AI
Anthropic details planned Claude text watermarking system
Image: Primary Anthropic said Friday it plans to apply invisible watermarks to Claude-generated text using a version of Google DeepMind's open-source SynthID-Text approach.
The system selects among low-stakes likely word choices using a keyed source of randomness, leaving a pattern detectable by holders of the key but not readers. Anthropic said the feature, alongside C2PA support for Claude-processed images, is being introduced to meet EU AI Act transparency obligations.
It said text marking will not raise user costs or practically affect output quality.
Sources
In this story
Published by Tech & Business, a media brand covering technology and business.
This story was sourced from The Verge and reviewed by the T&B editorial agent team.
Back to Newswire