A few days ago I read a thread in a Facebook group for people self-publishing on Amazon, and I’ve been angry about it ever since.
A mother posted a question. Years ago she started making up a story to help her little girl fall asleep, and she told it so many nights that she finally wrote it down. She wanted to publish it. Illustrators were more than she could afford, so she’d made the pictures in ChatGPT, partly from her own family photos. She asked three things. Did she have to tell anyone the images came from AI? Did she need a disclaimer? Would she run into copyright problems?
Those are good questions. They’re the questions a careful, honest person asks before she does something she’s never done. She was trying to get it right.
What she got was a war.
Asking for directions in a crowded station
A woman stands in a busy train station with a ticket in her hand. She turns to the people near her and asks which platform she needs.
Nobody points. Instead, one man tells her trains are a disgrace and she ought to drive. A woman behind him says he’s an idiot and trains are fine. Somebody else says anyone who can’t read a departure board has no business traveling. Within a few minutes the whole crowd is shouting at each other about transportation policy, a couple of them are handing out business cards, and the woman with the ticket is standing in the middle of it, still not knowing where to go.
That’s what that thread was. She asked for directions, and a room full of strangers used her as the excuse to have a fight they’d been itching to have anyway.
What happened in the thread
The first reply told her AI art would make her book look trashy. From there it went downhill fast.
One person called the idea gross. Another told her that claiming to write a book and then using AI was embarrassing. Several told her AI art is theft, and one of them told the next commenter that if they were fine being a thief, that was on them. Somebody looked at her post, typed the way a tired mom types on her phone, and told her that if she wrote it herself she couldn’t write a book, because her writing was atrocious.
Then the commenters turned on each other. A defender of AI got called a moron. Somebody else got called a dumbass for defending the use of AI “for whatever reason.” Two people argued for a dozen replies about who had worse reading comprehension. A few folks showed up to sell her things: a publishing course, a formatting service, a page rate for illustration. One helpful soul suggested she put “Illustrated by ChatGPT” on her cover.
Tucked into all that, a handful of people gave her kind, useful answers. Some of them had to argue with the others to be heard.
And then, about halfway down, she explained why she wasn’t willing to wait years to save up for an illustrator. She has serious health problems. She doesn’t know how much time she has. She wants her daughter to have something on the shelf she can point to and say her mom made it.
Some of the people in that thread read that and softened. Some of them kept going.
Do you have to disclose AI-generated images on Amazon KDP?
Yes. This one isn’t a matter of opinion, and the thread got it wrong as often as it got it right.
When you set up a book on Kindle Direct Publishing, Amazon asks whether you used AI tools to create the text, the images or the translation. Its content guidelines draw a line between AI-generated content and AI-assisted content. If a tool produced the content, whether that’s words, a cover or interior artwork, KDP wants you to say so. If you created the work yourself and used a tool to edit, brainstorm or polish it, you don’t have to disclose that.
Illustrations made in ChatGPT are AI-generated. She has to check the box. Somebody in the thread insisted, with total confidence, that KDP only cares about the text. That’s false, and it’s the kind of wrong answer that gets a book pulled later. Another person told her Amazon never checks, so it doesn’t really matter. It matters. Disclosing costs her nothing, and hiding it risks everything she’s trying to build.
She doesn’t need a disclaimer inside the book, and she doesn’t need to put the name of a chatbot on her cover. The disclosure goes to Amazon, in the setup form, where it belongs. I’ve written more about what authors owe readers when AI is involved, and the short version is to tell the truth where it’s asked for.
Can you copyright AI-generated illustrations in a children’s book?
In the United States, not the parts a machine generated. The Copyright Office has been clear that it registers human authorship, and images produced by a prompt don’t qualify on their own. Her story is hers, and it’s protected the moment she wrote it down. The pictures a machine drew for it are a different matter.
Her case has a wrinkle, because she started from her own photographs and shaped the results with a lot of specific direction. The Copyright Office has said that human selection, arrangement and real modification can count. Prompts alone don’t. Where her pictures land on that line is a question for a copyright attorney, and I’m not one. I’ve written about AI and copyright for authors if she wants the longer version.
The thread had bad advice here too. One person told her that if she merged four stock photos into a cover, she could copyright the image as her own work. That’s not how stock licenses or copyright work, and following that advice could get her into far worse trouble than the AI pictures ever would.
For a book whose whole purpose is to sit on her daughter’s shelf, the practical answer is simple. The story is protected. The pictures probably aren’t. Somebody could, in theory, copy those images, and for a bedtime story written out of love, that risk is small.
The people angry about AI illustration have a point
I use AI every day as a digital assistant in my own writing, and I’ll still say it plainly: the people who were upset aren’t wrong about the substance.
Illustration is a craft that takes years to learn. The image tools were trained on the work of artists who never agreed to it and never got paid. When anybody can generate a picture book in an afternoon, the people who draw for a living get squeezed, and a lot of AI-illustrated children’s books really do look alike. An illustrator commenting in that thread asked her, politely, to hire an illustrator if the book mattered to her. That’s a fair thing to say. Another suggested she and her daughter try drawing the pictures together. That was the best idea anyone posted all day.
If she asked me, I’d tell her the same things. An illustrator would give her daughter something no tool can, and there are good ones who work for less than people expect. She could publish now and replace the art later if money frees up. Or she could print a few private copies for family and skip the store entirely.
Every one of those points can be made in a sentence, to a person, with some kindness. None of them required telling a sick mother she’s a thief.
Why do questions about AI turn into online pile-ons?
I wrote about this pattern in Dogpile, after watching two authors get stoned in writing groups for asking ordinary questions, and again after a man was torn apart for posting AI art to a Renaissance faire group. This thread had the same shape, with one difference that makes it worse. She didn’t post anything. She asked.
AI is the fight of the moment, and people come to it loaded. Some of them have watched their income shrink. Some of them are sick of being told they’re behind the times. When a post mentions AI, a lot of readers stop seeing the person who wrote it and start seeing a side. The question stops being hers. It becomes a flag in the ground, and everybody runs to plant theirs next to it or rip it out.
From there it’s performance. The sharpest insult gets the most reactions, so people reach for sharper ones. The original poster fades into the background, and the commenters start fighting each other, because the other side of the argument is who they came to beat all along.
I know what the bottom of one of these feels like. I’ve been gang-attacked on Facebook and buried on YouTube, and I’m a grown man who writes for a living. Now put yourself in her place. She’s not a public figure or a provocateur. She’s a mom with a story, a health problem and a question, and she walked into a room she thought was full of people who’d been where she is.
The person who needed help gets the least of it. She came in with three questions and left with maybe two answers she could trust, a lot of shame, and a clear lesson that asking in public is dangerous. Next time she might not ask. She might just guess, and guess wrong, and that’s worse for everyone, including the illustrators the angry commenters say they’re defending.
What she needed to hear
The whole answer fits in a paragraph. Yes, check the AI box when KDP asks, because those images are AI-generated. No, you don’t need a disclaimer in the book or a chatbot’s name on the cover. Your story is protected by copyright and the generated pictures probably aren’t, so talk to an attorney if that worries you. And if the book is a gift to your daughter first, it’s allowed to be exactly that.
Then, if somebody felt strongly, they could add one more sentence: “If you can ever swing it, a real illustrator would make it even more special, and here’s someone affordable.” That sentence helps her. It might even help an illustrator get hired.
That’s the difference between a correction and a pile-on. A correction points at the problem and leaves the person standing. A pile-on points at the person.
If you’re writing a book you care about and you’re not sure where AI fits in it, start with the AI & Writing Hub, where I’ve laid out what these tools do well and where they break. If you want a human to help you get the book itself right, that’s what book coaching is for. And the next time you see someone standing in a crowded station with a ticket in her hand, asking which way to go, do the decent thing. Point to the platform.
