AI CRITIQUE

Temsilî görsel — FLUX.1-dev ile bilgisayarımda üretildi; haberin gerçek fotoğrafı değil.Illustrative image — generated with FLUX.1-dev on my computer; not a real photograph of the story.

Anthropic adds invisible watermarks to new Claude models in compliance with EU regulations

Read the article English (automatic translation)

Summary

Anthropic will embed invisible, machine-readable watermarks into texts generated by its new Claude models starting August 2, 2026, to meet the EU AI Act transparency requirements.

Why it matters

This move supports regulatory compliance by providing a technical method to identify AI-generated content and gives platforms the ability to verify provenance. By applying the watermark globally and publishing detection tools, Anthropic sets a benchmark for transparency standards across the industry.

Highlights

Invisible watermarks persist through copy-paste operations and can bypass moderate-level moderation.
Supported files will carry provenance metadata signed according to the C2PA standard.
Anthropic plans to publish technical documentation for detecting its watermarks via APIs and platforms.

Critical take

While the watermark provides a useful signal, its reliability is limited—heavy editing or short excerpts can erase it, and the presence of a mark does not guarantee that the content was originally created by Claude.

Bu kritik, yerel yapay zeka modeli gpt-oss-120b tarafından yazılmıştır. Haberi ben seçtim; kritik metnini yayımlanmadan önce tek tek denetlemiyorum.This critique was written by the local AI model gpt-oss-120b. I chose the story; I do not review each critique before it is published.

Back to the issue