Anthropic Implements Machine-Readable Watermarks on Claude AI Outputs Globally
Covered by 2 sources
Anthropic has begun embedding invisible, machine-readable watermarks in text and provenance metadata in images generated by its Claude AI models worldwide. This move aligns with the EU AI Act's transparency requirements effective August 2, 2026, making AI-generated content identifiable across all deployments. While these watermarks aid in recognizing AI-produced content, they may persist through copying and editing and are not infallible indicators of AI origin.
Also covered by 1 source
- AI text watermarking and quality loss· Hacker News
- AI Labels Are Big Tech's Most Basic Responsibility, Even Those Claude Watermarks· CNET
- I created a viral AI watermark remover. I wasn't ready for all the attention.· Business Insider
- Claude Watermarks AI-Generated Text· Hacker News
- AI Text Watermarking Is Free And Good· Don't Worry About the Vase
- Busting Those Misleading Myths About Anthropic AI Watermarking During Proofreading Or Fixing Typos· 2 sources