Claude users cancel subscriptions over watermark in generated text

Since Anthropic announced a watermarking system for text generated by Claude, some users have begun canceling their Claude Max subscriptions ($100 monthly), concerned that their work may be permanently marked as AI-processed. Interviews with canceling users revealed specific concerns: that even original texts only edited or verified by Claude could retain the watermark; that clients or institutions could question authorship of professional projects; and that the marker could violate contractual clauses or compliance policies. Some users mentioned exploring alternatives such as Chinese AI models, fearing excessive dependence on a proprietary vendor. Despite concerns raised, Anthropic stated it has not observed a significant overall increase in cancellations. The company clarified that the watermark only indicates that Claude processed text, not that it created it, and that it persists even through editing, verification, or summarization. While some view the watermark as contributing to transparency, others see it as restrictive for professional and academic workflows.