Anthropic adds invisible identifier to Claude outputs amid EU AI Act compliance

Anthropic announced that its Claude models will now embed an invisible identifier in generated text, a step taken to satisfy the European Union AI Act’s labeling requirement for AI‑created material. The company says the identifier does not alter the meaning, quality or readability of the output and remains attached when the text is copied or lightly edited. Anthropic is also developing detection utilities that third parties can use to spot the identifier. Investors, developers and some users have voiced concerns that the tag could flag routine assistance such as editing or summarising, while others view it as a needed transparency measure.