Anthropic Plans to Watermark Claude-Generated Text
Anthropic announced that future Claude models will watermark generated text to comply with the EU AI Act, effective August 2, and the method will not affect output quality, add hidden characters, or i…