Anthropic has announced it will watermark text produced by its models, including Claude, to comply with European regulations. The company updated its support documentation to describe the change.
The European Union’s AI Act transparency code, which came into force on August 2, 2024, requires that content created or edited by AI be marked in a way that other systems can detect.
How Anthropic will implement watermarking
According to Anthropic, every model released after August 2 will automatically include technology to watermark both machine-generated text and files. For files, the company is using the C2PA open standard.
Anthropic also said it will extend watermark support to older models. The support page states that because the watermark is part of the text, “it will travel with the text when it’s copied and pasted elsewhere, and may persist through some editing. Watermarking will be applied at the model level, which means it will be present no matter which Claude product or surface the text comes from.”
It is not yet clear how extensive edits must be to remove the watermark; we have asked Anthropic for clarification and will update the report if they respond.
Products affected
The company noted that watermarking will apply across multiple offerings, including the Claude platform API, Claude, Claude Code, Claude Cowork, and Claude Tag.
Industry context
Platforms are moving to mark AI-generated content in response to user backlash and to avoid regulatory scrutiny. For example, AI music platform Suno said it will flag tracks created on its platform after a series of legal challenges. Newsletter service Substack recently partnered with Pangram to identify AI-generated content; Substack CEO Chris Best highlighted the problem of “Claudefishing,” a term used for deceptive use of AI-generated material.
Beyond Anthropic, other organizations such as Black Forest Labs, Google, Meta, Microsoft, OpenAI, and Synthesia have committed to following the EU’s transparency code.
Why the change matters
The transparency code that took effect on August 2 aims to improve visibility into AI-generated and AI-edited materials so both users and automated systems can distinguish them from human-created content. Watermarking is one of the technical approaches companies are adopting to demonstrate compliance.
