Anthropic to Embed Invisible Text Watermarks in Claude Outputs Globally
量子位 · wechat · 2026-08-11
Anthropic announced that to comply with the EU AI Act, new Claude models will embed invisible watermarks directly into generated text. These watermarks will persist through copying and moderate editing, and will be implemented globally, not just in the EU.
In addition to text watermarks, Claude will append C2PA-based signed metadata to generated files like images. This mechanism covers the entire product line, including the API and Claude Code. Anthropic stated the watermark won't affect text quality and plans to release detection tools later.
The move sparked intense community backlash. While a reliable AIGC detection system is necessary, many users are concerned about covert tracking, and developers are already brainstorming third-party tools to strip the watermarks.
More from Models
- First Model to Pass the Video Turing Test: Half of Users Thought It Was Human — Puzzleheaded-King584 · 2026-10-03
- 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
- Post-training Yandex AliceAI-80B-A3B from scratch: a NaN bug in custom V100 kernels killed one run — jjusko20 · 2026-10-03