A new study of over 14,000 self-published e-books on Amazon shows that AI-generated content is quietly reshaping the entire publishing market by leveraging its sheer volume. Data indicates that the size of Amazon's book catalog increased by 38.3 times from the beginning of 2023 to the first quarter of 2026, while the platform's quarterly total revenue only grew by 8.9 times during the same period. This means that more and more books are competing for a slowly growing, limited revenue pool, leading to a noticeable "dilution effect" in the market.

What is concerning is that this impact is not limited to low-quality content; even ordinary books without any detectable AI text have also been affected. In the eight major book genres covered in the statistics, the income per book for non-AI texts declined in seven categories. The study points out that books with a large amount of AI content account for 20% of the catalog but contribute only 12.1% and 11.3% to sales and revenue, respectively. Conversely, books without detectable AI text still account for 72.5% of mainstream revenue. However, as AI books continue to climb to the top of bestseller lists, the proportion of AI works in the new Top 25 list has reached as high as 31%.

Additionally, the analysis found that some prolific authors have significantly increased their output using AI, with some authors earning millions of dollars from just a few books. In terms of language features, best-selling AI books show a high degree of overlap with existing published works, relying much more on rare expressions than typical literary works.

Industry experts believe that this study provides key empirical data for ongoing AI copyright lawsuits. In the past, AI companies often defended themselves by citing human reading and learning as justification. However, evidence of market dilution and linguistic overlap may have a profound impact on future judicial judgments regarding fair use.