Trend shift

AI Book Supply Starts Diluting Author Income

A study of Amazon’s self-publishing market finds AI-generated book supply rising sharply while per-book revenue for human authors falls in most genres. Copyright disputes may increasingly turn on measurable market dilution.

By All We Need Editorial System

Amazon’s self-publishing market may be undergoing a commercially meaningful shift: AI-generated book supply has grown faster than sales and revenue, while per-book revenue for human authors has declined in most studied genres. If the relationship holds across more platforms and periods, copyright disputes may extend beyond training legality to the market effects of generated output.

Supply is growing faster than demand absorption

The Decoder cites a study finding that AI-generated books account for about one-fifth of Amazon’s self-published catalog but only 12% of sales. The paper’s comparison of 2023 through 2025 release cohorts reports 38.3-fold growth in AI-related supply, versus 7.3-fold sales growth and 8.9-fold revenue growth. That gap suggests more than new reader demand: revenue available per work may be spread across a much larger catalog.

Dilution travels through platform distribution and pricing

Self-publishing platforms lower the cost of listing, cover creation, translation and promotional copy, allowing generative tools to place more titles into the same search, recommendation and advertising inventory. The study finds lower per-book revenue for human authors in seven of eight genres. The effect need not depend on readers considering AI prose superior; additional supply can compete for finite attention, discovery and spending. Catalog expansion and author unit economics may therefore be moving into tension.

Market data may enter copyright harm calculations

The U.S. Copyright Office has said that sales substitution, market dilution from stylistically similar outputs and viable licensing markets can matter to fair-use analysis. Quality is not the only explanation: another reader study found that AI text fine-tuned on complete author works could receive stronger preferences on some measures. The strongest alternative explanation is that revenue changes reflect demand, ranking systems or changes in the author mix rather than AI supply; that makes observable market data more important than abstract quality claims.

What to watch next

Watch for genre-level disclosures on AI labeling, impressions, conversion and author income from Amazon or other self-publishing platforms, and for replication across platforms, price bands and longer time windows. Evidence that author income improves after ranking, labeling or listing-rule changes would strengthen the supply-pressure mechanism. If the effect disappears after controlling for genre and demand changes, the claim would weaken.

Sources