For a moment I entertained the idea that these are intentionally bad to get people to buy them as gag gifts. My kids and I certainly had a good laugh looking at the pictures in blog. That second picture of the jaw sticking out had my son ROFL-ing.
This just seems like laziness vs AI = bad. It's not like publishers are putting out masterpieces with human writing. They're cranking out minimum viable content as well.
I've found that by putting meaningful effort into AI storytelling, I can create bespoke stories that my kids love night after night.
My workflow is below: Caveat that it costs about $0.25-$0.50 to weave a book like this with Claude Sonnet and Gemini Nano Banana Pro. But to me the cost is worth it for the quality.
- Use Claude structured output and ask for page1, page2 ... pageN instead of an array of pages or wall of text.
- Pass a story arc as a set of values to the prompt. I.e. say each page has an emotional beat between 0.0 and 1.0. For a "man in hole" type of story: page1 starts at 0.6, page2 = 0.5, page5 = 0.25, page10 = 0.85. This ensures page 5 lands the "crisis" and page10 resolves higher than the start.
- For illustrations, have Claude generate the story text and an illustration prompt per page. i.e. page1: { "text": "...", "illustration": "..." }.
- For art consistency, add an "Art Direction" key to the structured output. Feed this into Gemini/OpenAI and ask for an art board visual guide & character reference sheet.
- Feed the page text, illustration prompt, and the art board to Gemini/ChatGPT images. I'm constantly surprised at the quality of the output.
Here's an example set of pages from a magic school bus style story about the immune system
I skipped some of the native romanian books we read since there are no known translations that I am aware of. One of the main reasons I want to teach my toddler english is so that we can appreciate a wider selection of books, because there are many books not translated in Romanian.
I also noticed that the quality of the translation matter immensely, probably more than for normal books. And a lot of books just don't translate all that well because they rely on rimes or alliterations.
The Complete Calvin and Hobbes anthology by Bill Watterson - a formative part of every kid's childhood growing up in our house.
Dad: How's your math coming?
Calvin: I don't do math anymore. I decided I'm more of a "visual" person.
Dad: Good. Visualize being the only 45-year old in first grade.
From the time of its publication in 1947 until 1972, the book was "banned" by the New York Public Library due to the then-head children's librarian Anne Carroll Moore's hatred of the book.
According to children's literature expert Betsy Bird, Moore criticized Goodnight Moon due to the fact that she believed it lacked a meaningful narrative structure and educational value.
)
"Good night little house and good night mouse, good night comb and good night brush, good night nobody, good night mush"
My kids loved the mush part. I still remember it more than a decade after I last read the book for the bajillionth time, often more than once per night, to a kid that wouldn't go to sleep.
I've been wondering what the long term result on peoples perceptions of reality will be after all this AI slop. I've noticed a lot of the times I can spot AI slop videos because they just don't match what I know to be true, I can think "That's an AI video of a fox because I know foxes don't act/move like that" But then the only reason I know that is because I've seen hundreds of videos on the internet before AI generated video was a thing. But someone who grew up seeing AI slop from the start doesn't have that firm grasp on reality to spot fake content from.
One of the scariest things to me about getting older is that there's entire generations younger than me who don't realize a growing number of things we used to have and lost. Because, like you said, they grew up without it.
And they don't believe things even can be better because they regularly hear one of the dumbest ideas of our time: that the past wasn't actually better, we only remember it that way.
the lack of effort has been the main thing for me since this all started.
you give people a tool to do something easier and instead of doing more WITH the tool they do this instead.
is anyone out there using AI to make more higher quality children's books than were possible before?
I don't think there has ever been an appetite amongst corporations to improve the quality of their products if they can easily get away with reducing costs.
I think its always been a thing. Give a society any new technology and the distribution curve of human effort doesn’t disappear: a slice of people will aim it at entertainment, shortcuts, and the lowest common denominator (this book), and a smaller slice with high discipline and curiousity will use the exact same tools to become 10x more capable. The tech changes; the distribution of how people use it mostly doesn’t
The lack of effort is the point. This is the equivalent of SEO spam, meant to extract money from the ecosystem by crowding out books that earned their position in the marketplace.
The difference is that multimodal generative AI means you can crank you SEO spam-style content across all media, at any scale.
As the other commenter here says, I suspect whoever did skim through found it amusing. Besides, it's not like this genre was particularly accurate anyway --- I have some old purely-manual-human examples of "how things work" books which distort, exaggerate, or grossly simplify for illustrative purposes.
The lack of effort is the point. The intent is to automate the entire pipeline and churn out huge numbers of these for whatever the top selling topic of the week is.
Because openai and other ai companies spend billions convincing people that they dont need to put in effort as long as they use AI. They literally think they are interacting with a hyperintelligence that is so smart it will destroy the planet eventually. Why would you spellcheck a digital god? Why not just push straight to publish and "automate" everything.
I assume these aren't being published by major publishing houses but rather microbrands and print-on-demand services. They're, like, bypassing gatekeepers and democratizing knowledge, man. Why do you hate freedom so much?
There was a window where new authors could break in with blogging and self publishing. Andy Weir (The Martian, Project Hail Mary, and the bad one that shall not be named) got started this way I think.
That window is now closed. If I wanted to be an author I’d probably try to get a real publisher, with all the downsides that entails.
These are not made by people who care. It’s a scam basically. Spamming Amazon with slop is a current hustle culture thing. There’s guides, probably AI generated and not proofread, explaining how to do it. They obviously have tricks to game the rankings since these books get recommended like mad in every category.
It’s today’s hot successor to the big drop shipping craze, which is also still happening, and has destroyed Etsy. That was another hustle culture thing. I remember hearing something about it being one of the get rich scams Andrew Tate was teaching at his thing.
You could use AI to help make a good book like this, but you would proofread and fact check it and sit there and converse with the AI and tell it all the stuff to fix… just like vibe coding.
I understand you can use AI to make a good book but you can also make a good book without AI. Why does AI have to be involved at all? Were we running out of children books that we need to optimize a factory assembly line for them with AI?
It's like there are some things that do not even need AI and thats okay. Children's books also don't need a hurculean effort to write/create (the part ai tries to automate and fs up). In fact, its almost entirely about the concept and direct execution.
You mention vibe coding but this is fundamentally different and it doesnt apply
Just like Vibe coding, I think this article is trying to convey that, when children’s books important not to rely solely on AI, but to consciously strive to create high-quality works.
The concept of consuming AI generated content for children has always baffled me.
We collectively have a virtually infinite collection of already existing hand crafted quality content filtered over the years in the form of children stories and tales that we can pick and chose from to read to our children. We love telling stories especially to our children.
Why would ANYONE be enticed by the idea of using AI to generate tales when there are so many out there to tap from is really beyond my comprehension.
While I agree with your main point, this isn't exactly true.
The quality content in children's media does NOT survive through the ages. There are so many other incentives in children's publishing that quality for children is but one signal among many. Like how a parent will buy a book that teaches a 'good lesson' as a proxy for a good book, which is harder to determine.
On top of that, there are systems at play that limit the impact of curators who really put the work in to identify good children's books. For example, a children's librarian has to buy books through the city or county procurement process. Only certain vendors will have registered as a valid supplier to the procurement team, and then they have a chokehold on what can be bought for the library, so they can offer their shovelware with larger margin, along with a few compromises about the inclusion of known-good books.
And then to add to this, the rights to publish good books are more expensive, and require more work and negotiation.
Any parents who want an example of this should check out the works of Tomi Ungerer. Really some of the best picture books ever made, and often not available to be purchased at all. Phaidon, a niche and fancy publisher finally secured some rights, and is releasing some nice editions, but you won't find them in most public libraries. And even then, some of the his best work isn't available due to complications (like The Hat, only available in anthology or used books from the 70s)
This is so apparent as a parent that loves to read. It feels like things are even worse than Sturgeon's law would make you think.
Sturgeon’s law is absolutely true. Just look at the religious ideas and worldviews children have been taught for thousands of years, and at the hatred, wars, massacres, and slavery religion has brought into the world. Modern dictatorships like North Korea and China also subject children to carefully engineered indoctrination. I think modern AI, and future AI guided by humans, can do much better than Sturgeon’s law would suggest.
> We love telling stories especially to our children.
And, for a large number of parents, "we" love sitting our children down in front a screen and letting it be their primary source of entertainment before they can even utter one word.
I'd bet that the majority of parents feeding their children AI slop don't even know it's AI slop because they couldn't identify it as such...even if they even cared to, which most of them don't.
You could say that about pretty much any form of media, people just like new stuff more than old stuff. There's more 9 and 10/10 movies than most people would watch in their lifetime already but people will go see some forgettable trash movie in the theater instead.
The category of thing is not new. The level of industrialization and quantity is new.
Slop on social media also predates AI but at least back then someone had to make it… usually people in poorer countries using it to game algorithms for monetization.
>Last week, I posted a visual demonstration of the sameness of AI-generated content. This makes the output easy to spot even if all the individual pieces are perfect facsimiles of what a human could create:
I think a point missing is that this output all looks the same because the prompters are not specifying much more than the barest minimum to get what they want. If you just prompt "generate a cover for my book 100,000 whys which is childrens book that answers their questions about science" then you get images like from TFA using the models default style. However, the models are capable of reproducing any great artists style and any content you want.
If you have seen the prompts for images on communities of enthusiasts you may notice that they can be quite long and specify considerable detail about both the content and the style of the output.
Here is one of the four above the fold on the front page of CivitAI for me right now, it has both a positive and negative prompt. Not that long because this is a fairly simple image. However the image doesn't look like the slop in the 100,000 Why's book covers or the many commercial signs and advertisements I'm seeing when I leave the house.
The only way this ends well is that we end up with extreme regulations in all places to enforce ownership, from software (because of insecure code due to excessive vibe coding) to children's books (because we end up teaching them crap).
Totalitarian as hell, but I don't see any other way.
It's like surreal absurdist art.
I've found that by putting meaningful effort into AI storytelling, I can create bespoke stories that my kids love night after night.
My workflow is below: Caveat that it costs about $0.25-$0.50 to weave a book like this with Claude Sonnet and Gemini Nano Banana Pro. But to me the cost is worth it for the quality.
- Use Claude structured output and ask for page1, page2 ... pageN instead of an array of pages or wall of text.
- Pass a story arc as a set of values to the prompt. I.e. say each page has an emotional beat between 0.0 and 1.0. For a "man in hole" type of story: page1 starts at 0.6, page2 = 0.5, page5 = 0.25, page10 = 0.85. This ensures page 5 lands the "crisis" and page10 resolves higher than the start.
- For illustrations, have Claude generate the story text and an illustration prompt per page. i.e. page1: { "text": "...", "illustration": "..." }.
- For art consistency, add an "Art Direction" key to the structured output. Feed this into Gemini/OpenAI and ask for an art board visual guide & character reference sheet.
- Feed the page text, illustration prompt, and the art board to Gemini/ChatGPT images. I'm constantly surprised at the quality of the output.
Here's an example set of pages from a magic school bus style story about the immune system
[image] https://media.discordapp.net/attachments/839188039229112353/...
I'll start: John Rocco, How We Got to the Moon. (http://www.howwegottothemoon.com/)
"I won't give up my rubber band" is a sweet, imaginative, thoughtful exploration of the thing we get attached to in our lives. https://www.goodreads.com/book/show/58739625-i-won-t-give-up...
"I wonder where I am" is a exploration of maps in various forms, a bit over my 2.5 year old's cognitive abilities, but I think it's great. Can't wait until he's able to get it.: https://www.goodreads.com/book/show/204810909-i-wonder-where...
The books from Julia Donaldsson are classics. I am partial to Gozzle:
Gozzle https://www.amazon.co.uk/Gozzle-Julia-Donaldson/dp/152907641...
The snail and the whale: https://www.amazon.com/Snail-Whale-Julia-Donaldson/dp/150983...
My kid had a loooong "The hospital dog" phase: https://www.amazon.co.uk/Hospital-Dog-Julia-Donaldson/dp/150...
I skipped some of the native romanian books we read since there are no known translations that I am aware of. One of the main reasons I want to teach my toddler english is so that we can appreciate a wider selection of books, because there are many books not translated in Romanian.
I also noticed that the quality of the translation matter immensely, probably more than for normal books. And a lot of books just don't translate all that well because they rely on rimes or alliterations.
So mainstream, it has a Wikipedia page: https://en.wikipedia.org/wiki/The_Way_Things_Work
( with bonus book ban credentials:
"Good night little house and good night mouse, good night comb and good night brush, good night nobody, good night mush"
My kids loved the mush part. I still remember it more than a decade after I last read the book for the bajillionth time, often more than once per night, to a kid that wouldn't go to sleep.
And they don't believe things even can be better because they regularly hear one of the dumbest ideas of our time: that the past wasn't actually better, we only remember it that way.
The difference is that multimodal generative AI means you can crank you SEO spam-style content across all media, at any scale.
Over time, i hope the chickens come to roost.
That window is now closed. If I wanted to be an author I’d probably try to get a real publisher, with all the downsides that entails.
It’s today’s hot successor to the big drop shipping craze, which is also still happening, and has destroyed Etsy. That was another hustle culture thing. I remember hearing something about it being one of the get rich scams Andrew Tate was teaching at his thing.
You could use AI to help make a good book like this, but you would proofread and fact check it and sit there and converse with the AI and tell it all the stuff to fix… just like vibe coding.
It's like there are some things that do not even need AI and thats okay. Children's books also don't need a hurculean effort to write/create (the part ai tries to automate and fs up). In fact, its almost entirely about the concept and direct execution.
You mention vibe coding but this is fundamentally different and it doesnt apply
Afaik, parents are super protective of their children and would never do something that could inhibit a childs learning
We collectively have a virtually infinite collection of already existing hand crafted quality content filtered over the years in the form of children stories and tales that we can pick and chose from to read to our children. We love telling stories especially to our children.
Why would ANYONE be enticed by the idea of using AI to generate tales when there are so many out there to tap from is really beyond my comprehension.
The quality content in children's media does NOT survive through the ages. There are so many other incentives in children's publishing that quality for children is but one signal among many. Like how a parent will buy a book that teaches a 'good lesson' as a proxy for a good book, which is harder to determine.
On top of that, there are systems at play that limit the impact of curators who really put the work in to identify good children's books. For example, a children's librarian has to buy books through the city or county procurement process. Only certain vendors will have registered as a valid supplier to the procurement team, and then they have a chokehold on what can be bought for the library, so they can offer their shovelware with larger margin, along with a few compromises about the inclusion of known-good books.
And then to add to this, the rights to publish good books are more expensive, and require more work and negotiation.
Any parents who want an example of this should check out the works of Tomi Ungerer. Really some of the best picture books ever made, and often not available to be purchased at all. Phaidon, a niche and fancy publisher finally secured some rights, and is releasing some nice editions, but you won't find them in most public libraries. And even then, some of the his best work isn't available due to complications (like The Hat, only available in anthology or used books from the 70s)
This is so apparent as a parent that loves to read. It feels like things are even worse than Sturgeon's law would make you think.
Anyway, check out the caterpillar for the fifty-seventh time.
Yeah, it really does.
There is a mountain of human care and creativity to draw from; and nothing wrong with adding to the mountain.
But why bother with the statistical simulacra of the mountain (or raise your children on it).
I think the history of children's literature may be shorter than you think.
And, for a large number of parents, "we" love sitting our children down in front a screen and letting it be their primary source of entertainment before they can even utter one word.
I'd bet that the majority of parents feeding their children AI slop don't even know it's AI slop because they couldn't identify it as such...even if they even cared to, which most of them don't.
AI slop is just a more complete reimplementation of the "shovelware" from the 90s
https://en.wikipedia.org/wiki/Shovelware
https://en.wikipedia.org/wiki/One_Hundred_Thousand_Whys
Slop on social media also predates AI but at least back then someone had to make it… usually people in poorer countries using it to game algorithms for monetization.
I think a point missing is that this output all looks the same because the prompters are not specifying much more than the barest minimum to get what they want. If you just prompt "generate a cover for my book 100,000 whys which is childrens book that answers their questions about science" then you get images like from TFA using the models default style. However, the models are capable of reproducing any great artists style and any content you want.
If you have seen the prompts for images on communities of enthusiasts you may notice that they can be quite long and specify considerable detail about both the content and the style of the output.
Here is one of the four above the fold on the front page of CivitAI for me right now, it has both a positive and negative prompt. Not that long because this is a fairly simple image. However the image doesn't look like the slop in the 100,000 Why's book covers or the many commercial signs and advertisements I'm seeing when I leave the house.
https://civitai.com/images/134444826
Totalitarian as hell, but I don't see any other way.