Anthropic Signs EU AI Act, Adds Invisible Watermarks to Claude Outputs

MicahBerkley · x · 2026-08-11

By signing the EU AI Act, Anthropic is implementing invisible watermarks for the text outputs generated by Claude. This mechanism is designed to trace the origin of the content and is reported to persist through some degree of editing.

While this move supports AI safety and copyright protection, it has sparked a debate within the community. Critics argue it introduces unnecessary friction for builders and creators who rely on AI for legitimate workflows, questioning whether this represents genuine safety or just excessive control.

Related event: Anthropic to Add Dual Watermarks to Claude Outputs for EU AI Act Compliance(20 posts)→

Original post →

More from Models

Models channel →