Anthropic to embed invisible watermarks in AI-generated text to meet EU rules

Anthropic has confirmed it will add watermarks to text produced by its AI models, including Claude, in response to European regulations. The company updated its support page to announce the change, which aligns with the EU AI Act's Transparency Code that took effect on August 2. That code requires AI companies to make AI-generated or edited content identifiable by other systems.

According to Anthropic, all models released after August 2 will automatically include technology that watermarks both computer-generated text and files. For files, the company is using the C2PA open standard. The firm also said it plans to extend watermarking to older models, and that the mark will stay with the text even when copied and pasted elsewhere, potentially surviving some edits. The watermark is applied at the model level, so it appears regardless of which Claude product or surface generates the text, including the Claude platform API, Claude Code, Claude Cowork, and Claude Tag.

It remains unclear how much editing might remove the watermark. Anthropic has been asked for clarification but has not yet responded. The move comes as platforms rush to mark AI-generated content after user backlash and growing regulatory pressure. Last week, AI music platform Suno said it would label tracks created on its service following legal challenges. In July, newsletter service Substack partnered with Pangram to flag AI content, with its CEO Chris Best criticizing what he called "Claudefishing" - using AI to generate posts. Other companies, including Black Forest Labs, Google, Meta, Microsoft, OpenAI, and Synthesia, have also committed to the EU's code.

Source: TechCrunch