Two books a week: Amazon pulls the plug on AI slop

Or: why the new KDP rules are the best news in a long time for honest authors — and the worst for people with a Lamborghini in their thumbnail.

Quietly and without fanfare: the revolution came without a press release

I’ll admit it: even I almost missed it. While we were busy at the Frankfurt Book Fair debating the future of reading, AI licensing deals and whether the coffee in the hall can really be called coffee, Amazon had already acted. Since September 21, 2026, a new limit has been in place at Kindle Direct Publishing: no more than two new titles per week per book format may be created. In early October, a second brake was applied — submissions for publication are now also capped at two titles per week per format. The counter resets every Sunday at midnight UTC.

The announcement was made roughly as loudly as an algorithm whispers: a brief post in the KDP community forum, a few lines on the help page. No fanfare, no banner, no webinar. For a company that normally celebrates every new Kindle color scheme like a moon landing, that’s remarkably understated. Perhaps they simply didn’t want to wake anyone up. Especially not the coaches.

A brief look back: I did say so (twice)

Anyone who read my two pieces on buchmarkt.de will know the diagnosis: Amazon KDP had — and still has — a serious slop problem. Coloring books featuring unicorns with six legs. Children’s books in which the story repeats itself on every page because the prompt demanded it. Self-help guides on topics the „author“ didn’t even know existed. And all of it published at a pace that would have made even Simenon envious.

The problem was never AI itself. The problem was the sheer volume. When someone can upload twenty „books“ in a single afternoon, the three years a novelist poured into her work simply drown in the noise. Visibility in the store is a finite resource — and it was being systematically harvested by those who had invested the least effort.

What has actually changed

For anyone who doesn’t follow every forum update by heart, here is a summary of the key rules:

Upload limit: A maximum of two new titles per week per format (e-book, paperback, hardcover) may be created — and a maximum of two per week may be submitted. Updates to already-published books do not count toward the limit.

AI disclosure remains mandatory: Anyone who uses AI to generate text, images or translations must declare this to Amazon — cover art and illustrations included. And, a nice touch: generated is generated, even if you tinker with it extensively afterward.

AI-assisted is fine: Authors who write themselves and use AI only for proofreading, refinement or brainstorming are not required to disclose this. That is a fair distinction — and precisely the one I have been advocating in my courses and consultancy work for some time.

Low-content, clearly labeled: Notebooks, planners, journals and logbooks must be marked as „low-content.“ Miss the checkbox and you’re out.

No clone armies: Content that barely differs from other books, is too short, repeats itself incessantly or is sold with misleading metadata is classified as a „poor customer experience“ — and will be rejected or removed. Fifty notebooks that differ only in cover color are not a product line. They are an imposition.

Hard times for the gold-rush coaches

And now to the real losers of this reform. You know who they are. You’ve seen their videos, whether you wanted to or not. Young men in front of rented sports cars, grinning broadly and explaining how „low-content books“ can generate passive income. Coloring books, activity books, hastily assembled children’s titles — upload once, collect forever. They’re happy to share the secret, of course. For 997 euros. Today only 497.

The miracle formula was never really substantiated. There were screenshots. There were dashboards showing large numbers. There were testimonials from people who, suspiciously often, happened to be selling a course themselves. What was rarely provided: verifiable figures on costs, returns, deleted titles and the many people who gave up in frustration after three months. The law of large numbers is, after all, usually the most seductive argument of all. I’ll write about that in more detail another time — for now, just this: anyone whose business model is built on volume has a mathematics problem.

Let’s do the arithmetic quickly: two paperbacks a week comes to around a hundred a year. For someone who actually writes, that is an absurdly high figure. For someone whose strategy is „flood the market with 500 variants of My First Dino Coloring Book,“ it is the end of the business model. I can already hear the new course titles: „From two books a week to a million.“ I look forward to seeing the math on that.

The mockery is warranted, but this is about more than a few bruised YouTube channels. It is about the people who believed these promises and sank their savings into courses and tools. All one can say to them is: the shortcut was never a shortcut. It was a dead end with a toll booth.

Why this is good for everyone who actually writes

I consider these rules more than reasonable. They prevent those who put genuine work into their books from simply being drowned out by AI slop. Self-publishing already has an image problem — in bookshops, in the literary press, sometimes even within one’s own family („Oh, you did it yourself?“). Every cloned how-to book with made-up sources has fed that prejudice.

By putting the brakes on volume, Amazon benefits:

Authors, whose new releases will no longer disappear among thousands of throwaway titles.

Readers, who can once again browse and find actual books rather than prompt outputs.

The self-publishing scene as a whole, which can finally define itself again by quality rather than upload speed.

And yes, Amazon itself — because a store full of disappointments is, in the long run, a bad store.

**That said, I’m not without reservations**

A little water must go into the wine. First: the AI disclosure is an internal declaration made to Amazon. The customer browsing the shop never sees it. Transparency for the corporation is nice; transparency for readers would be nicer.

Second: the limit catches the wrong people too. Small publishers with large backlists, authors who want to release a completed series all at once, translations into multiple languages — they all now have to plan in weekly instalments. That’s annoying, but manageable. And honestly: anyone publishing fourteen volumes at once should probably ask themselves whether they’re squandering their own launch impact anyway.

Third: rules are only as good as their enforcement. Anyone with enough accounts, pseudonyms and creativity will look for workarounds. Amazon will have to stay on top of this — through genuine scrutiny, not just counting.

**Conclusion: Less is more again**

Amazon reacted late, reacted quietly, but reacted correctly. Two books a week is enough for anyone who is genuinely writing. And it’s too few for anyone who is merely producing. That’s exactly where the line falls — and it falls in the right place.

To all the authors who pour their evenings, weekends and nerves into a manuscript: this one’s for you. To the coaches with the rental car: perhaps now is the moment to write a book of your own. A real one. It takes longer than an afternoon. But it’s worth it.

And for everyone now feeling nervous: check the current KDP help page before every publication. Amazon likes to change its rules — and communicates them, as we have learned, in little more than a whisper.

This English version was created with the help of AI (translated from the German original) and may differ slightly from a professional human translation.