What happened
Anthropic has released a global watermarking feature for Claude. The system embeds hidden signals into text generated by the model. These markers identify the output as AI-created content. The feature applies across global Claude deployments.
The context
Watermarking embeds tracking directly into model outputs to address misinformation and plagiarism. Major AI developers are adding these features to meet growing demands for content provenance.
Sources
- Computing UK ↗ via Google News Reported
See Anthropic's Vikshy Score → · Anthropic's models & pricing →