Analyst Reviews Claude's Content Watermarking: A's Most Disastrous Self-Harm Decision

August 11th, Citrini analyst Jukan commented on the "New Claude Model Embeds Invisible Text Watermark" that "Anthropic's announcement of this move on the eve of its IPO is the most serious self-inflicted decision."

Earlier, Anthropic started adding machine-readable "invisible watermarks" to the text generated by Claude. Claude, Claude Code, Cowork, API, as well as models supported on AWS, Google Cloud, and Microsoft Foundry will all be covered. The new model is already online-ready, and existing models are being gradually added.

It is reported that Anthropic took this action to comply with Article 50 of the EU AI Act's transparency requirements, with the watermark directly embedded in the text content, capable of being propagated through copy and paste and globally applicable, without affecting readability.