News Summary

• Anthropic has signed the EU's Code of Practice on Transparency of AI-Generated Content.
• It plans to watermark Claude's output across its ecosystem.
• Text will carry an invisible statistical watermark, while supported visual formats will receive signed metadata noting that Claude was involved.
• Because the information is cryptographically signed, it can also help reveal whether the credential or file has subsequently been tampered with.
• Anthropic says the technique adds no characters, needs no extra tokens, doesn't change output quality, and carries nothing that identifies a user, an organization, or a chat.