Anthropic Watermarks Claude AI Text Outputs

Anthropic is adding watermarking to Claude AI-generated text to distinguish AI from human content, addressing academic integrity, AI slop, and legal requirements like the EU AI Act. The method embeds an imperceptible mark during generation, but editing or reprocessing can dilute it.