On August 2, 2026, the obligations of Article 50 of the EU AI Act took effect, requiring generative AI providers to mark the content their systems produce and ensure it can be detected as AI-generated. Days later, Anthropic disclosed that every Claude model released after that date embeds a watermark based on SynthID-Text in all generated text, enabled by default with no user opt-out; Google has deployed SynthID-Text in Gemini since 2024. Users objected that the watermark degrades quality, particularly for code, that it secretly encodes identifying information, and, in mutual contradiction, that it is easily removable and inescapable; the vendor answered with assurances of unchanged quality, no identifying information, and robustness to light editing. In this work, we argue that neither the objections nor the assurances can currently be verified and that this unverifiability, rather than watermarking itself, is the substantive governance failure. We sort the contested assertions by what it would take to settle each and evaluate the open-source SynthID-Text implementation on two open-weight models, because no public tool can test the deployed systems. On prose, the measured effect of the watermark does not exceed that of changing the sampling seed. On code, the cost is three points of correctness on one model and below measurement on the other, while detection remains near chance, a limitation of detectability rather than quality. The remaining gaps trace to withheld access or missing institutions and we map each to a requirement: release of matched outputs, configuration disclosure, accredited audits, a shared evaluation protocol, and interoperable detection.\n    \n\n### Current browse context:\n\ncs.CY\n\n### References & Citations\n\nLoading...\n\n# Bibliographic and Citation Tools\n\nBibliographic Explorer \n\n*(*[What is the Explorer?](https://info.arxiv.org/labs/showcase.html#arxiv-bibliographic-explorer))\nConnected Papers \n\n*(*[What is Connected Papers?](https://www.connectedpapers.com/about))\nLitmaps \n\n*(*[What is Litmaps?](https://www.litmaps.co/))\nscite Smart Citations \n\n*(*[What are Smart Citations?](https://www.scite.ai/))\n# Code, Data and Media Associated with this Article\n\nalphaXiv \n\n*(*[What is alphaXiv?](https://alphaxiv.org/))\nCatalyzeX Code Finder for Papers \n\n*(*[What is CatalyzeX?](https://www.catalyzex.com))\nDagsHub \n\n*(*[What is DagsHub?](https://dagshub.com/))\nGotit.pub \n\n*(*[What is GotitPub?](http://gotit.pub/faq))\nHugging Face \n\n*(*[What is Huggingface?](https://huggingface.co/huggingface))\nScienceCast \n\n*(*[What is ScienceCast?](https://sciencecast.org/welcome))\n# Demos\n\n# Recommenders and Search Tools\n\nInfluence Flower \n\n*(*[What are Influence Flowers?](https://influencemap.cmlab.dev/))\nCORE Recommender \n\n*(*[What is CORE?](https://core.ac.uk/services/recommender))\n# arXivLabs: experimental projects with community collaborators\n\narXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.\n\nBoth individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.\n\nHave an idea for a project that will add value for arXiv's community? [**Learn more about arXivLabs**](https://info.arxiv.org/labs/index.html).", "url": "https://wpnews.pro/news/watermarks-without-verification-ai-text-watermarking-after-the-eu-ai-act", "canonical_source": "https://arxiv.org/abs/2609.09604", "published_at": "2026-09-11 02:07:09+00:00", "updated_at": "2026-09-11 02:22:37.004794+00:00", "lang": "en", "topics": ["ai-policy", "ai-safety", "generative-ai", "large-language-models", "ai-ethics"], "entities": ["Anthropic", "Claude", "SynthID-Text", "Google", "Gemini", "EU AI Act", "Article 50", "arXiv"], "alternates": {"html": "https://wpnews.pro/news/watermarks-without-verification-ai-text-watermarking-after-the-eu-ai-act", "markdown": "https://wpnews.pro/news/watermarks-without-verification-ai-text-watermarking-after-the-eu-ai-act.md", "text": "https://wpnews.pro/news/watermarks-without-verification-ai-text-watermarking-after-the-eu-ai-act.txt", "jsonld": "https://wpnews.pro/news/watermarks-without-verification-ai-text-watermarking-after-the-eu-ai-act.jsonld"}}