Copy-paste no more: Anthropic puts invisible watermarks on Claude text under EU rules

Anthropic has announced the implementation of an invisible watermarking feature for text generated by its Claude models. This move is in response to EU regulatory requirements, which mandate that AI-generated content be identifiable to prevent misuse and ensure transparency. The watermark is embedded in a way that is imperceptible during normal reading, but can be detected using specific technological tools, allowing the origin of the text to be verified. This initiative is part of broader industry efforts to increase accountability in AI outputs.