Anthropic says it has introduced an invisible watermarking system for Claude AI that can be embedded into AI-generated text and files. The aim is to help identify the origin of content and make it possible to trace AI-produced material even after it is shared across different platforms.
Medium reports that the new method “quietly” marks content without adding visible text for users, positioning watermarking as a way to support provenance and attribution. NDTV similarly describes the development as an update that enables Claude to embed invisible watermarks in AI-generated content.
Across the two outlets, the central focus is the same: Anthropic’s Claude adds an invisible marker intended to persist through distribution. Neither outlet, in the provided excerpts, specifies the technical approach, how robust the watermark is to editing or copying, or when the feature becomes broadly available. The reporting also emphasizes attribution rather than any user-facing changes to how outputs are generated or displayed.
