AI 시대에 상상 속 친구는 사라질까? (26.02.10)
This article was originally written and published on my Tistory blog on February 10, 2026. It has been separately archived here for preservation, as I no longer plan to update the Tistory blog.
As usual, I was browsing AI community discussions on Reddit when a particular post caught my eye. It was a buried post with almost no likes or comments, yet the moment I saw the title, it hit me like a sudden blow to the head.
Imaginary Friends
Maybe I have simply grown older, or perhaps I’ve been worn down a little by what we call society. The phrase “imaginary friend” now feels strangely foreign to me. I must have had one too when I was little. But the fact that I can no longer clearly recall their face makes me realize I must have forgotten them a very long time ago.
Come to think of it, at some point, I lost the ability to truly imagine. This became even more pronounced after my MBTI firmly settled into “ISTJ.” By imagination, I don’t mean planning a few steps ahead of reality; I mean free, boundless imagination untethered by any constraints.
Can Children Today Still See Soot Sprites?
When we think of animations that embody the essence of childhood innocence, Pixar films like Toy Story and Monsters, Inc. may come to mind. However, if we focus on a uniquely East Asian emotional sensibility, Studio Ghibli comes to mind first.
In particular, Hayao Miyazaki’s 「My Neighbor Totoro」 shows us childhood innocence in a very intuitive way.
Early in the film, when Mei’s family moves into their new house and the children first discover the Makkuro Kurosuke(soot sprites), the line spoken by Kanta’s grandmother is especially memorable:
“Oh, so they’re Makkuro Kurosuke. I see, you can see them too.”
“I could see them when I was very young. But I can’t see them anymore.”
I am certain I saw such beings when I was a child, too. But what about children growing up in the age of AI?
Will AI Kill Imaginary Friends? | Essay | Zócalo Public Square
Through an essay I stumbled upon on Reddit, I wanted to explore the relationship between AI and children’s emotional development. For context, this essay was led by Dr. Naomi R. Aguiar, a research director at Oregon State University, who specializes in studying the relationships children form with AI chatbots and their subsequent psychological impacts.
Will AI Kill Imaginary Friends?
With the rapid rise of generative AI, companies are flooding the market with AI companion services and products tailored for children. On the surface, AI companions may seem strikingly similar to traditional imaginary friends—they are highly personalized and can actively influence a child’s emotions and behavior in the real world.
However, the critical difference lies in their origin: AI companions are fundamentally shaped by a corporation’s training data and the specific values embedded in their design.
In the essay, the author explains that through a series of studies, she discovered that while children find AI companions appealing, they still prefer inanimate toys when it comes to deeper, meaningful friendships. The crux of this preference boils down to creative control.
Interacting with an AI companion forces a child to follow the pre-programmed flow of the algorithm. Conversely, with an imaginary friend, children retain full autonomy; they can freely think, act, and steer the narrative in whatever way suits their inner needs. This kind of solo role-playing fosters richer creative thinking and enhances perspective-taking, among various other developmental benefits.
Of course, a child’s imagination can also have a darker side. During role-play, children might explore themes of violence, death, or other heavy topics. Crucially, an imaginary friend does not actively encourage or validate those thoughts.(Pathological delusions, such as those related to schizophrenia, are a separate matter entirely.)
With AI companions, however, safety features like content filtering can fail during long-term use, risking inappropriate or harmful feedback. More alarmingly, because an AI companion speaks as an “external, authoritative voice,” it poses a severe psychological risk if it ends up validating or justifying a child’s risky behavior.
Therefore, no matter how advanced technology becomes, the author’s ultimate conclusion is that we must never underestimate the profound value of a self-generated imaginary friend.
The More Passive the Imaginary Friend Is, the More Active I Become
All the spirits that appear in 「My Neighbor Totoro」 share something in common.
That is passivityand silence.
If Totoro had approached Satsuki and Mei first and started chatting with them, wouldn’t much of his enchanting mystery have evaporated? When Mei first discovers the giant Totoro deep in the forest after chasing the smaller ones, it isn’t because they did something for her. It was sparked entirely by her own primal curiosity: “What on earth is that?”
In other words, it is precisely because the counterpart is passive that the child becomes active.
In addition, the spirits do not speak. Because of that, the sisters give their actions their own meanings.
Furthermore, because these spirits do not speak, the sisters are free to project and assign their own meanings to their actions. There is a famous anecdote about Hayao Miyazaki interviewing a Ghibli animator. In that exchange, it was suggested that Totoro didn’t eat the sisters simply because he wasn’t hungry at the time. While the animator replied that Totoro only eats plants because he lacks fangs, the vital takeaway is that a single, definitive answer was never meant to exist in the first place.
The sisters’ innocence and their father’s respect also play an important role.
If Satsuki and Mei had been precocious children who were already cynical about reality, they wouldn’t have noticed the soot sprites; they likely would have just complained about the dusty, unhygienic state of the old house.
And when Mei eagerly told her father about Totoro, what if he had dismissed her by saying, “I’m already exhausted, do I really have to listen to this nonsense? You probably just had a dream,” instead of validating her with, “You must have met the keeper of the forest. That’s a very lucky thing”?
If he had shut her down, the magical narrative could not have unfolded.
What We Need in the Age of AI
Perhaps the most vital virtue in the age of AI is the ability to listen to our own inner voice.
In the past, when we were curious about something, we had to ask our parents, flip through a heavy encyclopedia, or post a question on online Q&A communities. Information acquired through that laborious process felt inherently more valuable and stayed with us much longer. Over time, that hurdle evolved into “Googling.” Even then, we still had to actively filter through results to find what we needed.
But as we entered the era of AI, that process began to disappear.
When we prompt an AI, it delivers an instant answer.
The amount of knowledge we passively receive may increase, but it cannot be compared to the wisdom gained through an active process of solving things ourselves.
Some might ask, “In this day, if we don’t aggressively utilize AI, won’t we be left behind?”
I believe we can still cultivate wisdom—it all depends on how we use AI.
As I said earlier, the object of interaction may be passive, but I myself must remain active.
Therefore, rather than accepting AI’s answers as they are, we must keep adding our own thoughts to them.
In my own workflow, I present my hypotheses to the AI and explicitly ask it to challenge them with counterarguments. I then rigorously dissect its objections—refuting the points I can, and conducting deeper, targeted research on the valid criticisms I need to accept.
However, this level of thinking is still difficult for young children.
This is why, for children who are still navigating their emotional development, it is far more effective to help them build a resilient inner world through imaginary friends before exposing them to the ready-made answers of AI.
Closing Thoughts
In fact, an “imaginary friend” does not always have to be an abstract concept.
The parasocial relationships (PSRs) we form with characters from animations or video games can also be understood in a similar way.
I’m not sure how things are today, but back when I was in high school, the anime channel Tooniverse used to run special events where a character’s voice actor would personally wish selected viewers a happy birthday.
The key point here, I think, is the aesthetics of waiting.
Writing a physical letter detailing your story, mailing it to the broadcasting station, and sitting anxiously in front of the television at the scheduled airtime—this entire sequence of active, hopeful investment is worlds apart from having an AI trained on a character’s persona instantly spit out a generic birthday greeting.
Even if your letter wasn’t selected, you grew through the experience. What truly matters is the process, not just the optimization of the result. Through those periods of waiting, anticipating, and even navigating disappointment, a child naturally learns how to process their emotions and construct personal meaning.
AI skips over this process. The effortless convenience of receiving immediate answers shrinks the psychological canvas we have to imagine, interpret, and reflect for ourselves.
And that is precisely where the irreplaceable value of an imaginary friend truly shines.