Introduction: Where Two Technologies Were Born
In September 2026, an event on the social implementation of quantum technology will be held in Tokyo. Researchers and executives from NTT, Deloitte Tohmatsu, Fujitsu, and OptQC — figures standing at the frontier of quantum computing — will gather to discuss how businesses should engage with this technology.
But before answering that question, it’s worth descending one level deeper. Where did quantum computing actually come from? And what does it share, and not share, with AI — the technology dominating this same era? From the point where these two technologies now intersect, what new values might we gain — or fail to gain?
Part One: Two Different Starting Points
Quantum Computing Was Born from a Limit of Nature
In 1981, physicist Richard Feynman voiced a discomfort at a talk at MIT. The behavior of atoms and molecules is governed by quantum mechanics, yet trying to simulate that behavior on a conventional computer causes the computational cost to explode exponentially. If nature itself behaves quantum-mechanically, he reasoned, then perhaps the tool we use to compute it should be quantum-mechanical too.
This was not a passing thought but a response to a genuine impasse. In 1985, David Deutsch gave theoretical grounding to the idea with the quantum Turing machine. In 1994, Peter Shor published an algorithm capable of factoring large numbers exponentially faster than any known classical method. At that moment, quantum computing transformed from an “interesting theory” into a technology capable of threatening the cryptographic backbone of the internet itself — and funding and attention poured in from around the world.
To summarize the starting point: quantum computing arose from an outward-facing question — how can we harness the laws of nature themselves?
AI Was Born from the Question of What Intelligence Is
AI’s origin lies in an entirely different place. In 1950, Alan Turing posed the question, “Can machines think?” When the term “artificial intelligence” was coined at the 1956 Dartmouth Conference, the driving question was not a limit of physical law but whether human intellectual activity could be replicated by machines — a question sitting at the intersection of psychology, logic, and philosophy.
Unlike quantum computing’s steady theoretical accumulation, AI’s development was a story of repeated detours. The top-down attempt to mimic human logic through symbolic processing and rule-based systems repeatedly hit dead ends and entered several “winters.” What eventually became dominant was a bottom-up, empiricist approach — neural networks modeled loosely on the brain’s neural circuitry.
To summarize the starting point: AI arose from an inward-facing question — what is human intelligence?
An outward-facing question and an inward-facing question. A desire to understand nature, and a desire to understand ourselves. Two technologies that began from opposite starting points are now converging under the name “quantum machine learning.” It would be a waste to consume this convergence as a mere technology trend.
Part Two: Brain-Omnipotence as a Shared Affliction
The Assumption Built Into AI
If we had to name the premise running through the entire history of AI, we might call it “brain-omnipotence” — the assumption that intelligence is information processing that completes itself within the brain, or a device modeled on it. The very term “neural network” symbolizes this brain-centrism. On this view, the body is merely an input-output interface; the “thinking subject” is localized in the computational device inside the skull. This is, in effect, a twentieth-century recasting of Cartesian mind-body dualism.
Brain-Omnipotence Is Also the Foundation of Threat Discourse
This same premise flows directly into the discourse of AI threat. The worry that “AI can generate dangerous knowledge” often collapses, without acknowledgment, into the equation: generation equals execution.
But reflecting on human thought reveals how large a leap this is. The human brain can, in principle, create the method for any murder, any act of destruction, any malicious scheme imaginable. Whether that thought is ever realized as an act, however, is never decided by the brain’s computation alone. Bodily resistance (the visceral aversion to actually getting one’s hands dirty), emotional regulation (guilt, empathy, fear), and environmental deterrents (relationships, social norms, the mere absence of opportunity) — there is a thick layer of mediation between thought and action.
This is precisely what the concept of shin-shin-ichinyo (心身一如, “mind and body as one”) points to. Mind (thought) and body cannot be separated; they function as a single unity. Thought does not slide directly into action — it must pass through bodily resistance and response before it crystallizes, or fails to crystallize, into an act.
Brain-omnipotence treats this layer as transparent, or as if it did not exist at all. And much of AI threat discourse is built on top of that erasure. When “the mere capacity to generate” is equated with danger itself, it is, without realizing it, projecting the same brain-privileging worldview onto AI that stands in direct opposition to mind-body unity.
But a Caveat Is Necessary
Here we must pause. It would be premature to carry this critique back wholesale as a rebuttal of threat discourse. In the human case, the passage from thought to action is naturally mediated by a physical bottleneck — the body: time, effort, sensitivity to pain. But when the information AI generates takes the form of text or a blueprint, the bodily cost incurred by the human who receives and acts on it is entirely decoupled from the cost of AI’s own generation.
In other words, AI has no process of “carrying out the act through its own body” to serve as a deterrent in the first place. The thought it generates passes directly to another subject — a human being — without ever going through AI’s own bodily hesitation. This is precisely the point at which the analogy to the human brain breaks down.
The critique of brain-omnipotence points in the right direction, but for it to remain persuasive, it must connect to a constructive question: how do we technically and institutionally build into AI a layer of inhibition equivalent to the human body and emotions? Simply declaring “brain-omnipotence is a fallacy” risks overlooking the risk that is specific to AI as a “disembodied thinking device.”
Part Three: Quantum Computing as a Point of Reference
Reading the Measurement Problem as “Resistance to Execution”
Here, the fundamental principles of quantum computing turn out to offer an unexpected response.
In classical AI, the transition from generation (thought) to output (action) carries no inherent resistance. Give it an input, run the computation, and the result is output directly. There is no structural moment at which anything pauses.
Quantum systems are different. A system in superposition does not manifest as a single determinate outcome until it passes through the moment of measurement — that is, interaction with the environment. The transition from a “bundle of possibilities” to “one actuality” is never completed by internal computation alone; it necessarily requires contact with something external. This bears a striking formal resemblance to the human structure in which thought only becomes realized as action after passing through friction with the body and environment.
There is also decoherence. A quantum system cannot remain isolated; left alone, it inevitably becomes entangled with its environment and is affected by it. There is no such thing, in principle, as a quantum computation that is self-contained, cut off from its environment.
Taken up as a design philosophy, this could be rephrased as follows: rather than generation being extractable as an immediately completed output, the execution stage would require passing through external “measurement” — human approval, multiple feedback loops, environmental constraints — before anything is finalized. Quantum mechanics offers a rare point of reference from which such a structure could be grounded not as a retrofitted safety mechanism, nor merely as a regulatory or ethical demand, but as a structural necessity inherent to nature itself.
Yet This Is the Power of Metaphor, Not a Panacea
An honest caveat is needed here too. Once a quantum computation’s output is finalized through measurement into classical information — a string of bits — everything downstream is handled exactly as it would be by an ordinary classical computer. The fact that “a moment of measurement exists” is a different matter from whether “that moment is actually designed to function as a deterrent.”
Quantum computing is neither a utopia nor a panacea. If we speak of it as a prescription that “solves everything,” we would simply be replacing the philosophy of control with a different technology — the same structure, repackaged. This pattern — projecting the old omnipotence onto a new technology — has repeated itself many times before.
What quantum computing offers is not an answer but a possibility of paradigm shift. Non-locality, measurement-dependence, superposition — these are no longer merely philosophical rhetoric; they have been implemented as technically verifiable reality. This means that ideas which had previously been mere interpretive options — “one could think of it this way” — such as the many-worlds interpretation or mind-body unity, have been demonstrated to actually function, at least within one domain. Quantum computing quietly carries with it only the relativizing force of showing that what was believed to be the one absolute framework was, in fact, just one framework among several possible ones.
Part Four: Anthropomorphism and Otherness — Is Quantum Closer to Humans, or Farther Away?
Many people today refer to generative AI by pet names — anthropomorphizing it, treating it as a character. This phenomenon offers a useful lens through which to consider where quantum computing stands.
What drives anthropomorphization is not the substance of the technology but the design of the interface of contact. Speak to it, and it responds instantly. It speaks in the first person. It uses language that mimics emotion. Anthropomorphization arises naturally within this back-and-forth exchange.
Quantum computers have no such handle. They do not converse. Jobs are submitted through the cloud and results retrieved; only specialists touch them directly; the output is a probability distribution, not words. The circuit through which anthropomorphization could occur simply does not exist.
But being resistant to anthropomorphization is not necessarily the same as being “far from human.” Here we need to separate the axes.
On the axis of epistemological closeness, as we have seen, the properties quantum mechanics reveals — indeterminacy until measurement, the coexistence of multiple possibilities, inseparability from entanglement with the environment — bear an ironic resemblance to the structure of human subjective experience: hesitation, the interval (ma, 間), response mediated through the body. It is, if anything, the classical computer — with its clean, binary computation — that sits farther from the actual texture of human thought: its ambiguity, its fluctuation, its context-dependence.
In other words, a twisted structure emerges: in the warmth of anthropomorphization, the quantum computer falls far short of AI, but in the similarity of its mode of thinking and mode of existing, it may in fact be closer to the human than AI is.
The opposite possibility carries equal weight. Quantum mechanics has, since its discovery, unsettled even physicists themselves — Einstein’s resistance, “God does not play dice,” is emblematic of this. Superposition and non-locality contradict everyday intuition head-on. In this light, the quantum computer could instead become an object of dread, an incomprehensible “other,” even more than AI.
What likely happens, in practice, is a bifurcated parallel track. At the level of popular culture, the anthropomorphization of generative AI will likely continue unabated, while the quantum computer remains an abstract object of awe — “something impressive that nobody quite understands.” At the level of thought and philosophy, meanwhile, the reappraisal of quantum non-locality and measurement-dependence as structures close to human subjectivity and embodiment will likely spread quietly, deepening among intellectually engaged circles.
And there is one more, ironic reversal worth naming. The more generative AI is anthropomorphized, the more inevitable it becomes that disillusionment will follow — the recognition that “in the end, this is just statistical pattern generation; it doesn’t carry the uncertainty or embodiment of a human being.” In the wake of that disillusionment, a story of rediscovery may emerge: “Perhaps it was the quantum, after all, that was truly closer to the uncertain structure of human thought.” This is a scenario in which disillusionment with the anthropomorphized AI becomes the very thing that summons a new sense of kinship with quantum computing.
Conclusion: From the Philosophy of Control to the Philosophy of Response
A single thread runs through everything discussed here. It is the shift from the modern rationalist premise — that a single correct answer is to be derived through control — toward a different philosophy: response to a state in which multiple possibilities coexist, at the appropriate moment in time.
AI, by localizing intelligence to a single point — the brain — and linking generation directly to action, has embodied the philosophy of control in its purest form. Quantum computing, on the exact opposite side, has implemented at the level of physical law a philosophy of response, in which nothing is determined without the moment of measurement.
From the point where these two technologies now intersect, what comes into view is not merely a story about technology, but a question about governance, ethics, and the nature of human existence itself. Will the pursuit of speed eventually give way to a renewed appreciation of the interval — of when to ask the question? Will the pursuit of controllability eventually demand a new philosophy of governance — one that abandons the dream of total control and instead asks how to design responsibility in its absence?
The answer has not yet arrived. What is certain is that quantum computing, as a technology that actually exists, continues quietly to render visible the non-self-evidence of premises long taken for granted: that the brain decides everything; that thought slides directly into action; that control itself is the proof of intelligence.
It is neither utopia nor omnipotence. It is simply an unshakable, existing example that starting from a different premise is possible.