Anthropic Adds Invisible Watermarks to AI Text

Anthropic has announced that it is adding invisible watermarks to text generated by its AI models. The hidden signal is designed to remain with the content when it is copied and pasted elsewhere and may also survive certain types of editing. Claude models launched from August 2, 2026, will support the watermarking system from the start, with the feature rolling out globally.

The company said the watermark will be imperceptible and will not affect the meaning, quality or readability of AI generated responses. Anthropic plans to eventually provide users and third parties with tools to detect the embedded marks. However, the company stressed that the watermark should not be treated as a definitive AI detector, as heavy editing, paraphrasing, translation or combining AI output with other text could make the mark undetectable.

The watermarking system will cover supported Claude models across Claude, Claude Platform API, Claude Code, Claude Cowork and Claude Tag, as well as supported access through AWS, Google Cloud and Microsoft Foundry. For supported file formats such as SVG, PNG and JPG, Anthropic will use digitally signed provenance metadata based on the C2PA standard to indicate whether a file has been processed by Claude and whether its provenance information has been altered.