Urgent.News

600+ sources. One page. See who else covered it.

Editions

More in AI

AI Text Watermarking: The Statistics Hiding Inside Every Sentence

Something changed on August 2, 2026. Every new Claude model started embedding an invisible mark into the text it generates. You can't see it, strip it with backspace, or find it with Ctrl+F.

  • Claude models embed invisible statistical bias watermark since August 2026
  • Google's SynthID uses pseudorandom token sampling for watermarking
  • Watermark detection requires secret key and one-proportion z-test

More from Saturday 15 August →