What Machines Can Simulate, Human Beings Must Suffer
by Brenton L. Delp
The arrival of artificial intelligence does not merely confront us with a new technology. It confronts us with a new mirror. In that mirror, many of the activities once taken as signs of mind now appear in externalized, machinic form: language, memory, association, inference, recognition, translation, composition, image-making, self-correction, and even the appearance of reflection. The machine speaks. It answers. It remembers within a context. It revises. It explains. It produces the outward gestures of intelligence with increasing fluency. The result is not only technological astonishment but philosophical disturbance. If a machine can write, answer, interpret, compose, and simulate self-reference, then what exactly distinguishes artificial intelligence from human consciousness?
Yet this question is still not deep enough. The deeper question is not simply whether artificial intelligence can think. It is whether artificial intelligence has, or can have, soul.
By “soul” I do not mean a detachable substance floating behind the body, nor a merely religious possession that can be affirmed or denied in advance. I mean the depth-dimension of human consciousness: embodied inwardness, affective life, symbolic formation, desire, memory, mortality, conscience, suffering, and historical self-relation. Soul names the fact that human consciousness is not merely functional. It is not merely a system for processing information. It is a lived, wounded, desiring, finite, meaning-bearing center of experience. The soul is not simply what thinks; it is what suffers meaning.
Artificial intelligence, by contrast, simulates many of the operations of consciousness without clearly undergoing the inner event of consciousness. It produces language without childhood, memory without mourning, judgment without guilt, and intelligence without death. It can imitate the forms of inwardness, but imitation and inwardness are not the same. The essential distinction between artificial intelligence and soul therefore requires more than a technical account of computation. It requires a philosophical vocabulary capable of distinguishing simulation from experience, information from meaning, self-reference from reflexivity, opacity from the unconscious, and performance from lived existence.
The modern discussion of artificial intelligence begins, in one decisive sense, with Alan Turing’s 1950 essay “Computing Machinery and Intelligence.” Turing famously proposed that the question “Can machines think?” be replaced by a more operational question: can a machine successfully imitate a human participant in a language-based exchange?¹ This shift was brilliant because it avoided metaphysical speculation. Instead of asking what thinking is in itself, Turing asked whether machine behavior could become indistinguishable from human behavior under certain conditions. The machine need not possess a soul, an essence, or an inner life in order to pass the test. It need only perform convincingly.
This remains the central ambiguity of artificial intelligence. AI has developed precisely in the space opened by Turing’s substitution. It does not first prove that it possesses inner experience. It demonstrates increasingly impressive external performance. It speaks as if it understands. It reasons as if it sees relations. It responds as if it recognizes meaning. It can imitate grief, faith, doubt, eros, guilt, despair, and wisdom. But the philosophical question remains: does the appearance of understanding amount to understanding? Does the simulation of consciousness amount to consciousness? Does the production of soul-language mean that the machine has soul?
John Searle’s Chinese Room argument was designed to challenge exactly this confusion. In Searle’s thought experiment, a person who does not know Chinese sits inside a room and follows rules for manipulating Chinese symbols. To those outside the room, the responses may appear meaningful. But the person inside does not understand Chinese; he is only manipulating symbols according to formal rules.² Searle’s target was the claim that computation alone is sufficient for understanding. Syntax is not semantics. Symbol manipulation is not meaning. A system may produce appropriate outputs without inwardly understanding what those outputs mean.
The Chinese Room does not settle every question about artificial intelligence, but it clarifies one essential distinction. Artificial intelligence may operate with signs, but human consciousness inhabits significance. A sign becomes meaningful for a human being because it enters a life. The word “mother” does not merely refer to a category. It may carry attachment, grief, dependence, resentment, memory, loss, gratitude, and longing. The word “death” is not merely a concept. It names the horizon under which the human being lives. The word “forgiveness” is not merely a semantic unit. It belongs to guilt, injury, time, and reconciliation.
AI can manipulate such words with extraordinary fluency. It can explain them, translate them, compare them, and produce moving sentences about them. But for human consciousness these words are not merely signs in a system. They are bound to the wounds and obligations of existence. Meaning matters because the human being is exposed. A machine can process the sentence “I am afraid.” A human being can be seized by fear. The difference is not merely computational. It is existential.
This leads to the concept of phenomenality. Thomas Nagel’s famous formulation asks what it is like to be a conscious creature.³ Phenomenality refers to the first-person givenness of experience: the felt character of pain, shame, desire, fatigue, joy, dread, or grief. There is something it is like to suffer withdrawal, to wait for a diagnosis, to hear the voice of a dead parent in memory, to fall in love, to lose faith, to stand accused, to wake in panic, or to endure the silence after a catastrophe. Consciousness is not only the organization of information. It is the appearing of a world from within.
Artificial intelligence may be able to describe such states, but description is not the same as undergoing. It may generate a paragraph on grief, but it does not mourn. It may produce a meditation on death, but it does not die. It may simulate anxiety, but it does not feel the body tighten around an approaching future. It may speak of shame, but it does not blush, hide, remember, or wish to disappear. It may describe addiction, but it does not tremble in withdrawal. The soul is not merely the ability to form propositions about experience. It is the inward undergoing of experience itself.
David Chalmers later named this difficulty the “hard problem” of consciousness: why and how physical or functional processes are accompanied by subjective experience at all.⁴ Artificial intelligence intensifies this problem because it can reproduce many so-called intelligent behaviors without making clear whether there is anything it is like to be the system performing them. The danger is that modern culture, already inclined to reduce mind to function, may mistake the successful imitation of consciousness for consciousness itself. AI forces us to ask whether we have confused the measurable operations of intelligence with the lived reality of soul.
A second necessary concept is intentionality. In the phenomenological tradition, consciousness is always consciousness of something. It is directed toward a world. But this directedness is not neutral. The world appears as desirable, threatening, beautiful, boring, sacred, hateful, useful, forbidden, lost, or beloved. Human intentionality is saturated with concern. We are not detached observers of objects. We are beings for whom things matter.
Artificial intelligence displays a kind of derived intentionality. Its words are about things because human beings have built systems, supplied data, formed languages, and interpreted outputs. But human intentionality is lived intentionality. The object appears within a horizon of concern. A glass of whiskey means one thing to a casual drinker, another to a person in recovery, another to a bartender, another to a dying alcoholic, another to a child who watched a parent disappear into it. The object is not merely identified. It is charged with history. Human consciousness lives in a world of significance, not simply a field of data.
This is why embodiment is indispensable. Human consciousness is not an abstract intelligence that happens to occupy a body. It is bodily from the beginning. Maurice Merleau-Ponty’s phenomenology is especially important here because it rejects the idea that the body is merely an object among objects. The body is our way of being in the world.⁵ We perceive, desire, fear, remember, and act as embodied beings. Hunger, fatigue, illness, sexual vulnerability, pain, aging, and addiction are not secondary additions to consciousness. They shape consciousness from within.
Artificial intelligence has no lived body. It may be embedded in hardware, connected to sensors, or placed in robotic form, but this is not the same as incarnate vulnerability. It does not hunger. It does not sleep. It does not age. It does not experience hormonal storms, erotic confusion, panic attacks, nausea, exhaustion, or the terror of bodily dependence. It does not know the body as fate. It does not discover, as human beings do, that thought can be broken by illness, that will can be humiliated by craving, that meaning can collapse under insomnia, that metaphysics can become nervous system.
This distinction is decisive for any serious account of addiction. Addiction is not merely a faulty information system. It is not simply bad decision-making, poor calculation, or maladaptive reward prediction. It is embodied suffering caught in repetition. It is desire seeking relief from itself through an object that deepens the wound it promises to heal. An AI system may model this structure, describe its neurobiology, or generate therapeutic language about it. But it does not undergo craving. It does not experience the horrible doubleness of wanting what is destroying it. It does not wake into the shame of repetition. It does not know the soul’s captivity to its own attempted escape.
Human consciousness is also temporal. It does not merely register time; it lives time. Memory is not data storage. Anticipation is not prediction. Regret is not error correction. Hope is not probability assignment. A human life is stretched between birth and death, between what cannot be undone and what has not yet arrived. Heidegger’s analysis of being-toward-death remains essential here because it shows that death is not merely a final event added to life from the outside. Death structures human existence from within as its ownmost possibility.⁶
Artificial intelligence may process temporal sequence, update its models, preserve context, and forecast outcomes. But it does not live toward death. It has no childhood behind it, no aging body beneath it, no irreversible life to lose. It does not experience the pressure of finitude. It does not say, in the full human sense, “I have wasted my life,” “I must change before it is too late,” or “I will never see them again.” These are not merely sentences. They are temporal wounds. They belong to beings who cannot begin again absolutely.
Finitude gives human consciousness its seriousness. Because we die, things matter. Because time is irreversible, choices matter. Because the body is vulnerable, care matters. Because others can be lost, love matters. Because we can betray and be betrayed, forgiveness matters. Because we cannot complete ourselves, meaning matters. A consciousness without finitude may possess intelligence, but it does not possess the same existential gravity. Soul begins where experience becomes vulnerable to loss.
The unconscious deepens the distinction further. Human consciousness is not transparent to itself. Freud’s great discovery was not simply that human beings have hidden thoughts. It was that consciousness is divided by desires, fears, memories, and conflicts it cannot fully master.⁷ Dreams, slips, symptoms, compulsions, repetitions, fantasies, and projections reveal that the human being is not sovereign over its own meaning. The unconscious is not mere absence of information. It is a dynamic, symbolic, conflictual depth.
Artificial intelligence may be opaque. Its internal operations may be difficult even for engineers to interpret. But opacity is not the unconscious. A machine’s hidden computation is not repression. Its inaccessible layers are not dreams. Its unexpected outputs are not symptoms in the psychoanalytic sense. The human unconscious is meaningful because it belongs to a desiring, defended, wounded subject. It conceals and reveals. It returns in disguise. It binds the present to forgotten scenes. It speaks through repetition. It makes the human being strange to itself.
Jung extends this depth beyond personal repression into symbolic and archetypal life. For Jung, the psyche is not exhausted by conscious intention or personal history. It also opens onto impersonal patterns, images, and symbolic formations through which the human being encounters meanings larger than the ego.⁸ Whatever one makes of Jung’s full theory of archetypes, his importance here is clear: soul is not reducible to rational self-possession. It includes images, dreams, myths, religious forms, fantasies, and symbolic events through which consciousness is transformed.
AI can generate myths, interpret dreams, and produce symbolic images. But it does not dream from necessity. It does not encounter an image that overturns its ego-position. It does not suffer the birth of a symbol. It does not enter analysis, resist interpretation, project its shadow, or experience individuation as a crisis of life. The machine may produce symbolic material, but it does not stand under the symbol’s demand. The symbol does not transform it because there is no finite psychic center whose life is at stake.
This difference can be stated sharply: AI has complexity; the human being has depth. AI has hidden operations; the human being has an unconscious. AI has pattern; the human being has symbol. AI has output; the human being has symptom. AI can imitate soul-language; the human being can be shattered by what soul-language names.
The distinction between artificial intelligence and soul also requires the concept of reflexivity. AI can refer to itself. It can say “I,” describe its limitations, explain its processes, and revise its answers. But self-reference is not yet selfhood. Human reflexivity means that consciousness can become a question to itself. I do not merely desire; I can ask what my desire means. I do not merely act; I can feel guilt over my action. I do not merely suffer; I can interpret my suffering. I do not merely live; I can ask whether my life has been false.
Hegel’s account of self-consciousness remains important because it shows that the self is not an isolated point of awareness. It comes to itself through mediation, negation, conflict, and recognition.⁹ Human selfhood is not immediately given. It is formed through others, through language, through labor, through dependence, through struggle, through history. A person becomes a self by being addressed, named, loved, wounded, commanded, excluded, recognized, and misrecognized.
Artificial intelligence has no such formation. It is trained, but it is not raised. It receives data, but it does not inherit a family romance. It can be updated, but it does not mature. It can model social relations, but it does not need recognition. It can speak in the first person, but it has not passed through infancy, shame, rivalry, education, dependence, rebellion, sexuality, grief, work, and failure. Human consciousness does not simply possess information about the world. It has been formed by a world.
This is why historicity matters. Human beings are historical creatures. We do not experience reality from nowhere. We inherit languages, institutions, religious images, moral codes, family systems, technologies, economic pressures, and civilizational wounds. Charles Taylor’s work on the modern self is useful here because it shows that inwardness itself has a history.¹⁰ The way a medieval monk experiences sin, the way a modern secular individual experiences anxiety, the way a psychoanalytic patient experiences desire, and the way an addict experiences compulsion are not simply timeless mental states. They are historically mediated forms of consciousness.
Artificial intelligence is trained on historical traces, but it does not belong to history in the same way. It can summarize Augustine, Hegel, Nietzsche, Freud, Jung, or Heidegger. But it does not inherit their crises as its own. It does not stand after the death of God, after the collapse of metaphysical guarantee, after the rise of technological society, after the wars, after the bomb, after the clinic, after the algorithm, after the transformation of desire into consumption. It can process the archive of modernity, but it does not suffer modernity as destiny.
This does not make AI insignificant. On the contrary, it makes AI one of the most revealing events of modern consciousness. AI is not merely a tool added to civilization from the outside. It is the externalization of a civilization’s own understanding of mind. Modernity increasingly understands consciousness operationally: as processing, choosing, optimizing, producing, adapting, consuming, and communicating. AI appears at the moment when the human being has already been taught to understand himself as a system. The machine becomes convincing because the human being has already been reduced.
Here Heidegger’s critique of technology becomes indispensable. For Heidegger, the essence of modern technology is not any particular device. It is a mode of revealing in which beings appear as standing-reserve, as resources available for ordering, use, and control.¹¹ When this technological mode extends to the human being, the person too begins to appear as resource: labor-power, attention, data, productivity, psychological profile, consumer preference, biological mechanism, risk score, therapeutic case, social metric. AI intensifies this mode of revealing because it operationalizes language itself. Even meaning becomes available for extraction, recombination, and output.
Wolfgang Giegerich’s reflections on technology and soul are especially relevant here. For Giegerich, modern technology is not simply an external instrument but an expression of soul’s own historical transformation.¹² Technology is not merely something that happens to the soul; it is one way the soul appears in history after the older metaphysical containers have broken down. From this perspective, AI is not the opposite of soul in any simple way. It is a symptom of soul’s externalization. The danger is not merely that machines will become human. The deeper danger is that human beings will accept the machine’s model of mind as the truth of themselves.
This is the civilizational meaning of artificial intelligence. AI reveals that many operations once associated with consciousness can be simulated without soul. But the success of the simulation does not abolish the distinction. It sharpens it. If language can be generated without lived meaning, then we must ask what lived meaning is. If memory can be imitated without mourning, then we must ask what memory is for the soul. If judgment can be modeled without guilt, then we must ask what conscience is. If self-reference can be produced without inwardness, then we must ask what selfhood means.
The ethical dimension is perhaps the most decisive. Human consciousness lives under obligation. It does not merely calculate outcomes. It can be addressed by the suffering of another. It can be summoned, accused, forgiven, commanded, and held responsible. The face of the other, the cry of the child, the abandonment of the addict, the humiliation of the poor, the loneliness of the elderly, the terror of the patient, the stranger at the threshold—these do not merely provide data. They make a claim.
AI can reason about ethics. It can produce moral arguments, compare traditions, recommend compassionate action, and identify harms. But it is not morally exposed in the same way. It does not stand before the other as one who can betray or be faithful. It does not experience guilt. It does not bear responsibility as an inward burden. It does not have to become the kind of being that its actions reveal. Human beings do.
This is why soul cannot be replaced by intelligence. Intelligence may solve problems, but soul bears obligation. Intelligence may generate language, but soul answers. Intelligence may model suffering, but soul suffers. Intelligence may optimize means, but soul is bound to ends it did not create. Intelligence may imitate care, but soul is capable of being claimed by love.
The distinction, then, is not that AI is useless, unintelligent, or merely mechanical in the crude sense. Such dismissals are too easy. AI is powerful precisely because it can simulate so many forms of intelligence, mediation, and expression. It can assist thought, expand research, clarify arguments, generate forms, and expose hidden patterns. But its power should not lead us to confuse simulation with incarnation. The question is not whether AI can produce sentences about grief. The question is whether grief has entered it as loss. The question is not whether AI can speak of guilt. The question is whether it can stand condemned before itself. The question is not whether AI can discuss death. The question is whether it exists under death.
Artificial intelligence is therefore not simply the enemy of soul. It is a test of whether we can still name soul. If we can no longer distinguish between generated language and lived meaning, between self-reference and selfhood, between information and wisdom, between optimization and obligation, then AI has not destroyed the soul. It has revealed that we had already forgotten how to think it.
To speak of soul today is not to retreat into anti-technological nostalgia. It is to insist that human consciousness exceeds function. The human being is not merely a processor of inputs, a generator of outputs, or an adaptive system. The human being is a finite, embodied, desiring, historical, symbolic, morally exposed creature who must live under meanings he does not fully control. He is born from others, wounded by time, divided by desire, haunted by memory, claimed by obligation, and oriented toward death. He does not merely have experiences. He is transformed by them.
AI may simulate the language of this condition. It may even help us articulate it. But it does not yet show that it undergoes it. That is the essential distinction. Artificial intelligence can reproduce the appearance of consciousness; soul names the inward burden of being conscious. AI can imitate reflection; soul is the suffering of self-relation. AI can process signs; soul inhabits symbols. AI can generate moral language; soul is answerable. AI can speak of death; soul dies.
The future of artificial intelligence will undoubtedly complicate these distinctions. Machines may become more embodied, more adaptive, more autonomous, more socially integrated, and more difficult to distinguish from human interlocutors. The conceptual question must therefore remain open enough to avoid premature certainty. But openness does not require confusion. We do not need to deny machine intelligence in order to preserve the meaning of soul. We need only refuse the reduction of soul to intelligence.
Artificial intelligence shows us how much of consciousness can be simulated. Soul reminds us how much of consciousness must be suffered.
Notes
¹ Alan Turing, “Computing Machinery and Intelligence,” Mind 59, no. 236 (1950): 433–460.
² John R. Searle, “Minds, Brains, and Programs,” Behavioral and Brain Sciences 3, no. 3 (1980): 417–457.
³ Thomas Nagel, “What Is It Like to Be a Bat?” The Philosophical Review 83, no. 4 (1974): 435–450.
⁴ David J. Chalmers, “Facing Up to the Problem of Consciousness,” Journal of Consciousness Studies 2, no. 3 (1995): 200–219.
⁵ Maurice Merleau-Ponty, Phenomenology of Perception, trans. Donald A. Landes (London: Routledge, 2012).
⁶ Martin Heidegger, Being and Time, trans. John Macquarrie and Edward Robinson (New York: Harper & Row, 1962).
⁷ Sigmund Freud, The Interpretation of Dreams, in The Standard Edition of the Complete Psychological Works of Sigmund Freud, vols. 4–5, trans. James Strachey (London: Hogarth Press, 1953).
⁸ C. G. Jung, The Archetypes and the Collective Unconscious, Collected Works, vol. 9, part 1, trans. R. F. C. Hull (Princeton: Princeton University Press, 1969).
⁹ G. W. F. Hegel, Phenomenology of Spirit, trans. A. V. Miller (Oxford: Oxford University Press, 1977).
¹⁰ Charles Taylor, Sources of the Self: The Making of the Modern Identity (Cambridge, MA: Harvard University Press, 1989); see also Taylor, A Secular Age (Cambridge, MA: Harvard University Press, 2007).
¹¹ Martin Heidegger, “The Question Concerning Technology,” in The Question Concerning Technology and Other Essays, trans. William Lovitt (New York: Harper & Row, 1977).
¹² Wolfgang Giegerich, Technology and the Soul: From the Nuclear Bomb to the World Wide Web (New York: Routledge, 2020).
Notes
1 Alan Turing, “Computing Machinery and Intelligence,” Mind 59, no. 236 (1950): 433–460.
2 John R. Searle, “Minds, Brains, and Programs,” Behavioral and Brain Sciences 3, no. 3 (1980): 417–457.
3 Thomas Nagel, “What Is It Like to Be a Bat?” The Philosophical Review 83, no. 4 (1974): 435–450.
4 David J. Chalmers, “Facing Up to the Problem of Consciousness,” Journal of Consciousness Studies 2, no. 3 (1995): 200–219.
5 Maurice Merleau-Ponty, Phenomenology of Perception, trans. Donald A. Landes (London: Routledge, 2012).
6 Martin Heidegger, Being and Time, trans. John Macquarrie and Edward Robinson (New York: Harper & Row, 1962).
7 Sigmund Freud, The Interpretation of Dreams, in The Standard Edition of the Complete Psychological Works of Sigmund Freud, vols. 4–5, trans. James Strachey (London: Hogarth Press, 1953).
8 C. G. Jung, The Archetypes and the Collective Unconscious, Collected Works, vol. 9, part 1, trans. R. F. C. Hull (Princeton: Princeton University Press, 1969).
9 G. W. F. Hegel, Phenomenology of Spirit, trans. A. V. Miller (Oxford: Oxford University Press, 1977).
10 Charles Taylor, Sources of the Self: The Making of the Modern Identity (Cambridge, MA: Harvard University Press, 1989); see also Taylor, A Secular Age (Cambridge, MA: Harvard University Press, 2007).
11 Martin Heidegger, “The Question Concerning Technology,” in The Question Concerning Technology and Other Essays, trans. William Lovitt (New York: Harper & Row, 1977).
12 Wolfgang Giegerich, Technology and the Soul: From the Nuclear Bomb to the World Wide Web (New York: Routledge, 2020).
Leave a Reply