Watermarking
5 pieces on Watermarking, including 1 step-by-step guide.
Guides
News & Analysis
SynthID Bio Watermarks AI Proteins That Still Work in the Lab
Google DeepMind's SynthID Bio embeds verifiable signatures into AI-generated protein sequences and 3D structures while preserving biological function across three wet-lab targets.
Read more →
Claude Watermarks: Two Systems, Three Output Types, One Gap
Claude uses SynthID-Text for prose and signed C2PA metadata for images — but code sits in a weak middle ground with fewer statistical degrees of freedom.
Read more →

Anthropic Explains How Claude's SynthID-Text Watermarks Work
Anthropic details Claude's SynthID-Text watermarking: how it survives edits, why code gets negligible signal, and what a detection API means for developers.
Read more →

Anthropic to Watermark All Claude Text and Images for EU Compliance
Anthropic will embed invisible watermarks in Claude-generated text and C2PA metadata in images, in response to EU AI Act transparency rules active from August 2nd.
Read more →