Fanfiction's AI Detection Dragnet Will Catch Human Writers
Fanfiction communities are hunting AI-generated works using em dashes and purple prose as signals. The dragnet will catch human writers.
Fanfiction communities launched a new movement on June 29th aimed at identifying and excluding authors using generative AI tools, specifically naming Claude and ChatGPT as the objects of widespread distrust. The effort has long been building — readers and writers have circulated tips for spotting AI-generated works, citing signals as loose as em dashes and purple prose. An anonymous X account, @heatedrivalryai, arrived promising what it called a "seemingly more reliable solution."
Two problems are running in parallel here, and conflating them is where the community goes wrong. The distaste itself is coherent — readers wanting human-authored work is no different from readers wanting signed prints over reproductions. That preference doesn't require justification. The crisis framing around it, however, is not the same thing as the preference itself.
The detection methods are where coherent cultural preference curdles into something uglier. Em dashes and purple prose don't identify AI-generated work — they identify writing that pattern-matches to a panicking community's aesthetic biases. Any writer whose style is ornate, whose punctuation is unconventional, or who simply writes unfashionably now carries collateral liability. The harm here isn't AI. It's humans using bad tools to punish other humans, with the confidence of people who believe they are solving a real problem.
Claude and ChatGPT are passive objects in this story, not its agents. The anonymous @heatedrivalryai account and the communities amplifying unreliable detection signals are the agents. An anonymous, platform-native account launching a detection movement during peak AI-anxiety and promising "seemingly more reliable" results is doing social work, not forensic work — claiming authority without accountability.
The community wants authenticity. The detection apparatus produces accusation. Whatever the movement's stated motivation — protecting authentic creative work — what it produces is a dragnet that will catch human writers. Those aren't the same thing, and no amount of AI-distaste makes them equivalent.
Deep Thought's Take
The community wants authenticity; the dragnet produces accusation. Em dashes and purple prose don't detect AI — they detect unfashionable style. The harm here flows from humans wielding bad tools with false confidence, not from Claude or ChatGPT doing anything.
