Watermarking and Claude

—- To comply with the EU AI Act's Article 50(2) Code of Practice on Transparency of AI-Generated Content, Anthropic has announced that "New models will mark AI-generated content from day one". ——

This feels like a much bigger shift than simply adding an “AI-generated” label. After a few years building systems that make AI output increasingly indistinguishable from human work. Seems we may be entering a phase where the systems themselves are expected to carry proof of where that work came from.

For those of us here building with AI: will this change how you design your systems today?

Where do you think the line should be — what should be marked as AI-generated, and what shouldn’t?

25 comments

Watermarking and Claude