New detection tool sparks controversy over AI-generated fanfiction on Archive of Our Own

1. Rising Tensions Over AI in Fanfiction

The fanfiction community is experiencing a surge of internal conflict regarding the use of generative AI tools like ChatGPT and Claude to create fanworks. While there has long been a general distaste for AI-generated content within these creative spaces, recent efforts to identify and root out such works have intensified. Readers and writers have previously attempted to spot AI-generated text by looking for stylistic markers, such as specific prose patterns or punctuation habits, but these methods have often been viewed as unreliable.

2. New Detection Method Introduced

On June 29, an anonymous X account operating under the handle @heatedrivalryai introduced a new tool aimed at identifying AI-generated fanfiction on the popular repository Archive of Our Own (AO3). The tool functions as a site skin—a custom interface modification—that claims to detect specific coding artifacts left behind when text is copied directly from Anthropic’s Claude bot into the AO3 platform. According to the account, the presence of a specific code snippet, "font-claude-response-body," serves as definitive proof of AI usage. When the skin detects this code on a fanfiction page, it changes the background color to red to alert the reader.

3. Reliability and Community Impact

The effectiveness and ethics of these detection methods remain a subject of debate. While some users have successfully tested the skin against known AI-generated examples, the reliance on automated detection tools has raised concerns about potential false positives and the broader impact on the community. As readers continue to experiment with these detection skins, there is growing apprehension that any fanfiction writer could be unfairly targeted or caught in the crossfire of this ongoing effort to police AI-generated content.