- Anthropic announced it would embed invisible watermarks in some versions of Claude's text output, applied globally since August 2.
- Since Monday, users have been posting on X that they have canceled their Claude subscriptions, citing the watermark.
- Business Insider spoke with four users who confirmed their cancellations, including Arturo Villarroya, Vladislav Rajtmajer, Richard Echols, and Santiago Alessandro Rivera Martínez.
- Anthropic stated it hasn't seen a trend of an uptick in cancellations and clarified that the watermark indicates Claude processed text, not necessarily authorship.
- Some users expressed concerns about the watermark persisting in client work, academic settings, and code, while others supported watermarks for transparency.
- Arturo Villarroya, who works in public policy, canceled his Claude Max subscription after becoming "uncomfortable" with the watermarking plan.
- Villarroya expressed concern that the watermark could remain even when the underlying writing was original, potentially causing issues in academic and professional settings.
- Vladislav Rajtmajer, a freelance developer, also canceled his subscription, citing concerns about the watermark affecting authorship and potential contract penalties.
- Richard Echols, an AI consultant, canceled his subscription due to worries that the watermark could appear on his original work.
- Santiago Alessandro Rivera Martínez expressed concerns about becoming too dependent on Claude, leading him to cancel his subscription.
- Anthropic added the watermark to comply with the European Union's AI Act and plans to roll out a free API for users to check for the watermark.
- Google and Elon Musk's X also use watermarking technologies for AI-generated content.
Anthropic's recent introduction of an invisible watermark in its AI assistant Claude has led to a wave of subscription cancellations among users concerned about the watermark's potential to persist in their work.125617
The AI lab announced that some versions of Claude will embed a watermark directly into their text output, which may survive editing and travel with the content when copied and pasted.
Arturo Villarroya, a public policy professional, expressed his discomfort with the watermark, stating, "I canceled my Claude Max subscription — a service that starts at $100 a month — after becoming uncomfortable with Anthropic's new watermarking plan." He fears that the watermark could remain even when the writing is original, complicating matters in academic and professional environments with strict AI usage rules.491011
Despite the backlash, Anthropic claims it has not observed a significant increase in cancellations since the watermark announcement. The company reassured users that the watermark does not alter the meaning, quality, or readability of Claude's responses.
However, many users remain skeptical, with some sharing their cancellation confirmations and voicing concerns about the implications of a persistent marker in their work, particularly in coding and client projects.
As the debate continues, the future of Claude's watermarking feature remains uncertain, with users weighing the benefits of AI assistance against the risks of compromised originality.
“Anthropic says the watermark doesn't affect quality and is meant to comply with the EU's AI Act, with a free API planned for checking. However, users like Arturo Villarroya, who pays $100 a month for Claude Max, fear the mark could appear even when Claude only proofreads or translates original work.”
