Anthropic Plans Claude Watermarks for AI Generated Content

Anthropic will add machine readable marks to content generated by Claude models launched in the EU on or after August 2, 2026, detailed in a Claude support article.

The company says the marks are part of its commitments under the EU AI Act Article 50(2) Code of Practice on Transparency of AI Generated Content. New Claude models launched in the EU from that date will support marking at launch, and Anthropic says it is working to add marking support to older models during the transition period.

Claude generated text will carry embedded watermarks that are not visible to users. Anthropic says the watermark is part of the text, so it can remain when users copy and paste output and may persist through some editing. The marking will apply at the model level across supported Claude products, including Claude Platform, Claude, Claude Code, Claude Cowork, and Claude Tag.

For supported files, including .svg, .png, and .jpg files, Claude will attach signed provenance metadata using the C2PA open standard. Anthropic says this metadata can show that a file was processed by Claude and can help detect whether the file has been changed.

The company says marks will apply worldwide wherever supported Claude models are offered. It also says it will provide detection support for users and third parties, with more technical documentation to come. Anthropic notes that a detected mark is not conclusive proof of authorship, and the absence of a mark does not prove that content was not AI generated or processed by Claude.