Anthropic has announced plans to implement watermarking technology across its AI text generation models, including Claude, in response to European Union regulatory requirements. The move comes following the EU AI Act’s Transparency Code, which became effective on August 2 and mandates that AI companies clearly identify artificially generated or modified content in ways that other systems can detect.
The company stated that all models released after the regulation’s effective date will automatically incorporate watermarking capabilities for both text and files, with the latter utilizing the C2PA open standard. Anthropic confirmed it would also extend watermarking support to its older models. The watermarks are designed to remain embedded within the text, persisting when users copy and paste content across different platforms and potentially surviving minor edits.
The watermarking functionality will apply uniformly across Anthropic’s various products and services, including the Claude platform API and Claude Code. The initiative reflects a broader industry trend, as other major technology firms including Google, Meta, Microsoft, and OpenAI have similarly committed to adhering to the EU’s transparency requirements. Recent weeks have seen multiple platforms, from music streaming services to newsletter platforms, implementing their own content identification systems following increased user concerns and legal challenges.
Originally reported by Techcrunch →
About the Author
