Anthropic adds invisible watermarks to Claude text to meet EU transparency rules

Anthropic has begun adding invisible watermarks to text generated by newer Claude models and signed provenance metadata to some files, a move the company says is meant to comply with European Union transparency rules. The marks are designed to be hard to notice, to travel with copied text, and to help identify content processed by Claude across its products, including the API, Claude, Claude Code, Claude Cowork, and Claude Tag. Anthropic says the rollout applies worldwide to models launched on or after August 2, 2026, and that it is working on extending the feature to older models. The change matters because it could make AI-generated content easier to detect, but it has also sparked concern from users worried about being identified as using AI at work or in school.