Anthropic is watermarking Claude text, and panicking users are taking it personally

If you copy and paste text straight from Claude into your schoolwork or job reports, your secret might soon be out.

Anthropic started inserting invisible code into Claude's editorial text outputs to mark them as AI-generated, according to TechCrunch. The company rolled out the tags to comply with the EU AI Act's Transparency Code, which requires tech companies to label AI-generated content. On Reddit, an upset minority of users is calling the hidden tags a "digital tattoo" that will catch people using the chatbot for daily assignments.

Why it matters: Passing off raw AI text as original work just got much harder. While a few critics argue that watermarking text is hypocritical given how AI models scrape training data, most commenters are pointing out the obvious: hidden tags only bother you if you are pretending the chatbot's writing is your own.

Know this: The watermarks target direct outputs. Users can still obscure the hidden tags by paraphrasing or rewriting the text before turning it in.

Bottom line: if you want to use Claude at work, you might actually have to edit the output yourself.