This story is related to the first of the week. I promise(ish—breaking news depending) no AI on Saturday, but the issues around Tuhin Chakrabarty's hugely impactful study on AI in self-published novels are so important and widespread I think it is warranted. Especially as this great piece in the New York Times brings out another angle of the story (with thanks to Jane Friedman, whose newsletter gifted the link from behind the NYT paywall).
ALLi News Editor Dan Holloway
The article takes a deep dive into Chakrabarty's paper, which you can access for yourself here. The paper is titled “Generative AI Floods and Dilutes the Market for Books.” That makes very clear that its aim is not to name and shame indie authors (or others) for using AI but to describe a landscape in which the use of AI might diminish the ability of those who don't use the technology to sell to audiences the way they once did. The NYT article sets this into context through the lens of indie author Jess Michaels, who reports a cliff-style drop-off in sales since ChatGPT's sophistication grew.
The study's abstract makes its intention explicit. It aims to test the assumption that AI-generated books are “low-quality ‘slop' that buyers will ignore, and are assumed to carry little commercial weight.”
The setup was as follows: 14,419 self-published titles published between 2023 and 2026 were run through an AI detection program. No titles had declared AI use. Sales were also tracked.
The headline results all make for interesting and important reading. They are alarming but not (yet) the existential crisis one might fear.
First, when it comes to some numbers, the survey found what we had suspected from that four million books figure that made so many headlines earlier in the year. Titles increased almost 20 percent while sales only increased less than 10 percent. That is, the pie didn't increase at the same rate as the slices it was cut into.
Second, the detection software found a significant prevalence of (alleged) AI use (more than a quarter of titles—that is, about the number that would explain the increase people have been so worried about).
Third, and importantly, sales of the titles displaying signs of AI use were less than those of other titles, but nonetheless at a significant level, sufficient to suggest commercial issues. One might speculate that one reason for the lower sales is reader loyalty to established authors. That's encouraging, though Michaels's story suggests it's not salvation. And I would suggest that as time goes on that effect may get diluted as readers find new “authors.”
I highly recommend reading the whole study.
Thoughts or further questions on this post or any self-publishing issue?
If you’re an ALLi member, head over to the SelfPubConnect forum for support from our experienced community of indie authors, advisors, and our own ALLi team. Simply create an account (if you haven’t already) to request to join the forum and get going.
Non-members looking for more information can search our extensive archive of blog posts and podcast episodes packed with tips and advice at ALLi's Self-Publishing Advice Center.
