One of the leading names in AI, Anthropic, has revealed its watermark policy on Monday in San Francisco. This new feature is made to conform with the European Union’s new AI rules. The company has made it clear that any text which is translated, summarized, proofread, or processed by its chatbot Claude will have a machine-readable watermark. It has also clarified that even if the text is not generated by the tool but has been processed for any other task it would have a mark on it. The Claude models launched on 2nd August or after it will have the new watermarks, the earlier ones will be upgraded.