Anthropic to watermark text generated by its AI models
Anthropic says its AI models, including Claude, will now watermark generated text and files to comply with EU rules. The move follows the EU AI Act's Transparency Code, which took effect on August 2.
Anthropic has confirmed that its AI models will automatically watermark generated text and files, according to an update posted on the company's support page. The change is tied to new European Union transparency requirements for AI-generated content.
The EU AI Act's Transparency Code, which took effect on August 2, requires AI companies to mark AI-generated or edited content so it can be identified by other systems. Anthropic said all models released after that date will include watermarking technology by default. For files, the company relies on the open C2PA standard.
Anthropic also plans to extend watermarking support to its older models over time. According to the company, the watermark is embedded within the text itself, meaning it travels along when the text is copied and pasted elsewhere, and may survive some amount of editing. Because the watermarking happens at the model level, it will be present regardless of which Claude product or interface generated the text — including the Claude platform API, Claude, Claude Code, Claude Cowork, and Claude Tag.
It remains unclear how much a user would need to edit AI-generated text before the watermark disappears. TechCrunch has asked Anthropic for clarification and is awaiting a response.
Wider industry shift
Anthropic joins a growing list of companies moving to label AI-generated content, driven both by user backlash and a desire to stay ahead of regulatory scrutiny. AI music platform Suno recently announced it would mark tracks made on its service, while newsletter platform Substack partnered with Pangram last month to flag AI-generated content. Alongside Anthropic, companies including Black Forest Labs, Google, Meta, Microsoft, OpenAI, and Synthesia have also committed to following the EU's code.
