Photo by

Agnar Freyr Stefánsson

The case against using generative AI as a substitute for illustrators

If you entered a children’s section of a bookshop this holiday season, it probably didn’t take long before you saw a book utilising generative artificial intelligence (AI) for its illustrations. Whether you noticed a too-smooth cartoonish animal, a shadowless landscape, or a character’s features changing from page to page without reason, some books in this year’s Jólabókaflóð class reinforced a trend we’ve been noticing: AI-generated illustration is becoming a familiar tool for some publishing houses.  

Here, we want to explore what results from generative AI being used to illustrate children’s books. We won’t go into AI’s obvious copyright infringement, nor the staggering environmental impact AI has and will have on the world; we just want to ask: what is gained when AI-generated illustrations are used, and what is lost? 

It’s useful to understand where the decision to use generative AI as a tool in children’s books usually gets made. In a piece summarising the July 2025 issue of Swedish magazine Bazar Masarin — whose cover, coincidentally, is adorned with a beautiful illustration by celebrated Icelandic children’s book illustrator Lóa Hlín Hjálmtýsdóttir — Icelandic writer Arndís Þórarinsdóttir states that Bazar Masarin concludes from extensive interviewing about AI usage in children’s books that “publishers are more positive than artists.” Why are publishers more positive about generative AI in children’s books? It’s easy to take some guesses. 

“What is gained when AI-generated illustrations are used, and what is lost?”

The Icelandic publishing house most vocal about their use of generative AI in illustration and writing is Mosfellsbær-based Óðinsauga, so let’s use that as an example. Recently, the publishing house justified its usage of generative AI by sharing: “AI was also used when 300+ images were needed in one book, that many images are impossible to draw for such a small children’s book market.” Here, two justifications for using generative AI ring through: AI-generated illustrations are fast (“300+ images”) and cheap (“small children’s book market”).  

Although hearing any company’s hunger for quickly made and inexpensive art can make an artist or anti-capitalist seethe, it is hard to imagine a money-minded entity choosing the more time-consuming and expensive venture.  

If this trend continues and more publishing houses substitute illustrators for AI, then a decision falls on the consumer. We propose that you choose a human-illustrated book over an AI-illustrated one.   

Why? Why should you look through a spread of children’s books and choose one with imperfect lines instead of one with smooth, fantastical landscapes? 

It’s important to remember exactly what we’re talking about here: children’s books. We want to ask: what does it mean to hand children AI-generated illustrations? 

In a 2024 study at The University of Zagreb and The American International School of Zagreb in Croatia, children were presented with an AI-generated book and a human-created book side by side. Not only did the children significantly prefer the human-created book overall, but the young readers connected more with the human-created illustrations because of a text-picture relationship that was wholly lacking in the AI-generated book.

When AI-generated illustration is used in children’s books, a vital link is lost in the creative process. Prompting generative AI for an illustration only gives you a flat portrait of exactly what you typed, a formulaic image that inherently lacks uniqueness, creativity, and depth. A human illustrator produces original and emotional work.   

Further, children’s books are teaching tools, full of sentences and images. As the story teaches literacy, the illustrations teach creativity. Children’s book illustrations are crucial introductions to art. Giving a child an AI-illustrated book hands them something they could never create on their own.  

An artist’s illustration is always their first-hand interpretation, not second-hand information condensed into a result. Only a human illustrator knows the feeling of joy that comes from playing as a child, and only a human illustrator’s work is drawn from their lived experience.  

Children’s books often hold stories that teach complex lessons about life. Illustrations add to these stories, with the unique style and interpretation of the illustrator. When using generative AI — which prioritises the most popular style from the data it is fed — then all hope for nuance, playfulness, or poignancy is lost.

“Only a human illustrator’s work is drawn from their lived experience.”

Human illustrators can use their instincts and discretion while deciding how to depict a story in new and original ways. Meticulously drawing a forest full of monsters can immerse a child as if they’re alongside the main character, while just drawing an incomplete circle and a sliver can be enough to tell a story of self-acceptance.   

Children deserve work with thought and effort behind it, work that is nuanced, playful, and poignant. And works like this do exist in Iceland. The aforementioned Lóa Hlín won the Reykjavík Children’s Book Award in 2025 for her playful book Mamma sandkaka, which she wrote and illustrated herself; Rán Flygenring’s recently published Blaka is being praised for her inventive painting style that uses neon and black ink; Brian Pilkington’s vast catalogue of work continues to inspire and delight generations of children. 

As books illustrated by artists and books illustrated with generative AI sit on the same table at the bookstore, the decision to choose one over the other has to be made by the consumer. Choosing a book not illustrated by an artist pushes us further into a future with more books illustrated by generative AI. Will children not follow their uniquely creative instincts because their work looks nothing like AI-generated illustrations? Are we choosing not to foster the next generation of children’s book illustrators? We hope not. 

The next time you set out to buy a new children’s book and spot one illustrated with generative AI, ask yourself the question: Who is gaining from this, and who is losing something? 

Agnar Freyr Stefánsson is a graphic designer, and Ish Sveinsson Houle is a journalist at the Reykjavík Grapevine. The two also make the crossword for the Grapevine.