Anthropic adds hidden machine readable marks to Claude in EU
Technology
Anthropic is rolling out a new feature for its Claude AI models: hidden, machine-readable marks that show when text or files are AI-generated.
This move follows European Union rules and is all about making it easier to spot what's made by AI on different platforms.
Claude gets EU watermarks and metadata
Starting August 2, 2026, supported Claude models launched in the European Union will get invisible watermarks on generated text, and where supported, files generated by Claude will include digitally signed provenance metadata.
These tools will show up across supported Claude products, and Anthropic says more details on how to detect AI content are coming.