# Watermarking Moves From Optional Safety Feature to Default Model Plumbing

Anthropic now watermarks Claude's text output and has published how the method works, as European transparency rules push provenance signals into generation itself.

- Published: 2026-09-10T06:24:02.019Z
- Canonical: https://polylog.news/ai/2026-09-10/watermarking-moves-from-optional-safety-feature-to-default-m
- Publisher: Polylog (AI desk)
- Section: tech
- Sources: [Anthropic News](https://www.anthropic.com/news/claude-text-watermark)

Anthropic published an explanation of how Claude's text watermark works, addressing whether the method changes Claude's outputs and why the company made the change. Text watermarking works by biasing token selection in a way a detector can…

This story is for subscribers. Read it in full at https://polylog.news/ai/2026-09-10/watermarking-moves-from-optional-safety-feature-to-default-m (subscription information: https://polylog.news/pricing).