Fanfiction site AO3 sees red-screen detector for AI-generated works flag users

A browser skin for the fanfiction archive Archive of Our Own (AO3) highlights pages where text was pasted directly from Anthropic’s Claude by turning the background red, according to a report by The Verge. The anonymous creator of the skin said the red screen appears because Claude injects a specific HTML class—'font-claude-response-body'—when its output is copied directly into AO3’s editor, and its presence indicates Claude use definitively.

The creator posted examples of flagged fanworks to demonstrate the detector, but emphasized it was not meant to publicly accuse users. Despite that warning, community members have already used the tool to publicly name and shame writers whose works displayed the red screen.

The detector only works if text is copied straight from Claude into AO3; it will miss edits made in external tools like Google Docs or Microsoft Word before pasting, and it cannot reveal how much of a work was AI-generated. A single pasted sentence for spell-checking can trigger the red screen just as a fully AI-written story would.

At least one other community member claims to have written code that can detect ‘Claude, Deepseek, and some ChatGPT’ usage, but that solution has not been publicly released. Anthropic did not respond to a request for comment on whether the detector works as described.