Anthropic to Watermark Claude-Generated Text Under EU AI Rules

Anthropic will begin watermarking text generated by its AI models, including Claude, as the company moves to comply with new European Union transparency requirements for AI-generated content.

The company confirmed the change in an updated support page. The EU AI Act’s transparency requirements took effect on August 2 and require certain AI-generated or manipulated content to be marked in a machine-readable format.

Anthropic adds watermarks to Claude-generated text

Anthropic said models released after August 2 will automatically include technology for watermarking generated text and files. For files, the company is using C2PA, an open technical standard designed to provide information about the origin and history of digital content.

The company also plans to extend watermarking support to older models. Anthropic says the text watermark is designed to remain attached when generated material is copied and pasted and may survive some subsequent editing.

“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.”

Anthropic said watermarking will operate at the model level, meaning it will apply regardless of which Claude product or interface produces the text.

The change will cover products including Claude, the Claude platform API, Claude Code, Claude Cowork and Claude Tag. Anthropic has not specified how extensively text would need to be edited before the watermark could no longer be detected.

AI companies respond to transparency requirements

Anthropic is among several AI companies that have committed to following the EU’s transparency framework. Black Forest Labs, Google, Meta, Microsoft, OpenAI and Synthesia have also committed to the code.

Other technology platforms are introducing their own approaches to identifying AI-generated material. AI music company Suno recently said it would mark tracks created through its platform, while Substack partnered with Pangram to identify AI-generated content.

The growing use of watermarking reflects increasing regulatory and platform pressure to make synthetic content easier to identify as AI-generated material becomes more common across text, images, audio and other digital formats.