EQUIPThe Verge · 2h ago
Claude will apply invisible watermarks to AI text and images
Executive Brief
The 30-second read
Anthropic is implementing invisible watermarks and metadata in Claude-generated content to meet European AI transparency standards. These machine-readable identifiers will apply to both text and image outputs without affecting human readability.
- 01Generated files will include digitally signed provenance metadata to track content origins
- 02Watermarking implementation ensures compliance with emerging European Union AI regulatory frameworks
- 03Embedded markers remain invisible to users while remaining accessible to automated verification tools
- 04Metadata integration utilizes C2PA standards for cross-platform digital content authentication
Go deeper · Equip playbook
Best Carry-On Luggage for Business Travelers →AI-generated summary · Verify at source
Anthropic has pledged to start marking Claude-generated text and images with machine-readable data, in an effort to comply with European rules for AI transparency. "Generated text will carry embedded watermarks, and generated files will include digitally signed provenance metadata where supported," Anthropic says on a new Claude support page. The changes are invisible to human […]