Anthropic Adds Watermarks to AI-Generated Text as Rollout Begins
The story in brief
Anthropic is rolling out invisible watermarks to text generated by its Claude AI models, starting August 2, 2026. The feature, based on the C2PA standard, will travel with text when copied and pasted, surviving some editing. It aims to help identify AI-generated content without affecting readability or quality. The system covers multiple Claude products and interfaces, as the company continues to refine its detection tools.