Visualizing Children's Imagination with Generative AI!? Unleashing Kids' Boundless Imagination at the "AI LOVE YOU Exhibition"

Is AI a "bad" thing that will take away human jobs, or a "good" thing that shows us the future...? With AI finding more applications in recent years, some people might be feeling uncertain about how to perceive it.

As an attempt to explore the potential of AI in new ways, Dentsu Inc. collaborated with Wit Collective LLC, which handles AI production projects, to hold the "AI LOVE YOU Exhibition." This event used generative AI to visualize children's imaginative thoughts and foster their imaginative abilities. (Held April 18-20 at "WHEREVER" in Roppongi)

In this article, Nami Shimoho from Dentsu Inc.'s Creative Planning Division 1, who handled the exhibition statements and panel copy, shares insights into the event alongside the staff's thoughts, behind-the-scenes stories, and future plans, incorporating voices from the team.

We wanted to show everyone, in a way they could understand, what AI really is.

The impetus for this exhibition came from Kazuhiro Hisataka of Wit Collective, with whom we'd previously collaborated. He approached us with the idea: "How about we do something together themed around generative AI?"

While I was interested in generative AI, it felt a bit intimidating. With mixed reactions circulating mainly on social media, I often wondered, "What exactly is AI to humans?"

So, before starting the project, I took time to learn about what generative AI can and cannot do from the AI production team at Wit Collective.

The most fascinating point I learned was that "generative AI excels at giving form to human ideas." No matter how wild a person's "I want to do this!" imagination might be, by conveying it in words to generative AI, it can create photos, illustrations, songs, novels, and more through those words. In terms of language, generative AI can transform anything conceived in any era, like waka poetry, into various forms.

This led me to think that children might be the most imaginative people. I wondered if AI could help expand their fantasies. This sparked an exhibition concept centered on imaginative children, using generative AI to visualize their fantasies.

Personally, I was fascinated by what forms generative AI might express in response to children's imaginations. Since I'd never seen an AI exhibition aimed at families before, I was excited about the possibility of creating something accessible to everyone—an exhibition that, in the best sense, didn't feel like typical AI.

Is children's "space for imagination" shrinking year by year!?

As we refined the exhibition content, one question arose within the team: "Is expanding children's imagination with generative AI truly a good thing?" To seek confirmation, we decided to consult Professor Shohei Tomita from Mie University's Faculty of Education, who researches children's imagination.

According to Professor Tomita, imagination provides people with a sense of peace and fosters new creativity through deep immersion. However, in today's digital environment, where information and answers are instantly accessible, there is less room for children to engage in deep, sustained imaginative play compared to the past. Amidst this, he explained that generative AI holds the potential to expand children's imaginative worlds through words, pictures, music, and various other forms. It can create a cycle of imagination and creativity, becoming a flexible and powerful tool.

Through this exhibition, we believe generative AI can become a supportive presence for children's imaginative power. Furthermore, by embracing children's diverse personalities and imaginative abilities, generative AI can help children discover their own "favorite parts" and grow to like themselves more. This led us to name the exhibition "AI LOVE YOU Exhibition."

Three Contents Expanding Children's Imagination

1. "Your Imagination, Shaped by AI"
With the help of three children, we used generative AI to create visual works that felt like stepping into their worlds, based on their "I want to be! I want to do!" imaginings.

Each child had 7-8 visual works of their imagination. When combining generative AI with their imaginative ideas, we particularly focused on showing the children the finished pieces and asking them, "What else would you like to see?" Our goal wasn't just "create and finish," but to use the works to further expand the children's imaginations.

The "imagination" works of the three children. The exhibition also showed how the initial creations were refined and improved.

The works were created in collaboration with the team of Katsuta Shimizu, the creative conductor* at Wit Collective.

※creAItive conductor
A unique title within Wit Collective for producing across multiple generative AI platforms

Mr. Shimizu, who worked on the generative AI piece.

Shimizu:
What I focused on most with generative AI creativity was expanding children's imaginations in exciting new directions. Using information from interviews and photos of the children as references, I generated about 1,500 images per child. As a parent myself, I aimed to create impactful visuals that would make them exclaim "Wow!" at first glance.

2. "Imagination AI Photo Studio"
We wanted children visiting the venue to experience visualizing their own fantasies, so we created the "Fantasy AI Photo Studio" as a photo spot. Children could input their imagined fantasies on the spot, and generative AI would turn them into images. They could then have their photo taken in front of a projector with the generated world as their backdrop.

Scene from the Imagination AI Photo Studio.

Kazuhiro Hisataka of Wit Collective led the operation of the generation flow and projector setup.

Mr. Hisataka, who played a central role in operating the Imagination AI Photo Gallery.

Hisakazu Kudaka:
Initially, we considered a workshop-style approach where participants would generate images on PCs. However, we wanted children to feel the excitement of imagination itself and create a more immersive experience. That's why we proposed a photo spot using projectors to project images, allowing them to take photos that felt like stepping into an imaginary world.

Actually setting it up was quite challenging—creating an immersive space in a limited area where sunlight streamed in (laughs). We tested it repeatedly, switching projectors, and finally succeeded. Seeing the children's joy at the event confirmed this approach was the right choice.

3. "Generative AI Q&A Corner"
We displayed a panel-based Q&A corner addressing basic questions about generative AI, such as "What's so amazing about generative AI?" and "What should I be careful about when using generative AI?" This content was also included in the venue pamphlet so families could learn together at home.

We also displayed comments from Professor Tomita, who kindly assisted us, on "Children's Imagination and the Potential of Generative AI."

Generative AI Q&A Corner and Professor Tomita's panel photo.

Additionally, as a souvenir for visitors, we distributed "Letters from Future Children"—messages written by children in 2030 to you in 2025. The content of these letters was generated by AI. We created about 10 different patterns—children who are good at sports, children who love singing, shy children, etc.—and made them available for visitors to pick up randomly.

In the center of the exhibition space, we also installed a monument featuring images of fantasies collected beforehand from children other than the three who collaborated. These images were printed and affixed to the monument.

(From top photo) "Letters from Future Children," "Monument of Everyone's Imagined Homes."

Two Key Messages Conveyed Through This Experimental AI-Powered Exhibition

Exhibition Statement and Panel Text by Shim.

When creating the exhibition statement and panel text, I focused on conveying two key takeaways: "Generative AI can be a partner that helps children's imagination" and "Imagination is important for both children and adults."

Additionally, aiming for panel text that parents and children would want to read together, I selected simple sentences and kanji that even children in sixth grade or younger could understand. (I even had the generative AI input the panel text to help with kanji selection!)

Key visual displayed at the venue.

We commissioned Hitomi Matsushita, an art director at Dentsu Inc., to adjust the venue atmosphere and the tone of the exhibited works.

Ms. Matsushita, who served as Art Director.

Ms. Matsushita:
With motifs like sharks and the ocean, we aimed for a cute tone that wouldn't feel too raw. To unify the exhibition's overall impression, we used generative AI to create pastel-toned images and then unified the tone around those target images.

Generative AI inevitably incorporates unconscious stereotypes or biased imagery from past data. We prioritized carefully listening to the children's own words and feelings, and thoroughly reflecting them in the visuals.

I believe generative AI served not as a "restrictor of children's sensibilities," but rather as a "reliable ally that further expands their free imagination."

Scene at the venue.

The bright future possibilities created by AI interacting with children

Visitor numbers exceeded expectations by more than double. Multiple media outlets, primarily web-based, covered the exhibition. Influencers sharing Tokyo event information also featured the "AI LOVE YOU Exhibition" as a recommended event.

The visitors were primarily the anticipated family demographic. We received many positive comments from both adults and children, such as: "It's fascinating to see children's imagination take shape," "Generative AI might be something that brightens children's futures," and "Even when the result is different from what I imagined, it's still beautiful and fun!"

We also learned that capturing children's free-flowing imagination becomes challenging after age 10. Simply pairing a child's photo with their imagination to create a piece proved a positive experience for families.

These discoveries and the significance of the "AI LOVE YOU Exhibition" align with Dentsu Inc.'s unique AI strategy, "AI For Growth," which advocates contributing to the growth of client companies and society by combining "human intelligence" with "AI intelligence."

The "AI LOVE YOU Exhibition" team is planning to expand the project beyond this exhibition, including participating in other events and developing services to preserve children's imaginative creations. If you think, "I want to join! I can do this!", please reach out to us.

【"AI LOVE YOU Exhibition" Project Team】

Dentsu Inc.：
Shimohona Nami (Creative Planning Division 1) · Sato Yoshifumi (Creative Planning Division 3) · Matsushita Hitomi (Creative Planning Division 3) · Shigemasa Naoto (Creative Planning Division 2) · Shimizu Takahiro (Creative Planning Division 3) · Kitano Ayumi (Creative Planning Division 3)

Wit Collective:
Kazuhiro Hisataka・Katsuta Shimizu・Fui Takano・Ichizen Eda・Kota Mikuro

About Wit Collective LLC

We produce new value in entertainment under the motto "When hearts move, the world moves." Leveraging expertise in advertising, film, events, and more, we handle everything from concept design to production and deployment with the optimal team. In our AI Production division, our "CreAItive Conductors"—visual media professionals with deep expertise in generative AI—design outputs by cross-utilizing cutting-edge technologies. We expand the possibilities of visual expression through generative AI, delivering novel entertainment experiences. We also operate "DO/AI," an AI video production service that leverages generative AI to provide advertising videos and IP development. Website URL: https://wit-collective.jp/