tech π€ Anthropic is tagging Claude text with invisible watermarks
Anthropic will implement machine-readable watermarks on text from new Claude models starting August 2, 2026. This change is in response to transparency rules under Article 50 of the EU AI Act, and Anthropic has already signed the EUβs Code of Practice. The invisible text watermarks will apply globally across all Claude products, including APIs and major cloud platforms. Supported files will use signed provenance data based on the C2PA standard for tracking. Anthropic notes that these signals are aids, not absolute proof of authorship.