Anthropic Embeds Invisible Text Watermarks and C2PA Provenance in Claude
dr_cintas · x · 2026-08-11
Anthropic's official documentation reveals that Claude now features a dual invisible watermarking system designed for AI content traceability and tamper-proofing.
- Text Watermark: Applied directly at the model generation level by subtly weaving it into word choices. It is imperceptible to humans, preserves the original meaning and readability, and is active across all interfaces (API, web, and cloud partners).
- File Provenance: Automatically embeds signed metadata based on the C2PA open standard into generated files (e.g., PNG, JPG, SVG) to verify authenticity and detect tampering.
Models released on or after August 2, 2026, will include these watermarks from day one, while older models will gain support during a transition period.
More from Models
- Using System One models in Swift: fast, deterministic decisions via Apple Foundation Models — rxwei · 2026-10-03
- Linux Kernel CVEs Surge From ~500 to 1500+ Per Release, LLMs Blamed for Bulk of the Rise — burny_tech · 2026-10-03
- Developer Complains OpenAI's Coding Model Endlessly Scopes Creeps Instead of Finishing Tasks — DavidWells · 2026-10-03
- Sonnet 5 Spotted in Google Antigravity Backend, Which Still Runs Sonnet 4.6 — brandon_galang · 2026-10-03
- Every's Dev Day chat with Matthew Berman: budget gone the moment he tried Ultrafast — every · 2026-10-03
- Wish list: a Qwen4 27B with 100B+ Engram offloaded to RAM and NVMe for local users — casper_hansen_ · 2026-10-03