RESOLVING THE LEDGER
This is one outlet's own report from Mashable — the article as it was filed. Other outlets are covering the same event; open the full story to compare every source side by side.
See the full story · 3 sources
Anthropic's Claude will start watermarking AI-generated content in compliance with EU rules, the company announced on Tuesday.
In a Claude support document, Anthropic wrote that it has signed the EU AI Act's Article 50(2) Code of Practice on Transparency of AI-Generated Content , and explains how it plans to put those rules into practice.
SEE ALSO:
Claude Code’s auto mode will be on by default, Anthropic confirms
Models launched in the EU on or after Aug. 2, 2026, wrote Anthropic, will support watermarking at launch. This includes output from all models including Claude Platform (API), Claude, Claude Code, Claude Cowork, and Claude Tag. Generated text will have embedded watermarks, while generated files will include "digitally signed provenance metadata where supported."
Notably, it won't apply only to Europe; instead, the watermarks will appear wherever Claude is offered, worldwide, though Anthropic warns that some platforms or features may not support certain watermarking types.
Featured Video For You
Is ChatGPT Changing the Way We Write?
The watermarks won't be visible if you don't know what you're looking for. For text files, the watermark will be "part of the text" and it will "travel with the text when it's copied and pasted elsewhere," but Anthropic doesn't really go into the technicalities of how this will work. For files such as images, Anthropic will use the Coalition for Content Provenance and Authenticity (C2PA) open standard . Details on how to recognize these watermarks will be shared in "forthcoming documentation," wrote Anthropic.
As for existing models that were launched before Aug. 2, 2026 (that's pretty much all of them at writing time), Anthropic says it's working to add watermarking support to those as well.
Watermarking won't be foolproof, warns Anthropic. For example, heavily edited, paraphrased, or translated text could not carry a detectable watermark, and a file's metadata could be stripped through format conversion, screenshots, or other means.
Anthropic's move follows similar commitments by AI companies including OpenAI, Google, and Meta.
AIPROPX is an independent multi-source news index — we track, compare, and connect coverage from across the web into one place you won't find anywhere else.