A New York Times reporter recently got a text from a new acquaintance asking why she had never mentioned her own biography. She had never written one, never approved one, and had never heard of the publisher. Yet there it was on Amazon, 90 pages long, sold under the name of an author who appeared not to exist.
That story is one of the sharpest examples of a problem that has been building for years. Authors started raising alarms about AI copycat books back in early 2023, and the volume has only grown since.
This guide walks through 15 documented cases, the numbers behind them, and a checklist you can use before you buy. Every case comes from published reporting or research, and I linked the sources so you can read them yourself. AI can be a legitimate writing tool, and many authors use it openly. The cases below are about mass-produced, misleading listings, not about AI use in general.
“Amazon is only one corner of a much bigger content flood. See how the AI slop economy works across platforms.”
- The Seagull Book That Climbed a Romance Bestseller Chart
- The Reporter Who Found Her Own Fake Biography
- The Retiree Who Publishes Hundreds of AI Books
- The 60 Biographies From One Author Name
- The Jump to 300,000 New E-Books a Month
- The Self-Help Category Where 77% Looked AI-Written
- The Baseball Biography That Got Nine Lookalikes
- The Author Who Found Five Knockoffs in Launch Week
- The Fake Copies of a Tech Reporter's Book on Apple Books
- The AI Disclosure Rule Shoppers Never See
- Amazon's ID Checks and Upload Limits
- What the Authors Guild Wants Amazon to Change
- How AI Books Squeeze Kindle Unlimited Authors
- Seven Warning Signs to Check Before You Buy
- Your Pre-Purchase Checklist
- Frequently Asked Questions (FAQ)
- Final Takeaway
1. The Seagull Book That Climbed a Romance Bestseller Chart
In late June 2023, indie romance author Caitlyn Lynch scrolled through Amazon's Teen & Young Adult Contemporary Romance chart and noticed something wrong. By her count, only 19 of the top 100 titles looked like real books. The rest appeared to be machine-generated nonsense built to farm clicks.
One of them, titled "wait you love me," sat at number 90. Its cover was a black-and-white seagull photo with a neon yellow title bar. It had two reviews, and both were one-star.
- ✅ Both reviews called it a fake AI book
- ✅ It ranked above real novels written by real people
- ✅ Most of these titles dropped off the chart within days, yet stayed on sale
Lynch warned that Kindle Unlimited pays writers by pages read, so a flood of junk could shrink the payout for everyone. Her warning has aged well.
“Streaming platforms hit the same wall with music. Read what happened in Spotify's AI music purge.”
2. The Reporter Who Found Her Own Fake Biography
In July 2026, Kashmir Hill of The New York Times went to Amazon to investigate that text message. She found "The Biography of Kashmir Hill," released in August 2025 under the name John Crane Miller. Her own account of the discovery is a tour of how these listings work.
The book ran 90 pages. It mixed widely available facts, like where she grew up, with generic insights that could describe anyone. The author page showed a common stock photo of a man in a suit. Miller had published 10 biographies in a single week, all about journalists.
- ✅ The book had no reviews until Hill posted one asking the author to contact her
- ✅ Soon after that review, the listing and all his other titles disappeared
- ❌ She never got a reply or any contact details
Amazon told the Times it invests in enforcing its guidelines and removes books that break them.
“Fake personas are becoming a pattern online. See the rise and fall of Emily Hart, an AI influencer built on a made-up identity.”
3. The Retiree Who Publishes Hundreds of AI Books
Hill also profiled Bill Johns, a 70-year-old retired cybersecurity consultant from Maryland. He was open about what he does, which makes his story useful. According to the Times, he published his first ChatGPT-assisted book in March 2025 and now has 445 titles on sale.
The numbers he shared are modest but real:
- ✅ He sells a few hundred books a month, earning roughly $7 each
- ✅ One holiday season brought 821 sales and almost $6,000 in profit
- ✅ He described the income as "bar money"
- ✅ His goal is 10 new books a week, the maximum Amazon allows
Johns said he removes a book when a one-star review convinces him the reviewer is right. Readers who actually finished his books were often harsh. To be fair, nothing in the reporting suggests he broke any rule. That is the uncomfortable part of the story.
“Low effort and steady payouts drive this whole trend. The numbers in YouTube Shorts AI slop data show the same pattern.”
4. The 60 Biographies From One Author Name
While Hill was browsing, Amazon's recommendation system suggested another title, "Dan Rather: Truth and Tenacity" by an author named Diane W. Gray. The same Times report says Gray published 60 biographies in a single summer. The subjects ranged from Joe Rogan and Celine Dion to soccer star Paolo Maldini.
The Maldini book had a single review, and the reviewer wrote a blunt warning not to buy it. Mr. Rather is only the subject of the title. Nothing in the reporting says he read or endorsed it.
The lesson is about volume. A catalog of dozens of celebrity biographies from one name in one season is a pattern worth noticing before you click buy.
“Want the full spotting guide? Start with signs you're looking at AI slop.”
5. The Jump to 300,000 New E-Books a Month
Single cases are easy to dismiss, so it helps to look at volume. Two business school professors collected data on 10 million Amazon books published over five years. According to a reprint of the Times report, monthly e-book publishing roughly tripled after ChatGPT launched. It went from around 100,000 titles a month in 2022 to more than 300,000 by the end of 2025.
Amazon pushed back on that number. A spokesperson said its internal metrics did not show that level of growth, though the company would not share its own figures. Not every new title is AI-made, either, since self-publishing was already growing.
- ✅ The professors' dataset covered 10 million books
- ✅ The estimated rise was from about 100,000 to over 300,000 a month
- ❌ Amazon disputes the growth rate and has not published its own data
Even if the real number is lower, the direction matters. More uploads means more listings a shopper has to sort through.
“One person can now run a whole publishing operation. See how it works in the fake empire built on Facebook AI slop.”
6. The Self-Help Category Where 77% Looked AI-Written
In January 2026, the company Originality.ai published a study of Amazon's "Success" self-help subcategory. It analyzed 844 books published between August 31 and November 28, 2025. Using its AI detector, it estimated that 77% were likely written by AI.
The study compared those books with likely human-written ones:
- ✅ Likely AI books averaged 26 reviews, against 129 for human-written books
- ✅ They were priced about one dollar cheaper on average
- ✅ Around 12% of the books came from just 4% of the authors
- ✅ About 90% showed some likely AI text in their descriptions, bios, or samples
Read this one with care. Originality.ai sells AI detection tools, and detectors give probabilities, not proof. The word "likely" is doing real work in that 77%. Still, the review gap and the publishing pace are things you can check yourself on any listing.
“Big platforms are feeling the pressure too. Read how Facebook's AI slop problem grew into a company-level issue.”
7. The Baseball Biography That Got Nine Lookalikes
Pittsburgh author John W. Miller spent three years researching a biography of Orioles manager Earl Weaver, including more than 200 interviews. His book, "The Last Manager," came out in March 2025. According to Pittsburgh's WESA, several look-alike titles appeared around his release, including "Earl Weaver: The Science of Rage" and a German-language "Earl Weaver Biografie."
Some of the covers used AI-made images of men in Orioles uniforms who only vaguely resembled Weaver. One cartoonish cover made him look like a young man with a beard, which is not how anyone remembers him. The English "Science of Rage" title was credited to the Bill Johns name, which then had about 375 Amazon titles.
Miller said it was unclear how much the knockoffs cut into his sales, but he doubted they had no effect. He noted his book still became a bestseller.
- ✅ The real book took three years of research
- ❌ The lookalikes appeared within days of release
- ❌ Their covers and titles were built to catch the same searches
“Readers are getting tired of this. See why users are fleeing Pinterest's AI slop.”
8. The Author Who Found Five Knockoffs in Launch Week
The Weaver case is not unusual. When journalist Talia Lavin released her book "Wild Faith," she searched for it online the way most authors do in launch week. Rolling Stone reported that she found at least five complete rip-offs of her work on Amazon, almost all of them seemingly made with generative AI.
Rolling Stone described these knockoffs as one of the most widespread problems in online bookselling. Authors told the magazine the copies are frustrating and nearly impossible to stop. Lavin's frustration was practical: someone had spent about 20 minutes on prompts and was now competing with years of her work.
Here is what many authors now report about launch week:
- ✅ They search their own title the day it goes live
- ✅ They report each copycat by hand
- ❌ New copies can appear under different names soon after
This is the hidden cost. Real authors take on an unpaid second job just to protect their own names.
“Authenticity is a brand asset now. Read how AI content habits make brands look fake.”
9. The Fake Copies of a Tech Reporter's Book on Apple Books
Amazon is not the only store with this problem. Tech writer Joanna Stern found 10 fake versions of her book I Am Not a Robot on Apple Books within days of the real one going on sale. 9to5Mac's report says most were AI summaries that copied the structure of her chapters.
She got Apple to remove them, but three more fakes soon appeared. The same report describes lookalikes of Lena Dunham's Famesick and Haley Sacks's Future Rich Person. Some copycats even used AI images of women who resembled the real author on the cover.
- ✅ Ten copies appeared within days of launch
- ✅ Removal worked, but new copies followed quickly
- ❌ The copies were built to look like the original or a summary of it
The article also quotes the estimate that tens of thousands of these titles may exist across Apple Books and Amazon. That is a rough estimate, not an audited count. The pattern is what matters. Any successful book can attract copies within days.
“Copies that pose as the real thing use the same trick as AI deepfake scams, just in a bookstore.”
10. The AI Disclosure Rule Shoppers Never See
Amazon does have a rule about AI content, and it is easy to misunderstand. Publishers using Kindle Direct Publishing must tell Amazon whether text, images, or translations were AI-generated. Content that AI only helped with, like editing or brainstorming, does not need to be declared.
The catch is who sees the answer. Book Riot explained that the disclosure is made to Amazon and not to readers. There is no badge, filter, or warning on the product page. Amazon also added limits, such as a cap on how many books one account can publish per day.
So a shopper gets the same page for a book Amazon knows is AI-generated as for a human-written one. The rule exists, but it was built for Amazon's own oversight, not for the person deciding whether to spend money.
- ✅ AI-generated content must be declared to Amazon
- ✅ AI-assisted work does not need to be declared
- ❌ Nothing on the public listing shows the answer
“Labels and disclosure matter across the web. See what Google says about AI-written social captions.”
11. Amazon's ID Checks and Upload Limits
Amazon has added friction for people who publish at scale. In July 2024, KDP began asking some users to verify their identity, as Self-Publishing Advice reported. Amazon's own KDP help page on identity verification says it may ask for a government-issued ID, and that the name and country on the account must match it.
The idea is simple. It is hard to prove a manuscript was written by a person, but it is easier to prove a real person runs the account. That makes it harder to spin up endless anonymous accounts.
Be careful with claims beyond that. Amazon has not publicly explained how it detects AI-generated books, and talk of hidden "Verified Human" systems is speculation. What is documented is ID checks and upload caps. The upload cap of about 10 books a week is the one Bill Johns says he tries to hit.
- ✅ ID verification aims to cut fraud and abuse
- ✅ Upload limits slow down mass publishing
- ❌ Neither one shows shoppers whether a book is AI-made
“Platforms are getting stricter across the board. Read about YouTube's AI slop purge.”
12. What the Authors Guild Wants Amazon to Change?
The Authors Guild has pushed on this issue for years. In its article on the surge of sham books, the group said it wants Amazon to stop these books from being posted and to clearly label mostly AI-generated ones on the images or listings. It also noted that Amazon's disclosure requirement was not yet visible to customers.
That article dates from 2023, so treat it as the group's long-standing position and not a fresh 2026 statement. Its asks were practical:
- ✅ A visible label on listings for mostly AI-generated books
- ✅ Blocking sham listings before they go live, not after
- ✅ Legal tools, including a right of publicity covering an author's name and style
- ✅ Better AI-detection tools to help enforcement
Amazon did move on companion books. The Guild reported that Amazon disallowed that category except for limited exceptions. The core gap the Guild described, a public label, is still what readers ask about today.
“Labels matter most where the audience is vulnerable. See the truth about AI kids' videos.”
13. How AI Books Squeeze Kindle Unlimited Authors?
The money side explains why this keeps happening. Kindle Unlimited pays authors based on pages read, not on copies sold. Caitlyn Lynch warned in 2023 that junk titles could drag down the payout pool for everyone.
Writer Lincoln Michel described the incentive in his newsletter. His view is that some operators flood Kindle Unlimited with many titles aimed at different niche searches, hoping each one earns a few accidental page reads. That is one writer's analysis, not an Amazon finding, but it matches the volume patterns in the cases above.
For real authors, the damage shows up in three places:
- ✅ Copycats compete for the same search terms as their books, as Miller and Lavin found
- ✅ Review pages can get crowded with confusion about which book is which
- ❌ Every hour spent reporting copies is an hour not spent writing
The cost is not only money. It is trust in the marketplace itself.
“Automation without judgment hurts creators too. See how AI marketing fails destroyed brand trust.”
14. Seven Warning Signs to Check Before You Buy
No single sign proves a book is AI-generated, and plenty of good books have few reviews. Analysts of the self-help study suggested that reader vigilance about review counts, publication timelines, and author credentials can serve as practical quality checks. Put together, the reporting in this guide points to seven signals.
- A stock-photo author with a vague bio. Hill's fake biographer used a generic suit portrait and buzzwords like "seasoned biographer."
- A huge catalog published in a short window. Ten biographies in a week, or 60 in a summer, is not how research-based nonfiction gets written.
- Very few reviews on a book that claims expertise. The study found likely AI books averaged 26 reviews against 129.
- A price slightly below the real book. The same study found a gap of about a dollar.
- A title that mimics a real, recent release. Look for "companion," "summary," or a near-identical subtitle.
- Reviews warning you off. A one-star review saying "AI slop" is worth reading before you buy.
- Generic, horoscope-style writing in the sample. Hill described her own fake biography this way: facts anyone could find, plus insights that could fit anyone.
Use Amazon's "Look Inside" preview to read a few real pages. Vague, repetitive sample text costs you nothing to spot and can save you the price of the book.
“Spotting fakes is a skill. Build it further with signs you're looking at AI slop.”
15. Your Pre-Purchase Checklist
Here is a simple check before you hit "Buy Now." Reporting on the study also noted that some authors published several books within a few days, which is the kind of detail you can verify on the author page in seconds.
- ✅ Click the author name and scan the catalog and publication dates
- ✅ Compare the review count against the claims in the title and description
- ✅ Read at least two full pages of the "Look Inside" sample
- ✅ Search the title plus the word "review" to find outside opinions
- ✅ Check whether the book copies the title style of a recent bestseller
- ❌ Do not trust the cover or star rating alone
- ❌ Do not assume a familiar-sounding title is the original book
If three or more of these raise a flag, close the tab and look for the original. Buying from a publisher's own site, a library app, or a bookstore you trust also cuts the risk.
Frequently Asked Questions (FAQ)
Does Amazon allow AI-generated Kindle books?
Yes, as long as the book meets Amazon's content guidelines. Authors must tell Amazon when text, images, or translations are AI-generated. That disclosure is internal and does not appear on the product page.
How can I tell if a Kindle book was written by AI before buying it?
You cannot be certain from the outside. Check the author's catalog and publication dates, the review count, the price against similar books, and the "Look Inside" sample. Several warning signs together are more meaningful than any one alone.
Are AI-detection studies like the 77% figure reliable?
Treat them as estimates. The study behind that figure used an AI detector to label books as "likely" AI-written, and detectors give probabilities, not proof. The company that ran it also sells detection tools.
Is it illegal to publish an AI-generated biography of a real person?
I am not a lawyer, and the answer depends on the country and the content. Publishing false claims about a real person can raise legal issues. Paraphrase-heavy copies can also be hard to challenge under current copyright law, which is part of why authors say enforcement is difficult.
Is Amazon doing anything about this?
Yes. It has upload limits, AI disclosure for publishers, identity verification for some accounts, and it removes books that break its guidelines when they are flagged. What it has not done is show shoppers a public AI label.
Final Takeaway
AI-generated books on Amazon are not a hypothetical future problem. They have reached bestseller charts, copied real authors' launches, and put a journalist's name on a biography she never approved. At the same time, Amazon has added upload limits and ID checks, and the worst listings often vanish after enough readers complain.
For now, the strongest protection is a reader who checks a few things before clicking buy. Look at the author's catalog, read the sample, and trust the reviews that say what the book is really like.
Keep reading:
“How AI content addiction shows up in creators: AI content tools addiction signs.”
“Why sameness is hurting creators: AI-generated thumbnails all look alike.”
“The wider risks of AI on social platforms: hidden dangers of AI social media.”