Anthropic Adds Invisible Text Watermarks and C2PA Metadata to All Claude Outputs
dotey · x · 2026-08-11
Anthropic announced that starting August 2, all newly released Claude models globally will embed machine-readable markers in their outputs to comply with the EU AI Act's transparency guidelines.
The marking mechanism operates on two levels:
- Text Watermarks: Weaves invisible watermarks into generated text that persist through copy-pasting and moderate editing.
- File Metadata: Appends C2PA-compliant signed metadata to generated files like SVG and PNG.
This applies across APIs, the official website, and major cloud platforms. Anthropic is developing accompanying detection tools. This marks the first large-scale deployment of text-level watermarks by a major AI lab, contrasting with OpenAI, which has yet to deploy its developed watermarking tech.
More from Models
- First Model to Pass the Video Turing Test: Half of Users Thought It Was Human — Puzzleheaded-King584 · 2026-10-03
- Using System One models in Swift: fast, deterministic decisions via Apple Foundation Models — rxwei · 2026-10-03
- Linux Kernel CVEs Surge From ~500 to 1500+ Per Release, LLMs Blamed for Bulk of the Rise — burny_tech · 2026-10-03
- Developer Complains OpenAI's Coding Model Endlessly Scopes Creeps Instead of Finishing Tasks — DavidWells · 2026-10-03
- Sonnet 5 Spotted in Google Antigravity Backend, Which Still Runs Sonnet 4.6 — brandon_galang · 2026-10-03
- Post-training Yandex AliceAI-80B-A3B from scratch: a NaN bug in custom V100 kernels killed one run — jjusko20 · 2026-10-03