Generative AI floods and dilutes the market for books
This paper demonstrates that generative AI is reshaping the book market not by displacing human authors through superior quality, but by flooding the market with low-cost AI-generated content that captures significant sales volume and top rankings, thereby diluting revenue per book and disproportionately harming non-AI works in high-diffusion genres.
Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
Imagine a giant, bustling marketplace where people sell stories. For a long time, making a new story was like baking a cake from scratch: it took time, effort, and skill. But then, a magical machine arrived that could bake a cake in seconds for almost no cost. Suddenly, the market was flooded with thousands of these instant cakes. Some people worried that because these cakes were made so fast, they must taste terrible and nobody would buy them. They thought the market would just ignore the "slop" and stick to the real, human-made cakes. But here is the tricky part of science: just because something is cheap and low-quality on average doesn't mean it can't change the whole market. If you throw enough cheap things into a room, they might crowd out the expensive things, even if nobody actually likes the cheap ones very much. This paper dives into that exact question: When a flood of AI-written books hits the shelves, does it just sit there unnoticed, or does it actually push human writers out of the way?
The researchers decided to investigate this by looking at the biggest online bookstore in the world, Amazon. They gathered a massive collection of 14,419 self-published genre fiction books (like romance, sci-fi, and mystery) released between 2023 and 2026. They used a super-smart detector to read every single word of these books and figure out how much of it was written by a human and how much was written by AI. They didn't just look at the books; they also tracked exactly how many copies sold and how much money they made every single day.
What they found was a bit like a crowded concert where the stage is getting smaller. First, they discovered that AI books are everywhere. About 20% of all the new books in their sample had a lot of AI text (more than 25% of the words). However, these AI-heavy books didn't sell as well as the human-written ones. They only made up about 12% of the total sales. So, the idea that "nobody buys AI books" is sort of true, but not the whole story. Even though they sell less individually, there are so many of them that they are starting to take up the most valuable spots on the bestseller lists.
The real shocker, though, is what happened to the human writers. The researchers found that the number of new books released grew 19 times faster than the total money people spent on books. Imagine if the number of people trying to sell lemonade doubled every week, but the amount of money people had to buy lemonade only went up a little bit. The result? The money each lemonade stand makes drops. That is exactly what happened here. Because the market was flooded with so many cheap AI books, the average amount of money earned per book went down for everyone, including the human authors who wrote their books entirely by hand. The human writers didn't just lose to the AI books; they lost because the whole market became too crowded.
The study also looked at why some AI books did manage to become bestsellers. They found that the successful AI books were like masterful copycats. They didn't just write random gibberish; they used very specific, rare phrases and descriptions that they had "stolen" from other existing books. The more money a successful AI book made, the more it seemed to be recycling these unique phrases from old stories. In contrast, the best human-written books didn't show this same pattern of recycling rare phrases to get rich.
So, the big takeaway is that Generative AI is reshaping the book market not by being better than humans, but by being more. It's a game of volume. By producing thousands of books that are "good enough" and cheap to make, AI is crowding out human writers, lowering their earnings, and taking over the top spots on the charts. It's not that the AI books are winning because they are masterpieces; they are winning because there are just so many of them that they are drowning out the competition. This is a big deal for copyright law, too, because it shows that flooding a market with cheap, substitutable copies can hurt the original creators, even if those copies aren't perfect.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.