- The self-publishing market is witnessing a wave of books written by generative AI, causing many authors to fear that their chances of being discovered by readers are shrinking.
- Author Jess Michaels has published 117 novels since 1999 and has had 10 bestsellers according to USA Today, but her sales have recently plummeted.
- She believes her works were scraped to train generative AI, which then produced books in a similar style, saturating the market.
- Initially, many believed that fiction would be a difficult field for AI to replace because readers value the emotions conveyed by humans.
- A new study shows that a significant proportion of self-published novels on Amazon contain text generated by chatbots.
- The research team, led by Professor Tuhin Chakrabarty at Stony Brook University (SUNY), used the Pangram tool to analyze data.
- After excluding low-selling books, the group randomly selected 14,419 novels released between January 2023 and March 2026.
- The results identified 2,168 books with over 50% generative AI content, of which 970 books exceeded 90%.
- Researchers estimate that novels containing at least 25% generative AI text have brought in tens of millions of dollars in revenue over three years.
- The study has not yet been peer-reviewed but reinforces the observation that generative AI books have become an unavoidable part of the retail market.
📌 The speed at which generative AI is penetrating the publishing industry is having a direct impact on traditional authors. Data from 14,419 novels shows that 2,168 have over 50% generative AI content and 970 exceed 90%, while the group of books containing at least 25% AI text generated tens of millions of dollars in revenue over three years. This indicates that distinguishing between books written by humans and AI is becoming increasingly difficult.
