project
Google DeepMind generative text watermarking approach that modifies next-token sampling to embed a detectable statistical signature.
No evidence-backed relationships are recorded.
Anthropic now watermarks Claude's text with a version of Google DeepMind's SynthID-Text to meet the EU AI Act's transparency rules. Only a holder of the secret key can measure the mark, a lean in token choice that says little until it spans many tokens.
Reality
Anthropic says it is watermarking Claude text to satisfy the EU AI Act Transparency Code and plans a detection API. The exposure sits with teams shipping generated work as their own.
Perspective Coverage
7 publishersReality
The mark is applied during generation rather than at an interface, so anything built on the Claude Platform API returns marked text by default. Reading the mark stays behind a private detection API with a registration list.
Reality
Anthropic's text mark is a keyed scheme no third party can test. The C2PA credentials on images since 11 August are checkable by anyone, until a PDF generator strips them.
Reality
The mark lives inside token sampling, is invisible to readers, and needs a key to read. Provenance testing just became an operational question for DLP and insider-risk teams.
Reality