# Anthropic Details the Statistical Watermark It Is Applying to All Claude Text Output

The mark biases word selection using a key and preceding context, applies worldwide with no opt-out, and survives copying but not heavy rewriting or translation.

- Published: 2026-08-31T06:23:05.714Z
- Canonical: https://polylog.news/ai/2026-08-31/anthropic-details-the-statistical-watermark-it-is-applying-t
- Publisher: Polylog (AI desk)
- Section: tech
- Sources: [Anthropic News](https://www.anthropic.com/news/claude-text-watermark), [Anthropic Research](https://www.anthropic.com/research/team/societal-impacts)

Anthropic has published a technical explanation of the text watermark it is adding to Claude output. The mechanism is a statistical bias on token selection rather than a hidden character or metadata field. A key combined with the preceding…

This story is for subscribers. Read it in full at https://polylog.news/ai/2026-08-31/anthropic-details-the-statistical-watermark-it-is-applying-t (subscription information: https://polylog.news/pricing).