Jukan: Adding watermarks to Anthopic on the eve of IPO is serious self harm

According to Citrini analyst Jukan, Anthropic's decision to add an invisible watermark to Claude's model before its IPO was a serious self harming decision. Anthropic has started adding machine recognizable invisible watermarks to text generated by Claude, Claude Code, Cowork, API, as well as supporting models on AWS, Google Cloud, and Microsoft Foundry. The new model will be supported as soon as it goes live, and existing models will be gradually added. Anthropic embeds watermarks directly into text content to comply with the transparency requirements of Article 50 of the EU AI Act. Watermarks can be copied, pasted, and disseminated globally without affecting readability.