Claude text watermark: what changes on November 3, and what it means for your script
· AI · PLAYCODE3
On November 3, Anthropic starts adding its text watermark to four more Claude models: Claude Mythos 5, Claude Opus 4.7, Claude Opus 4.6 and Claude Sonnet 4.6. Newer models, like Claude Sonnet 5.5, already carry it. The watermark is a hidden pattern in word choice. You cannot see it, and it does not change the meaning, quality or length of the text. It holds no information about you or your conversations. To find it, a detection tool reads the text and looks for the pattern. Right now that tool is in a private preview for groups like regulators, newsrooms, schools and researchers, ahead of new EU AI rules in December. Anthropic says a match only means the text may have passed through Claude, not that Claude wrote it, and very short, heavily edited or rewritten text may show no mark at all.
In Screenplay Assistant, Format runs on Claude Sonnet 5.5, so the lines it puts on your page already carry the watermark today. A few smaller jobs use older models, Claude Sonnet 4.6 and Claude Haiku 4.5, and the rollout begins to reach Sonnet 4.6 on November 3. Our AI formats and polishes. It does not write your story. But a detector only sees the finished text, so it could find a mark on a page you wrote yourself. That is why we built the Writer’s Record. Screenplay Assistant was there for every word, so it counts what came from you and what came from the app while you work, and prints it on one page. It is a record of the work, not a guess. It already names the AI models we use and which ones carry a watermark, and we will update that line as this change rolls out.
Read more
These links go to other sites.
- How Claude marks AI-generated content (Anthropic's help center) (support.claude.com, opens in a new tab)