Anthropic shares more details about how Claude’s new watermarks will work

Anthropic has provided more information about the workings of Claude's upcoming watermarking, addressing how it functions, whether editing can hide it, and its impact on code.

By AIWI EditorialReviewed by Chayanika DivekarAI-assisted summary · human-reviewed1 min readPublished Aug 15, 2026

Single source1 source

Reality Check

Availability: Announced only
Hype-o-meter: Mostly substance
Who it's for: Developers · Enterprises
Do this now: Review Anthropic's details on watermarking if you use Claude for content or code generation.
Claims check: Input only indicates that Anthropic shared details; specifics of the mechanism are not provided.
Corroboration: Single source · 1 source

Key Takeaways

Anthropic detailed how Claude's new watermarking will operate
Open questions include whether watermarks survive editing
The effect of watermarking on code output was raised

Anthropic has released more details about the watermarking system planned for its Claude models. According to the report, the company addressed several practical questions, including how the watermarking will actually function, whether it can be removed or hidden through editing, and how it applies to code generated by the model.

Why it matters

Watermarking is a proposed method for identifying AI-generated content, and its resilience to editing and its behavior with different output types such as code are central to how useful it will be.

Who should care

Users of Claude, developers who rely on the model for code, and those tracking approaches to identifying AI-generated content may find these details relevant.

Was this helpful?