Anthropic Labs faced a wave of criticism and user cancellations after it introduced a “visible watermark” on all text created or edited using its artificial intelligence assistant Claude.
The new feature aims to differentiate between generated and human content to comply with EU artificial intelligence laws. However, professional developers, editors, and consultants have expressed serious concerns that tags will be retained even with minimal editing or translation of their own text. This could lead to issues with compliance with company contracts, academic standards and copyright, prompting users to migrate to other platforms.
Company representatives said the watermark only indicates the involvement of artificial intelligence in word processing, not authorship, and that an overall sharp increase in subscription cancellations has not yet been recorded.