Generative AI floods and dilutes the market for books

2026-07-05 · Source: cs.CL updates on arXiv.org · Field: Media & Entertainment — Publishing & Journalism, Content Creation & Production, Digital Media & Streaming · Depth: Expert, extended

Summary

A study analyzing 14,419 self-published genre-fiction e-books on Amazon from January 2023 to June 2026 reveals that generative AI is diluting the market, reshaping it through sheer volume rather than quality. Books with substantial AI text (over 25% detected AI content) comprise 20% of the catalog but only 12.1% of sales and 11.3% of revenue. Despite lower average sales, their share of observed sales grew to approximately 20% by Q2 2026, increasingly occupying top-rank positions. The overall catalog expanded 38.3-fold, and selling books grew 19.2-fold, while quarterly revenue increased only 8.9-fold. This led to a decline in revenue per selling book across most genres, including for books with no detected AI text, particularly in genres with high AI diffusion and Kindle Unlimited availability. Successful AI-generated books also exhibit higher overlap with rare, distinctive language from existing works, suggesting a a derivative production channel.

Key takeaway

For legal professionals assessing copyright infringement and fair use defenses, this study provides empirical evidence of market dilution. The massive entry of AI-generated books, even if individually low-selling, demonstrably reduces revenue and top-rank positions for human-authored works. This supports the "market dilution" theory, suggesting that the sheer volume of substitutable AI content, enabled by upstream copying, can significantly harm existing markets, irrespective of comparative quality. You should consider these findings when evaluating claims of market effect in AI training litigation.

Key insights

Generative AI reshapes creative markets by volume, diluting returns for human-authored works, not by quality.

Principles

• Market dilution can occur via volume, not just quality.
• High AI diffusion correlates with reduced human author returns.
• Successful AI books reuse distinctive existing language.

Method

The study used Pangram v3.3 for full-text AI detection across 14,419 e-books, tracking daily sales and classifying content into no, light, or substantial AI text to analyze market effects.

In practice

• Monitor genre-specific AI diffusion rates for market impact.
• Evaluate content for rare expression overlap as an AI indicator.
• Consider Kindle Unlimited's role in AI-driven market saturation.

Topics

• Generative AI
• Book Market
• Copyright Law
• Market Dilution
• AI Content Detection
• Self-Publishing

Code references

• tuhinjubcse/BookSlop

Best for: Research Scientist, Legal Professional, Consultant

Related on AIssential

• Generative AI floods and dilutes the market for books — Generative AI is rapidly diluting the self-published book market, gaining sales and top ranks through sheer volume, not quality.
• AI has expanded the middle and lower parts of the book market much more than the top. — AI lowers content creation costs, increasing output but decreasing average quality, shifting value to trust and curation.
• How Gen AI is Disrupting B2B Buying Decisions — Generative AI is fundamentally shifting B2B buying into AI-mediated "dark funnels," demanding new go-to-market strategies.
• On the troubling rise of generative AI suspicion in academic publishing — Overextended suspicion of GenAI in academic publishing leads to false positives and an "AI witch hunt."
• ByteDance backpedals after Seedance 2.0 turned Hollywood icons into AI “clip art” — AI video generation models face significant copyright and ethical challenges when using existing intellectual property.
• Penguin v. OpenAI: AI-assisted production of market substitutes in the very formats that already plague children’s publishing (fake titles, lookalike covers, rapid self-publishing). — AI models reproducing copyrighted content on demand can constitute actionable copyright infringement.

See Counsel's argued verdicts on the open AI decisions leaders are weighing →

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by cs.CL updates on arXiv.org.