Claude AI Content Watermarks

Claude AI Content Watermarks is Anthropic's system for marking content generated by Claude so it can later be identified as AI-generated. It combines two techniques: an imperceptible watermark woven into generated text, and signed C2PA Content Credentials attached to supported files such as PNG and JPEG. Anthropic says the text watermark does not change meaning, quality or readability, travels when text is copied and pasted, and may persist through some editing.

The programme is tied to Anthropic's signing of the EU AI Act Article 50(2) Code of Practice on Transparency of AI-Generated Content. Claude models launched in the EU on or after 2 August 2026 support marking at launch, and Anthropic is adding marking to earlier models. Marking covers the Claude Platform (API), Claude, Claude Code, Claude Cowork and Claude Tag, and applies worldwide. Text watermarks also carry through AWS, Google Cloud and Microsoft Foundry. Support varies by model, and Sonnet 4.6 may not reliably carry the text watermark when temperature, top_p or top_k is non-default.

Detection is in private preview. It is offered to eligible organisations such as regulators, law enforcement, media, fact-checkers, independent researchers, educational organisations and EU civil society groups. It is also offered to enterprises that must check marks on their own Claude outputs. Access is by request form, and Anthropic plans to widen it over time. Files can be checked for Content Credentials with the free Claude Content Checker. Detection is not conclusive: a found mark indicates content may have been generated by Claude. Commentators expect workarounds and stripping tools. Operators deploying Claude should assess their own Article 50 obligations.