AI is changing literary production fastest in self-publishing, where platforms reward frequent releases. Romance writer Coral Hart says she has used AI to produce 370 novels under 21 pen names since 2025, earning a high six-figure income; she argues that tools speed up work but cannot replace an author’s understanding of craft and readers. An unreviewed study estimates that roughly one-fifth of recent self-published books contain at least 25% AI-generated text. The practice provokes anger over both threatened livelihoods and the use of writers’ work to train systems without permission. A settlement of $1.5bn in a case involving about 500,000 titles, alongside ongoing lawsuits, has sharpened those concerns. Seventy prominent authors have urged the biggest publishers not to release machine-created books, but received no such commitment.
With no settled rules, AI accusations have led to withdrawn deals and prize exclusions. Hachette dropped the romance novel Shy Girl; Commonwealth prize winner Jamir Nazir, debut novelist Jerry Falade and Canadian-Haitian author Thélyson Orélien faced allegations, which they denied, though Orélien acknowledged plagiarism elsewhere. The cases raise questions about whether AI detectors can establish authorship and whether publishers should cancel books rather than investigate with writers. Although research cited in the account gives one leading detector high accuracy, no test definitively proves how a text was written; non-native English speakers and neurodivergent writers may face particular suspicion. The fact that all authors at the center of these scandals were people of colour also prompts concerns about unequal scrutiny. In response, agents advise writers to preserve drafts and document their process, while the industry remains divided over disclosure, detection and fair procedures.
Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.