Publishing an AI Book: One Week, Zero Revenue

Publishing an AI Book: One Week, Zero Revenue

Publishing a book no longer requires a typewriter, a publisher, or winning a literary contest. All it takes is an idea, an artificial intelligence program, and an afternoon to upload the file to Amazon's platform. The result: a 71-page novel written in one week, signed with an Italian-sounding pseudonym and generating zero euros in revenue to date. The author admits frankly that it was an "express novel," poorly developed, and the synopsis was so vague that readers who don't check the series description have no idea what the book is about.

This case encapsulates a broader debate about the true cost of publishing. While the technical barrier has fallen, the commercial barrier remains intact. The discussion over whether this democratizes culture or floods the market with disposable products is now fully open.

The Process: Own Idea, External Development, Home Editing

The method described by the author is artisanal in concept but automated in execution. She provides the general idea, the language model develops the story, and she then tweaks parts she finds unconvincing because, as she explains, AI "changes the thread significantly" and cannot be left alone. The manuscript is edited in a word processor and uploaded to Amazon's digital publishing platform, which allows launching a standalone book or the first episode of a series.

The economic cost is virtually nil. The opportunity cost is debatable. The author acknowledges she could have written a thousand pages instead of seventy-one, but she wanted to finish it, and she did. The cover, heavily criticized in the conversation, also bypassed professional designers.

How Much Can You Earn Publishing on Amazon?

For now, nothing. The author confirms it herself: "absolutely zero." Her plan is to republish when the series grows and to put more work into subsequent episodes, but without going bankrupt in the process. This is the fine print of self-publishing: the platform doesn't charge for uploading files, but it doesn't guarantee readers either.

Comparison with traditional publishing is inevitable. In the conventional circuit, according to experiences shared in the thread, authors often pay for editing only to end up with unsold copies once they've exhausted their circle of friends, family, and local fairs. Traditional publishers still take their cut. The difference is that now the intermediary is an algorithm, and the storefront is a digital shop with millions of titles competing for the same attention.

Text Quality: Pronoun Errors, English Names, Overloaded Style

The excerpt published in the conversation received specific criticism. A reader detected incorrect pronoun usage (using "lo" where "le" is required) and English character names within a Spanish text. The author justifies this by stating the characters are foreign and the story crosses borders. Regarding the pronoun errors, she concedes that it "might be possible."

The overloaded style also sparked division. Some argue that commercial literature doesn't require having read all the classics, while others maintain that the result shows when the text comes from a model trained on millions of pages without editorial criteria behind it. The hastily written synopsis doesn't help: it describes a series about a killer, his tortured mind, and potential redemption, with a tone promising more than seventy-one pages can deliver.

Pseudonyms and Author Identity

Signing with an invented name isn't new. The author defends it with a commercial argument: "Paquita García Heredia doesn't have the same hook." The chosen pseudonym, with Italian resonance, fueled speculation about references to journalists or TV personalities. The discussion shifted toward the right to anonymity and whether a foreign name sells better than a Spanish one in the digital catalog.

The answer is empirical and disappointing: with zero sales, the pseudonym hasn't proven anything yet.

Is Self-Publishing with AI Worth It?

It depends on what you're looking for. As a business, the available numbers don't invite optimism: one week of work, a 71-page book, and no revenue. As an experiment, the cost is low and the learning is tangible. As a literary strategy, the consensus is that professional editing, style correction, and a commercially appealing synopsis are necessary.

The market itself offers cheaper alternatives to traditional publishing: blogs, publishing platforms, and reader communities. AI lowers production costs, not promotion costs. And promotion, in a catalog of millions of titles, remains the bottleneck.

The prediction, with caveats: we will see more books like this, faster and cheaper. Whether this improves the available offering or simply bloats it is a question the market will answer without hurry and without mercy.