Anthropic Details How Claude Marks AI-Generated Content
TL;DR. Anthropic signed the EU AI Act's transparency code, detailing how Claude models will mark AI-generated content with embedded watermarks and provenance metadata. - New Claude models launching in the EU after August 2, 2026, will support machine-readable marking from release. - Existing models are being updated to include these transparency features, with details to follow in technical guidance. - Marking will apply globally across all Claude platforms and cloud partners like AWS, Google Cloud, and Microsoft Foundry.
- New Claude models in the EU from August 2, 2026, will feature machine-readable marking at launch.
- Generated text will include embedded watermarks; files will have digitally signed provenance metadata where supported.
- Anthropic is working to add marking support to existing Claude models launched before the EU AI Act's deadline.
Sources
- How Claude marks AI-generated content — support.claude.com
- theregister.com — theregister.com
- the-decoder.com — the-decoder.com
- mashable.com — mashable.com