Anthropic explains how Claude’s invisible text watermarks will work

Anthropic has clarified however it's readying to use invisible watermarks to Claude-generated substance successful bid to comply with Europe's AI transparency rules. On Friday, Anthropic announced that Claude's substance marking strategy is "a mentation of the SynthID-Text approach" - an open-source watermarking exertion developed by Google DeepMind that creates detectable patterns utilizing wording probabilities.

This watermarking feature, alongside C2PA support for Claude-processed images, is being introduced to conscionable Anthropic's obligations nether the European Union's AI Act, which requires synthetic audio, image, video, and substance to see machine-readable marks …

Read the afloat communicative astatine The Verge.