Remove Claude Watermarks

Anthropic recently announced it will be adding statistical watermarks to text generated by models released after August 2, 2026 (and will try to add them retroactively to earlier models) to make the text machine-identifiable as being Claude-written. I know it’s too early yet to know exactly how this will be implemented, but once it is implemented, I’m sure many writers would appreciate being able to remove these watermarks from their text (or at least determine whether they are detectable within a project so the text can be edited). https://support.claude.com/en/articles/16266773-how-claude-marks-ai-generated-content