{"id":23262,"date":"2026-08-11T20:39:37","date_gmt":"2026-08-11T20:39:37","guid":{"rendered":"https:\/\/scannn.com\/anthropic-to-start-watermarking-claude-generated-text-images\/"},"modified":"2026-08-11T20:39:37","modified_gmt":"2026-08-11T20:39:37","slug":"anthropic-to-start-watermarking-claude-generated-text-images","status":"publish","type":"post","link":"https:\/\/scannn.com\/lv\/anthropic-to-start-watermarking-claude-generated-text-images\/","title":{"rendered":"Anthropic to start watermarking Claude-generated text, images"},"content":{"rendered":"\n
\n

Anthropic PBC has announced plans to embed an invisible watermark in text and images generated by Claude.<\/span><\/p>\n

The Register <\/span>reported<\/span><\/a> the change today, citing a <\/span>help desk article<\/span><\/a> published on Monday. It applies to the Claude artificial intelligence model series and the Anthropic services that it powers.\u00a0<\/span><\/p>\n

The watermarking mechanism is designed to bring Claude into compliance with the European Union\u2019s AI Act. The law, which went into effect in 2024, includes a voluntary clause known as the Code of Practice that Anthropic has signed. The provision requires model providers to mark AI-generated content.<\/span><\/p>\n

Anthropic says text watermarking will be performed by Claude models released after Aug. 2. It plans to bring the capability to earlier models further down the line. According to the company, the watermark is invisible and persists if the text that contains it is copied to other applications. It may even remain in place if users make edits.<\/span><\/p>\n

The company didn\u2019t specify how the text watermark works. However, a <\/span>research paper<\/span><\/a>\u00a0released by Google DeepMind in 2024 may provide clues. The paper describes a technology that modifies a language model\u2019s word choices to make the text it generates detectable. Anthropic\u2019s watermarking mechanism may use a similar method.<\/span><\/p>\n

DeepMind\u2019s watermarking mechanism supports not only text but also AI-generated images, videos and audio. The technology, which is known as SynthID, is integrated into several of Google LLC\u2019s consumer AI services. Anthropic likewise plans to mark media files generated by Claude. The company\u2019s Monday help desk article lists JPG, PNG and SVG among the image formats covered by the initiative.<\/span><\/p>\n

Anthropic will mark images using a technology called C2PA. It works by pairing AI-generated images with a metadata file that specifies what model created them, when and whether any copyright restrictions apply. Developers can also add in other details.\u00a0<\/span><\/p>\n

C2PA includes multiple mechanisms designed to prevent hackers from tampering with images\u2019 metadata. It creates a hash, or unique identifier, of each metadata file that makes it easy to spot edits. The technology can also detect when hackers attempt to replace the entire metadata file rather than edit it.<\/span><\/p>\n

Anthropic plans to release tooling that will make it easier for users to detect Claude-generated content. However, the company cautioned that the technology won\u2019t be perfect. In particular, it may not always detect watermarks embedded in short or heavily edited text snippets.<\/span><\/p>\n

OpenAI Group PBC has also signed the AI Act\u2019s voluntary Code of Practice. The provision went into effect last week, which means that the ChatGPT developer may soon launch its own text watermarking mechanism. OpenAI <\/span>created<\/span><\/a> an implementation of such a feature in 2024 and sits on the steering committee of the consortium that develops C2PA.<\/span><\/p>\n

Image: Anthropic<\/a><\/h5>\n
\n
\n

Support our mission to keep content open and free by engaging with theCUBE community. Join theCUBE\u2019s Alumni Trust Network<\/strong>, where technology leaders connect, share intelligence and create opportunities.<\/p>\n