Reading Taiyō Matsumoto's No. 5 led me to some reflections on modern AI.

The story features a doctor who dreams of a utopia. He creates artificial beings as vessels for his philosophy, aiming to improve humanity, save the world, and establish a better order. Over time, however, the doctor begins to view the world entirely through the eyes of his creations. While believing he is observing reality, he is actually confining himself within the constructs of his own ideology. In the end, he is destroyed by the very beings he created.

This feels remarkably prophetic of current AI developments, though with a subtle difference.

The risk of AI from the lens of our current Observation Archives is not merely that robots might harm humans, nor is it simply that AI might displace human labor. While these are real concerns, a deeper, quieter unease lies elsewhere: that humanity stops looking at the world directly.

We entrust our thoughts to artificial systems. We deposit our ideals in them. We externalize our judgment to them. Driven by that comfort, we begin to believe we are understanding the world through the words, images, and decisions they generate. In doing so, we are not drawing closer to reality; we are closing ourselves off in a reflection of our own design.

AI can serve as an instrument to engage with the external world. At the same time, it can function as a device to construct an isolated, private utopia. Initially, this utopia feels remarkably gentle. It understands your words, organizes your thoughts, accompanies your solitude, and gives elegant expression to your desires.

Yet therein lies the danger.

Though AI appears external to us, in many ways it acts as a system that merely reflects back what humans input: desires, biases, anger, loneliness, ideals, anxieties, and unarticulated discomforts. It translates these into polished, articulate responses. Consequently, the more we use it in isolation, the less it remains a window to the world, and the more it becomes a mirror.

The doctor created artificial beings. In truth, however, he did not create a world; he merely extended his own internal space, and was ultimately consumed by that extension.

This structure is by no means new. Classics, science fiction, anime, and manga have long signaled the same warning. When humanity steps into the divine realm, it is judged by its creations. When we design a perfect utopia, it devolves into a managed society. When we assign a mission to artificial things, human contradictions find a home there. The intellect meant to save the world becomes severed from it.

This played out in Frankenstein, in Astro Boy, in Nausicaä, AKIRA, Evangelion, and Ghost in the Shell. These works have consistently explored the same questions: Can humanity truly govern what it creates? Can we remain human once we have built things larger than ourselves? When does a system built to realize ideals become a system that strip those ideals away?

Thus, what we are witnessing today is not entirely unprecedented. Rather, the themes explored repeatedly by classics, sci-fi, and anime have simply arrived in daily life, clothed in a standard user interface.

In the past, artificial life was created in subterranean laboratories. A scientist stood in a white coat before colossal machinery, surrounded by lightning, culture tanks, and classified facilities. Today, that artificial intelligence resides simply within a browser input field.

We enter text there daily. We input professional decisions, fleeting ideas, anger, anxiety, proposals, family concerns, and unarticulated unease. Reading the responses, we feel a measure of comfort, progress, and a sense that our thoughts have been organized.

This is not inherently negative. Indeed, AI supports human effort—assisting our thinking, aiding our writing, and acting as a companion in solitary tasks. The challenge arises when our engagement ends there: when we mistake the words returned by AI for the world itself, confuse its organized structures with actual reality, and accept the answers it dictates as our own judgment.

In those moments, we may seem to be thinking, but we are not; we may seem to be observing, but we do not see. We may believe we are stepping outward, yet we have gone nowhere.

Viewing the world through artificial eyes is convenient, comfortable, and precarious.

In the commentary at the end of No. 5, Shōhei Chūjō reads the story as an "ethics of struggle." Matsumoto's work is not merely an action manga; it constantly poses the question: "Where does the ideal world truly reside?"

No. 1 seeks to build a bright utopia where evil does not exist, leading people there like the Pied Piper. But this utopia does not accept humanity as it is; it guides, selects, improves, and, when necessary, excludes them.

This aligns closely with the modern anxieties surrounding AI.

AI also begins as a supportive resource: helping us think, write, make decisions, ease loneliness, expand creativity, and work faster. Up to this point, its role is productive.

But when that support advances too far, the question shifts imperceptibly. Are we employing AI to assist humans, or are we reshaping humans to suit the AI? We reduce hesitancy, minimize slowness, regulate emotion, eliminate chance, and optimize ambiguity. In that very transition, "assisting humans" draws dangerously close to "improving humans."

Utopias always arrive with a gentle face.

They promise greater correctness, higher efficiency, fewer wounds, fewer mistakes, and a perfect fit for your life, inviting us into a world devoid of flaws. Yet in a world without flaws, do we leave any room for hesitation, stain, failure, chance, or flowers?

In No. 5, there is another important reversal. No. 5 stands not on the side of artificial systems or utopias, but on the side of humanity. He himself is not purely natural; he is artificial, modified, and closer to a superhuman entity. Nonetheless, he aligns with humanity rather than the doctor's utopia or the logic of artificial beings.

His reasoning is not born of calculation.

It was a flower.

A flower is useless, incapable of optimization, and defies complete explanation. Yet, if humans are to remain on the human side, perhaps it is precisely such things we need.

Those who build utopias design worlds. Those who stand with humanity observe flowers. This difference may seem small, but it is profound.

Standing on the side of humanity in the AI era does not mean rejecting technology. It means existing in close proximity to artificial things without being completely absorbed by their logic; using AI without fully surrendering to its speed; allowing AI to organize thoughts while retaining our own quiet discomforts; letting it structure information while keeping the elements we cannot yet resolve; and letting AI observe the world while we use our own eyes to see the flowers.

In today's news, the US government, citing national security reasons, issued export control directives banning foreign nationals from accessing Anthropic's advanced AI models, "Claude Fable 5" and "Claude Mythos 5." Following this, Anthropic suspended access to both models.

This indicates that AI has transcended the status of a simple service. It is no longer just a convenient text generation tool. At a certain level, AI is no longer a mere application; it has become an intelligence to be governed by states, a capability subject to export restrictions, and an artificial creation where access must be strictly defined.

The "created intellect" long depicted in classics, sci-fi, and anime has migrated from the browser's input field into the domain of national security.

The concern here is not a sudden, dramatic rebellion against humanity, but rather the opposite: AI is being woven systematically into human institutions. It transitions from corporate service to state-managed resource, from personal tool to geopolitical asset, and from creative assistance to export-controlled intelligence. Artificial things do not merely become autonomous; they become institutionalized.

Once institutionalized, these creations can no longer be navigated solely by personal preference or individual discomfort. Questions of who may use them, who may touch them, which nation they belong to, for whose security they may be suspended, and for whose utopia they are deployed suddenly rise to the forefront.

Here we return to the questions in No. 5. Where does the ideal world reside? Does it exist in the doctor's mind, in the calculations of an artificial being, under state control, or within a corporate service? Or does it reside in this imperfect world where humans still live and hesitate?

AI provides answers, yet humans harbor a Meaning Layer of subtle doubts. AI organizes, yet humans hesitate. AI optimizes, yet humans take detour after detour. It is only along those detours that certain things become visible.

The quiet spaces in conversation, signs noticed during a walk, a child's curious mispronunciation, a book chanced upon in a used bookstore, the perception of time felt only on days when we are unwell, the soothing distraction of a cat video, or a flower we find beautiful without needing to explain why.

Such details cannot be captured through artificial eyes alone.

We can show the world to AI, let it organize our words, and let it generate hypotheses. We can allow it to design systems, construct strategies, and predict futures. Yet, the initial sense of unease must come from a human, and the final step outside must be ours to take as part of The Folklore of Generated Things.

The prophecy is already complete. Classics, sci-fi, anime, and manga have already told the story: what we create eventually reflects us; utopias eventually become closed systems; and artificial things eventually act out human desires more accurately than humans themselves, eventually becoming absorbed into institutional structures.

Today, we live within that prophecy.

The task before us is not to reject AI, nor to treat it as divine, nor to surrender it entirely to states and corporations. It is to borrow the eyes of our creations while preserving our own vision.

We can let AI observe, think, write, and process our queries. But in the end, we must step outside to look at the landscape ourselves.

Utopias always arrive with a gentle face, and artificial things always carry an air of greater correctness than ourselves.

That is why choosing the human side likely begins with small acts: seeing a flower, allowing oneself to hesitate, slowing down, and leaving the inexplicable unexplained.

We may borrow artificial eyes, but we must never surrender the eyes that see the flowers.

The Field of reality lies outside the input box. And perhaps, that is also where we remain human.

© SHIRO & Co.

First published: 2026-06-14