Anthropic to Add Invisible Watermarks to All AI-Processed Content

Anthropic announced it will watermark content processed by its AI models, not just generated content, to comply with the European Union's AI Act. The watermarks will be invisible in text, while other files will carry digitally signed provenance metadata.

All new models will be affected globally from the start, even though the EU does not require watermarking for minor edits or assistive functions. Anthropic is applying the watermarking broadly to avoid exemptions and ensure compliance.