Updated
Updated · Business Insider · Aug 14
Claude Watermark Triggers Subscription Cancellations as Anthropic Says No Uptick Since Aug. 2
Updated
Updated · Business Insider · Aug 14

Claude Watermark Triggers Subscription Cancellations as Anthropic Says No Uptick Since Aug. 2

3 articles · Updated · Business Insider · Aug 14

Summary

  • Dozens of Claude users have posted since Monday that they canceled subscriptions over Anthropic’s new imperceptible text watermark, and Business Insider verified cancellations from four users including several $100-a-month Claude Max subscribers.
  • Anthropic applied the watermark globally on Aug. 2 to supported Claude text outputs, saying it survives copy-paste and some editing and signals only that Claude processed content, not necessarily authored it.
  • Users said that distinction does not solve their problem: developers, consultants and researchers fear watermarked code or edited client work could trigger authorship disputes, compliance issues or contract penalties.
  • Anthropic said it has not seen a trend of increased cancellations, even as backlash continues and some former users say they are switching to rivals such as Cursor, Grok and Chinese AI providers.
  • The company says the change is meant to comply with the EU AI Act and plans a free API for checking Claude watermarks, joining Google, OpenAI and X in labeling AI-generated content.

Insights

Can invisible AI watermarks truly survive your edits, or is this just a false sense of security for regulators?
Could Anthropic's invisible tracking tags accidentally flag your original, human-written work as AI-generated?