Anthropic to Add Invisible Watermarks to Claude Output

Anthropic notified customers that starting November 3, Claude's text output will carry invisible statistical watermarks that survive copy-paste and some edits, with detection currently limited to regulators, journalists, and fact-checkers. Some users reacted by urging a switch to open-source models.

2026-10-31 ~ 2026-10-31 · 4 related posts

• Anthropic adds invisible text watermark to Claude outputs starting Nov 3 — petesena · 2026-10-31
• Claude text gets invisible watermarking Nov 3: token-level statistical skew, not hidden chars — petesena · 2026-10-31

2 near-duplicate retellings: petesena · alex_verem