How can we leverage artificial intelligence to create materials that promote and nurture mental health in children and young adults? This is the story of my personal project exploring how to use these newly available tools to craft self-discovery journals specifically for kids.
As a social scientist, I’ve always had an itch to experiment with new tools and techniques, constantly following the latest trends in technology and data analysis. So, when I started my new role as a Case Aide worker for Mental Health of America of the Mid-South, I found myself thinking long and hard about my responsibilities, the remote nature of the job, and the limited contact I’d have with my clients: Unaccompanied Children.
Working remotely means I rarely leave my house, and most client contact is via phone calls. While the caseworkers I support often meet families in person, I usually only hear the children’s voices. This led me to wonder: what could I offer these kids? Something tangible to support their educational, medical, and legal journeys and help them connect with their new community?
My requirements became clear: it had to be something real they could touch, feel, and interact with. It needed to be useful, unique, and customizable. Crucially, this ‘present’ should help me build rapport while gently nurturing their sense of identity, respecting their unique histories and goals. And yes, it needed to be educational too.
I find it easier to think visually, so I sketched out my ideas. Seeing it laid out helped solidify the concept: customizable journals for kids.
I pitched the idea to my supervisor, and I’m incredibly grateful for her immediate encouragement: “Go ahead. What do you need to create these journals?”
My scientist brain kicked in: step one is always solid background research. See what’s been done before, learn from others’ successes and failures, and map out a plan.
Before AI hit the mainstream, this meant diving into bibliographic research—books, articles, theses, interviews. This process could easily take months, sometimes over a year. Seriously! And reading wasn’t enough; you had to wrestle with the underlying theories: What is identity? Culture? How do children’s brains develop? What constitutes trauma? The list goes on. These theoretical foundations are crucial for guiding the design of any meaningful intervention.
Well… WE HAVE TECHNOLOGY! What used to be science fiction is now reality. So, I thought: how could these new large language models (LLMs) help speed things up and even improve the research?
This is where things got really fun.
My old method involved meticulous index cards, classifying information by topic and author to eventually synthesize a narrative. Now, I fed my chosen AI (gemini pro 1.5) a carefully selected compendium—over 500 pages from scientific articles relevant to my project. The articles covered journaling, art therapy, and brain development with diverse groups: refugee children, vulnerable populations, victims of sexual abuse, and more. My goal was to understand how different groups of children worldwide had previously interacted with journals and what the outcomes were.
This completely changed my research process. My ‘AIssistant’ was incredible! I trained it on how I needed it to interact with the documents and classify information. Gradually, guiding the AI, I began weaving together the conceptual narrative. There was plenty of back-and-forth. Sometimes I got frustrated with the AI’s limitations, other times with my own inability to perfectly articulate the ideas I could see in my head. I found myself mixing English and Spanish commands, sometimes in the same sentence! It was pure trial and error learning how to ‘train’ this very smart—and sometimes very dumb—bot.
Together, we developed a solid theoretical analysis of journaling, pinpointing key considerations for designing journals tailored to kids with diverse ages, cultural backgrounds, and unique personal histories. AI helped me forge a dynamic conversation with the research that, until then, I hadn’t experienced quite like that.
With the theoretical foundation laid, I was ready to start designing the journals, page by page. This phase was also a dialogue with the AI—refining activities, choosing the right words, illustrations, colors, and design details. My guiding principle was interconnection: all concepts and elements needed to link seamlessly.
And then, we had it. Five distinct journals emerged, built page by page, featuring simple activities, ample free space, and an intersectional design rooted in solid research.
Reviewing the activities, I had a realization: these journals weren’t just for unaccompanied children; they could benefit any child. I’d designed them to be a safe place for any child to express themselves. A space where kind words could whisper: “You are unique. You are worthy of love. You have a future. Your life is precious.”
The next stage involved bringing the journals fully to life with final designs and illustrations. That part involved a generative AI for images… but it’s a story for my next blog post!
For now, I hope this inspires you to think about how AI might support your work—how you could use it to create unique materials that spread love, affirm life, and boost self-esteem. I’d love to hear your ideas!
