Anthropic to watermark text generated by its AI models

Anthropic is reportedly set to add a special mark or watermark to text generated by its artificial intelligence (AI) models, including Claude, in order to comply with regulations in force in Europe.

The AI model maker confirmed the marking on a support page it has updated.

In a TechCrunch report on Tuesday (11/8) local time, the European regulation in question refers to the EU AI Act on code transparency, which comes into effect on 2 August 2026.

AI companies are required to mark content generated by AI or edited by AI in a way that can be identified by other systems.

Anthropic stated that all models released after 2 August will automatically have technology to place a special mark on computer-generated text and also on files. For files, the company uses the open C2PA standard.

“Because the watermark is part of the text, it will travel with the text when copied and pasted elsewhere, and may remain even after some editing,” read a statement from Claude.

The company noted that the addition of watermarks will be applied across various products such as the Claude platform API, Claude, Claude Code, Claude Cowork, and Claude Tag.

Various platforms are now racing to place special marks on AI-generated content after facing criticism from users and to avoid regulatory scrutiny.

For example, the AI music platform Suno said it would mark songs created on its platform following a series of legal challenges.

Then in July, the newsletter service Substack partnered with Pangram to label AI-generated content. The company’s CEO, Chris Best, criticised Claudefishing, a term used for people who use AI to generate content.

Besides Anthropic, other companies such as Black Forest Labs, Google, Meta, Microsoft, OpenAI, and Synthesia have committed to complying with the EU code of practice.