Anthropic to Embed Invisible Watermarks in All Claude-Generated Text

nptacek · x · 2026-08-11

Anthropic announces that starting August 2, 2026, new Claude models will embed invisible watermarks in all generated text. The watermark is part of the text, not metadata, and persists through copying and some editing. This is part of the EU AI Act code Anthropic signed, with rollout worldwide and ongoing work for current models.

Related event: Anthropic to Embed Invisible Watermarks and C2PA Signatures in Claude Outputs from Aug 2026(14 posts)→

Original post →

More from Models

Models channel →