Overcoming "I Can't Draw": The Story of a 42-Year-Old AI Picture Book Author
機械翻訳 / Machine-translated
機械翻訳 / Machine-translated
11 PM. Saeki's fingers trembled as she stared at her smartphone screen.
Just one centimeter left before she pressed "Publish." A dream she had given up on for 42 years was about to become reality in just two months. The screen displayed the title of the picture book she had created with ChatGPT and Midjourney. Fuwari and the Sky's Gift.
How did a woman who had never drawn a picture in her life make it this far?
Saeki is a company employee working at an IT firm in Tokyo. Fifteen years in a planning role. Consumed by daily tasks, she comes home to help her second-grade daughter with homework. Her husband is away on a work transfer, returning home only on weekends.
In the middle of that everyday life, she had forgotten what she wrote in her graduation essay in sixth grade: "Future dream: picture book author."
"I was serious back then. I borrowed picture books from the library every night, read them, and wrote stories in my notebooks. But I couldn't draw."
Her art grades in middle school were always a "3." In high school she didn't take art, and she majored in economics in college. The dream of becoming a "picture book author" was quietly tucked away in a corner of her heart as "a childhood fantasy."
The turning point came without warning. Saturday morning, January 18, 2026. Her daughter looked up from a YouTube video on her tablet and said:
"Mama, this person made a picture book with AI. They didn't draw anything."
The screen showed a housewife's account of self-publishing a picture book on Kindle using AI — ChatGPT for the story and Midjourney for the illustrations, completed in three weeks.
Something stirred inside Saeki. Maybe I can do this too.
That night, after putting her daughter to bed, she searched on her phone: "how to make an AI picture book." What came up was a 2026 guide to creating AI picture books. All she needed was ChatGPT Plus (¥20/month) and Midjourney (¥10/month). Initial investment: around ¥4,500.
Let's try it.
The next day, Saeki created accounts for ChatGPT Plus and Midjourney.
The first thing Saeki asked ChatGPT was a simple question:
"Please give me 5 picture book ideas for early elementary school children that leave readers feeling warm and gentle."
Back came five story concepts: a cloud child's adventure, a lost star, friendships among forest animals, and more. The one that caught Saeki's eye was "a story about a tiny cloud that descends from the sky and finds a friend on the ground."
My daughter would love this.
And so began a dialogue with ChatGPT.
Over the course of one week, a ten-page story was complete. Each page had just two or three lines of text — short, rhythmic, easy to read aloud. ChatGPT helped refine the phrasing again and again.
The next hurdle was illustrations.
She joined Midjourney's Discord server and typed in her first prompt.
/imagine a fluffy white cloud child floating down to a small park, soft pastel colors, children's book illustration --ar 16:9
Seconds later, four candidate images appeared. All beautiful — but too visually complex to work as picture book pages.
I want something simpler, easier for children to look at.
Drawing on a practical AI picture book creation guide, she adjusted the prompt:
simple illustration for children's book, a small fluffy cloud character landing in a park, minimal background, watercolor style, soft lighting --ar 3:2 --v 6
This time, a warm, page-sized illustration appeared.
She repeated this process for all ten pages — an average of five generations per page. More than 50 attempts in total. Night after night she opened Discord and fine-tuned her prompts.
Once the story and illustrations were ready, it was time for layout. Saeki chose Canva. Even on the free plan, Kindle-ready picture book templates are available.
She uploaded the Midjourney images, placed the ChatGPT text, chose a rounded sans-serif font at 24pt, and set the background to a soft cream color.
Let me show my daughter.
She opened the finished PDF on her iPad and read it aloud to her daughter.
"Mama, this is amazing! Fuwari is so cute!"
The moment her daughter's eyes lit up, tears welled in Saeki's eyes.
"I actually made a picture book."
On February 21, 2026, Saeki published her picture book through Amazon Kindle Direct Publishing (KDP). Priced at ¥250, with a cover also made in Midjourney.
Three days after publication, she received a LINE message from her daughter's nursery school teacher:
"Saeki-san, I heard you published a picture book! Would it be all right if we used it for storytime at school?"
Her daughter had apparently been telling everyone about it. The teacher purchased it on Kindle and read it aloud on an iPad — and the children kept asking, "Again! Again!"
47 copies sold in one month. Royalties came to about ¥3,200. A small sum in absolute terms, but for Saeki, it was proof that "a 42-year dream had turned into a number."
Saeki is now working on her second book. The theme: "A star falls from the night sky and gets lost." This time, her daughter is helping write the story.
"Thanks to AI, the excuse of 'I can't do this because I can't draw' disappeared. What I needed wasn't the tools — it was just that first step of saying, 'Let me try.'"
The HAIIA certification program offers a structured way to learn the skills of self-realization through AI. People like Saeki — who had "given up on their dreams" — are beginning to move again, all across Japan.
2026 is said to be a turning point when generative AI shifts from "a year to experiment" to "a year to integrate into work and tools." In personal creative pursuits as well, AI has become a practical tool for turning "can't" into "can."
From the time Saeki started to the day she published on Kindle: 45 days in total. One to two hours on weekday evenings, plus weekend mornings. No special skills, no expensive equipment required.
All it took was:
If you have something you've always wanted to do tucked away somewhere deep inside, search for it on your phone today. Try "AI + [your dream]."
You'll find your first step.
On the HAIIA community page, you can join a community of people who have started moving toward their dreams with AI. If starting alone feels daunting, you don't have to — walk the path with others who are doing the same.
Today, Saeki is spending another late-night hour on her second book. After her daughter falls asleep, she opens Discord and types a new prompt into Midjourney.
The title "picture book author" is no longer just a dream.
As of April 2026, copyright rules for images and text generated by AI vary by country and by service. Generally speaking, copyright is considered to belong to the user who created the prompt, but always check the terms of service for each AI tool before using it commercially. Paid plans for Midjourney and ChatGPT permit commercial use.
Personal AI-driven projects go far beyond picture books. AI is being used across all creative fields — novels, poetry, lyrics, music, illustration collections, manga, video, game scripts, and more. By 2026, AI agent technology has also matured, making it possible for individuals to tackle even more complex creative work.
We recommend starting with AI tools you can try for free. Ask the free version of ChatGPT for story ideas, or try making a simple picture book with Google's Gemini app. Once you're comfortable, consider upgrading to a paid plan and taking on more advanced creative projects. HAIIA's Three Pillars offers a concrete introduction to the steps of self-realization through AI.
This article is reprinted from HAIIA Notes.