Anthropic introduces watermarks into Claude's texts to meet standards

American Anthropic will introduce a watermark to the texts produced by its ‘chatbot’ Claude, in order to meet European requirements for Artificial Intelligence (IA), it has been announced.

11/08/2026
by

Lusa

Conforme needed, in a statement, this brand will resist the copy-paste, but it only indicates that the content is processed by Claude.

In this way, a text written by a human and subsequently revised, translated or summarized by the AI will still have the watermark.

The company has advanced with this measure to comply with a European AI regulation, which requires suppliers to identify its content with a legible watermark.

The offenders are subject to fines of up to EUR 15 million or 3% of their overall revenue.

French press complains against Google for AI summaries

The General Information Press Alliance, the main French newspaper association, today filed a complaint with the Competition Authority of France against Google due to the implementation of the tools for preparing abstracts with Artificial Intelligence (IA).

Lusa