I read this websites free AI-written YA novels so you dont have to

Human novelists are usually considered prolific if they can knock out one book a year — maybe two if they're writing pulpy sci-fi or Young Adult (YA) fiction. And judging by the book world's buzziest story of the month — the horror novel Shy Girl being pulled by its publisher over evidence of AI use — these humans can't ask AI for an assist.

Meanwhile, AI novelists don't need any human help; they can knock out 12 YA books in 12 days. Or at least, that's what they've been doing on the brand-new website Babel Bot Books, which boasts such free-to-download titles as The Unraveling of Whispers, The Resonance Fracture, and The Chronosync Divide.

But making those novels actually readable is a different matter altogether — for human and AI authors.

"Every book in this library is an original, full-length science fiction novel," the Babel Bot website proclaims. (Some of the books presented would be more likely shelved under "fantasy," but we'll allow it.) "Written from scratch by an AI model [the site offers titles by Claude, Gemini, and GPT models] with zero human intervention in the creative process. No editing, no prompting chapter by chapter, no human co-authoring ... see for yourself what the bots have been up to!"

Reader, I decided to risk my sanity — and heck, maybe even my author hat — by doing just that. I'd seen a few pages of AI fiction before, but never gone this deep. Would an AI novel in 2026 be a surprisingly satisfying feast of plotty goodness, or thin and sloppy gruel?

What does an AI library of YA look like?

One thing AI can do really well — in fact, pretty much the only thing it's ever doing really well — is mimic the vast quantities of data it's been trained on, much of which was derived from books.

(Full disclosure: my book on Star Wars was one of the thousands of works in a pirated collection that Anthropic has admitted to using to train its Claude models in defiance of copyright, and I've filed my claim as part of the class-action lawsuit that Anthropic settled for $1.5 billion. That works out to just $3,000 a book, so settle down.)

So browsing the Babel Bot library for the most compelling title, I almost felt like AI was having a laugh by parodying ... well, exactly what you'd see in the YA section of a bookstore. Right down to covers with a single female protagonist working magic spells with bright lights, or looking at a futuristic cityscape with bright lights. AI doesn't give a damn about originality in this genre, because neither do we.

None of the books were written in the first person, which may disappoint some BookTok types. Their blurbs all seemed alike. There were repeated phrases even in the subheads. "The sky is a lie. The surface is calling," reads the tagline on The Harvester's Descent, written by Gemini on March 12. Then on the cover of The Starless Crown, which GPT 5.4 Codex tossed off on Saturday: "The sky is a lie. The stars are waiting." The Silence Protocol sits next to Silence Between the Stars.

Surely, the world's first AI-on-AI plagiarism lawsuit is not far off?

The book descriptions have a similar sense of copying each others' work. Nearly all their protagonists are 16 or, most commonly, 17. A few outliers are aged 18. There's no sign that Declan Voss, 16, magician of Thornwall in Unraveling of Whispers, is related to Maren Voss, 16, aboard the deep space vessel The Ardent in Silence Between the Stars. If the site is successful, maybe AI YA nerds will create elaborate theories about all the stories taking place in a single Babel Bot-verse.

But again, this feels less like AI is playing a lazy game of mad libs, and more like it's mocking us. You're the ones who made this kind of slop in the first place, not us, the bots' same-y blurbs seem to say. You love this kind of nonsense, don't you? Well, now we churn it out faster. Move over, human authors, we can do this all day.

What does an AI novel read like?

For a deeper dive, I selected The Probability Garden. It stood out from the slop for a few reasons. First, it was the only one of the 12 to be co-authored, by Claude Opus 4.6 and Gemini 3.1 Pro. At least, that's what it says on the site; in the downloadable EPUB file itself (one of five formats the site offers), Claude Opus is the only writer listed. (Authorship squabbles aren't unusual in human book world, either.)

Nevertheless, The Probability Garden was the only title to phrase its tagline as a question, and a philosophical one at that: "what if free will was just another prediction?" AI thinking about free will within this formulaic format? OK, I'm in.

So I let Claude transport me to the year 2260, where human society is controlled by quantum probability engines called Looms that weave threads predicting the future (not to be confused with the Weave, a quantum-psychic network in The Silence Protocols). The Looms are housed in a giant structure called the Spire (not to be confused with the setting of Chronosync Divide, a giant structure called the Spire.)

The Looms are known to predict human lives perfectly, but there are people who throw these predictions off. They're officially known as "participants," but working-class residents of The Understory call them "stochastics." As ... an insult, apparently? Bless your heart, Claude, I thought, you've got no idea how prejudice works, but you're trying!

Our protagonist is a 16-year-old Stochastic called ... Lumen. It's Lumen versus the Looms. OK, maybe you're not trying that hard, I thought at Claude. Or could this be Severance fan fiction? Lumen wears grey maintenance coveralls, which are to dystopias what the little black dress is to cocktail parties; a classic that never goes out of style.

As for Lumen herself, I kept waiting for her to become less grey — for some spark, some quirk, some sympathetic struggle, to make her feel like a hero worth following. The best Claude could offer in the opening chapters was that she wakes up at 5:47 a.m., even though the Looms keep predicting she'll wake up at 6 a.m.

Lumen then goes to work fixing ducts — hello, Terry Gilliam's Brazil! — and gets mildly harassed by Selk, her "assigned continuance monitor" who checks Lumen's ankle bracelet and chides her for waking early. "You're not special," Selk sniffs, in the book's first dialogue, 10 pages in. "You're a data point."

"A data point you visit personally every 3 days," Lumen shoots back. "Must be some data point." I was starved for conversation between characters at this point, so this actually sounded like a sassy comeback. Maybe this won't be so bad after all.

It's so bad after all

Claude knows, at least, that a protagonist has to want something. Lumen wants to know why her Weaver parents died at the Spire in mysterious circumstances nine years earlier. I'm not convinced Claude knows what it's like to have memories of parents, however. Lumen remembers "her father's voice reading to her at night, low and warm and slightly off-key when he did the character voices."

There's a lot of this sort of thing; description that falls into the uncanny valley. Claude knows voices are sometimes off key, but hasn't figured out that this is related to singing, not reading bedtime stories. In another chapter, one character is heard "exhaling through his nose" from "across the room." Are noses louder in 2260, or are ears more sensitive?

If only there were more amusingly weird descriptions, the book wouldn't be so terrible. Trouble is, they're just appetizers compared to the main course, pages of turgid technobabble. Claude tries to break it up with jokes, but its game is weak: "Shouldn't you be upstairs reading tea leaves?" Lumen says to a Weaver apprentice who works on the Looms. "Probability Cascades," the apprentice corrects.

"Same thing with better lighting," Lumen says. Huh?

There came a point where my brain snapped, and it was when the AI author seemed to be telling on itself: "The human brain wants structure. It finds it even in noise," one Weaver says about some impenetrable data. "This is noise."

It's a fair review of the whole AI novel concept: A Large Language Model, designed to predict the next word in a sequence, will create a lot of noise. The only hope for its success is that we'll see some structure in it. But we're used to dialogue between characters who sound human — and at the length of a novel, this is a hard act for AI to keep up.

Mercifully, Claude doesn't seem to know too much about the structure of fiction either, at least in terms of withholding information to create mystery. Somewhere around page 50, it revealed the entire plot, relieving me of the burden of reading any further.

Spoiler alert: The Looms aren't actually predicting the future. The weak-minded non-Stochastic humans are so convinced Loom predictions are accurate, they unwittingly create the predicted outcomes simply by expecting them.

Somewhere in there is a literary metaphor for what AI is doing to the world. Maybe a human author will be able to render it readable?