Anthropic shares more details about how Claude’s new watermarks will work

Live updates

1. 29d ago

   Anthropic's Claude AI introduces invisible watermarks

   Anthropic's Claude AI now includes invisible watermarks on generated text. The watermarks are designed to help identify AI-generated content and comply with EU law. Users cannot opt out of the watermarks. The move has sparked controversy, with some users canceling their subscriptions.

   Why it matters

   The introduction of watermarks on AI-generated content is a significant development in the AI industry. It aims to address concerns around the use of AI-generated content, such as misinformation and copyright issues. The EU has been actively working on regulations to govern the use of AI.

   What is confirmed

   • Claude users cannot opt out of new watermarks.
   • The watermarks are designed to be invisible.
   • The move aims to help identify AI-generated content and comply with EU law.

   Still unconfirmed

   • Some users are canceling their subscriptions due to the new watermarks.

   What to watch next

   • Regulatory approval from the EU.
   • User reaction to the new watermarks.
   • Development of methods to detect and defeat the watermarks.
   Sources used for this update (8)

   1. Forbes — Claude Users Can’t Opt Out Of New Watermarks—Here’s What We Know
   2. Ars Technica — Claude’s new Scarlet Letter watermark is invisible—for now
   3. Business Insider — Claude users are canceling their subscriptions, citing Anthropic’s new AI watermark
   4. Anthropic — How Claude's text watermarking works
   5. TechCrunch — Anthropic shares more details about how Claude’s new watermarks will work
   6. Engadget — Anthropic Is Watermarking Text Generated By Claude To Comply With EU Law
   7. Search Engine Journal — Anthropic Reveals What The Watermark Is And How It Can Be Defeated
   8. Futurism — People Horrified That They’ll Be Busted Now That Anthropic Is Watermarking AI Content
   confidence 90%