AI-generated books gain market share in Amazon self-publishing despite lower per-title sales
A large-scale empirical study of self-published Amazon fiction from 2023-2026 reveals that AI-generated books, despite lower individual sales rates, are capturing growing market share and displacing human-authored works from top rankings. Using full-text detection across 14,419 titles, researchers found that books with over 25% AI content remain commercially marginal per title but collectively represent a scaling phenomenon, suggesting generative AI is reshaping publishing economics rather than remaining confined to low-visibility niches. The finding challenges assumptions that AI-produced content lacks commercial viability and signals a structural shift in how algorithmic ranking systems allocate visibility in digital marketplaces.
Modelwire context
Analyst takeThe more pointed finding isn't that AI books sell poorly per title, it's that Amazon's ranking algorithms appear to be amplifying volume-based flooding as a viable market strategy, effectively turning a discovery mechanism into a vulnerability. The study's 2023-2026 window captures a structural shift already in progress, not a hypothetical risk.
None of the related Modelwire coverage from this week connects directly to publishing economics. The closest thematic thread runs through the LKValues benchmark piece, which flagged how AI systems embed assumptions that disadvantage non-dominant groups, and a similar dynamic is visible here: ranking algorithms designed for human-authored scarcity behave differently under AI-generated abundance, with costs falling on mid-list human authors rather than on the platforms. The interpretability work covered in 'Train the Model, Not the Reader' is a reminder that we also lack reliable tools to audit how Amazon's recommendation and ranking models actually weight signals, which matters if regulators or publishers want to verify whether flooding is being algorithmically rewarded or merely tolerated.
Watch whether Amazon updates its self-publishing content policies or ranking criteria within the next two quarters. A policy change would confirm the platform recognizes the displacement dynamic; silence would suggest the current economics suit the marketplace even if they don't suit authors.
Coverage we drew on
This analysis is generated by Modelwire’s editorial layer from our archive and the summary above. It is not a substitute for the original reporting. How we write it.
Modelwire Editorial
This synthesis and analysis was prepared by the Modelwire editorial team. We use advanced language models to read, ground, and connect the day’s most significant AI developments, providing original strategic context that helps practitioners and leaders stay ahead of the frontier.
Modelwire summarizes, we don’t republish. arXiv cs.CL originally reported this story as “Generative AI floods and dilutes the market for books”. The full content lives on arxiv.org. If you’re a publisher and want a different summarization policy for your work, see our takedown page.