Can AI replace children's book illustrators?
Summary
AI-generated picture books for children are getting better with time. Illustrators tell us how they feel about this development
In the late 1990s, when Sudeshna Shome Ghosh was working with Puffin, the children’s imprint of Penguin Books India (now Penguin Random House India), she had an opportunity to publish one of the most exciting books of her career. “The Puffin Book of Magical Indian Myths (2000) by Anita Nair was a big book for us—one of the first of its kind to retell Indian mythology for young readers," says the editor and writer. She commissioned Atanu Roy, one of India’s finest illustrators and cartoonists, to illustrate it. “His approach was ‘old school’, he took no shortcuts, and worked on each illustration for as long as he needed to," Shome Ghosh adds. “Anita would lose patience once in a while, but then, Atanu would send an image, and it would be mind-blowing."
The piece de resistance was the depiction of Lord Vishnu’s matsya (fish) avatar for the cover, which remains iconic to this day. “It was a different era," Shome Ghosh adds. “If you needed time to produce good work, you could afford to take it." Cut to 2025, and you can write a children’s book in a weekend and publish it. All you need to do is compose a prompt for an artificial intelligence (AI) tool, provide a skeleton of a plot, along with a few references for illustrations, and you will have your illustrated book ready in a few hours.
All in all, AI can do a passably good job— but more often than not, it is hit and miss. “Although AI-generated images often have a highly finished and rendered quality, I am yet to see an AI-enabled book that offers consistency of style and design throughout," writer and illustrator Pankaj Saikia says. “But considering the speed at which it is evolving, it won’t be long before AI is able to produce better quality books.
Also read: Art in children's books comes of age
GEN AI AS AUTHOR
In 2022, Ammaar Reshi, a product designer in Silicon Valley, was one of the first people to publicly put out the idea of using AI to write and illustrate children’s books. On 9 December last year, he posted a thread on X, starting with, “I spent the weekend playing with ChatGPT, Mid Journey and other AI tools… and by combining all of them, published a children’s book co-written and illustrated by AI!" The finished product was called Alice and Sparkle. It is a story of a girl, Alice, who creates her own AI, Sparkle, and together they embark on an adventure to make the world a better place. In a feel-good pitch, the blurb described Alice and Sparkle as a story that “hopes to inspire children, encourage their curiosity and learning, in one of the most technologically exciting moments in our lifetime."
Since then, the impact of AI on the livelihoods of artists and illustrators has become far more palpable. Writer, educa tor and illustrator Parismita Singh, best known for her graphic novel The Hotel at the End of the World (2009), remarks on the ease with which AI-generated images are being used in textbooks and educa tional content. “Some of my friends are using AI to create teaching materials," she says. “They send me their work from time to time to check if the art looks alright."
Apart from sabotaging the careers of professional illustrators, the AI invasion may diminish the trust of organisations and NGOs that are funding projects to make children’s books more diverse and accessible through translations and wider dissemination, Singh adds. That’s why not for-profits, especially, have to be vigilant. “As a not-for-profit organisation, we curate our books very carefully. We want to find humans who have a unique story to tell," says Canato Jimo, writer, illustrator and art director at Pratham Books. “I haven’t yet encountered AI-generated images in my field. There is a trust I share with the artists I work with."
At the same time, it’s not unwarranted for a textbook publisher to improve the bottom line by using AI-generated images, Canato admits. Why pay for the labour of human illustrators, when it is more expedient as well as cost-effective to get AI to do the job? Yet, it is not easy to take a black-and white view of the role of AI in children’s books. For Reshi, generative AI was a tool for personal innovation. Excited by the possibilities he had opened up, many adopted AI to breach the gatekeeping of mainstream publishing. Even a child aspiring to write a book could get on to an AI-enabled self-publishing platform like BriBooks to fulfil their dream. If you want to correct historical wrongs, you can use AI to create stories that don’t smack of gender biases, as illustrator and author Karrie Fransman and her partner Jonathan Plackett did in their 2020 project, Gender Swapped Fairy Tales. Such egalitarian uses notwithstanding, these trends forecast an uncertain future for professional illustrators.
Also read: Isn't it time that Snow White learnt stranger-danger?
A NEW VOCABULARY
Rather than giving in to alarm, Saikia, whose recent work includes illustrations for Shome Ghosh’s middle-grade novel, A Home to Haunt, takes a different stance. “As a professional for almost a decade, it feels like the time has come to re-evaluate my approach to illustration as a practice," he says. “I do not feel threatened by AI art. Rather, it feels like an important juncture in art history, similar to the rise of photography that led to newer movements in the visual arts."
Indeed, illustrated books for young readers have evolved significantly over the few last years, becoming sophisticated tools for cognitive, behavioural and social development rather than didactic instructional manuals. Think of the work of writers and illustrators like Oliver Jeffers, Julia Donaldson, Quentin Blake. Or closer home, of Prabha Mallya, Priya Kuriyan, Rajiv Eipe and Adrija Ghosh, to name a few.
Instead of lamenting the rise of AI, a more useful response may be to bring the human hand strongly into the creative process. Shome Ghosh gives the example of Tsering Namgyal Khortsa, whose novel, The Tibetan Suitcase (2024), she has published at Speaking Tiger. To provide a reference image to the book’s cover designer, Khortsa had asked AI to create the imaginative suitcase he had in mind. But the final result, beautifully illustrated by Mohit Suneja and designed by Maithili Doshi, not only superseded the AI version but also had nothing to do with any AI tool. Even for the smartest AI tools, it is still hard to rival the richness of observation and lived experiences that humans bring to the creative process.
That’s why, as AI continues to improve, it is key for artists to “get their hands dirty, rediscover the joy of creating illustrations by hand," says Canato, instead of relying heavily on digital tools like PhotoShop. In the days when Roy was illustrating Nair’s book, illustrators had to go out into the world to collect ref erences, be it of a leaf or flower. “With the coming of the internet, we became armchair artists, sitting in our rooms and drawing based on the visuals available on the internet," says Saikia. “In the process, we have become the mythical snake eating its own tail."
Also read: What artists’ childhoods can tell us
In the end, the survival of any form is linked to what its audience expects of it. “The crux of the matter is visual literacy," Singh says. “As writers and illustrators, our job isn’t to simplify narratives for readers. To be able to read a graphic novel, you’ll need to understand how visual lan guage works." With their subjective access to the joys, sadness and wonder of the world, artists can thus become the conduits of precious knowledge. “Recently, I was in an interior village in Arunachal Pradesh to research a project. ...I realised that no amount of pre-existing material on the internet could have told me the way I needed to engage with the community," Saikia says. “I had to be present there to understand the lives of the people, which, in turn, will inform the work I’ll go on to make. No AI tool I know can do the job of an artist going out into the world and feeling things."
