Anthropic's text watermarks signal new front in AI detection

Anthropic is introducing machine-readable watermarks to text and files generated by Claude to comply with European Union transparency regulations and the Artificial Intelligence Act. These imperceptible patterns in text and digital signatures in media files will be applied worldwide for models launched in the EU after August 2. This move allows for AI detection, meaning documents formatted or translated by Claude may carry an AI signature, identifying the content as a product of artificial intelligence.