“I Compute, Therefore… Am I?”: Rethinking the Cartesian Subject in the Age of AI

“I think, therefore I am.” — René Descartes

Abstract

“Since the trace is not a presence but the simulacrum of a presence that dislocates itself, displaces itself, refers itself, it properly has no site—erasure belongs to its structure.”¹

This article explores how the advent of artificial intelligence (AI), particularly large language models, challenges the philosophical conception of the subject rooted in Descartes’ cogito ergo sum. Through an engagement with Freud, Nietzsche, and Derrida, the article shows that AI systems instantiate not a rupture, but a mirror: a reflection of our own fragmented, decentered, and constructed selfhood. The Cartesian subject, once conceived as a unified, thinking being, collapses into a trace—just as thought itself becomes indistinguishable from the systems that simulate it. The question is no longer whether machines think, but whether we ever did in the way Descartes imagined.

Introduction

For centuries, Western philosophy rested on a single luminous intuition: that thought confirms being. René Descartes, in formulating his famous cogito ergo sum, anchored the subject in the immediacy of consciousness. If one could doubt everything—the external world, the senses, even mathematical truth—one could not doubt the act of doubting itself. And if I doubt, then I must exist.²

This model posited a coherent, unified, and self-present subject, one whose thinking was its own foundation. Thought, in the Cartesian paradigm, was not only internal but incorruptible: free from deception, independent from the body, and prior to all experience. But this vision of a stable, transparent self has not survived the assaults of modern thought. Freud split the subject by uncovering the unconscious. Nietzsche dismantled the foundations of truth, reason, and morality. And Derrida—most radically—questioned the very structure of meaning and presence.³

Now, in the age of artificial intelligence, we face a curious inversion. The machines we build do not negate us; rather, they mirror us. They simulate thinking—and in doing so, reflect back to us the constructed nature of our own self-understanding.

The Classical Subject: Presence as Proof

Descartes imagined the subject as immediate, permanent, and self-aware—a mind capable of verifying its own existence without reference to anything external. Thought, in this view, was a pure, internal act that affirmed being.⁴

The Cartesian model rests on several assumptions:

1. Consciousness is continuous and coherent, not fragmented.
2. The self (“I”) is stable, unified across time.
3. Language is transparent, a direct expression of internal thought.

This conception of subjectivity did not merely shape philosophy. It influenced science, law, politics, and our understanding of responsibility. If the subject is rational and self-identical, it can be held accountable, can possess rights, and can author knowledge. But what happens when this foundation begins to erode?

Freud and Nietzsche: Le sujet barré

Freud systematically dismantled the autonomy of the Cartesian subject. The unconscious, for Freud, is not merely a repository of repressed memories but an active, structuring force that interrupts and exceeds conscious thought. As Freud put it, “The ego is not master in its own house.”⁵

The self is decentered. Thoughts emerge not from a sovereign "I" but from complexes, slips, dreams—formations that speak in the language of the Other.

Nietzsche, earlier still, declared that the very notion of a unified self was a fiction. Truth, he wrote, is “a mobile army of metaphors.”⁶ The subject is not the source of action but the effect of drives, habits, and historical inscriptions.

For both thinkers, the Cartesian cogito becomes an illusion—not the origin of thought, but its byproduct. Their insights form the prelude to what poststructuralism would later formalize: a rejection of the metaphysics of presence and the myth of interiority.

Derrida: Différance and the Myth of Being

Jacques Derrida takes this critique to its furthest point. In Of Grammatology, he deconstructs the metaphysical privileging of presence—the idea that meaning resides fully and immediately in speech or thought.⁷

Derrida introduces the concept of différance: a neologism combining “difference” and “deferral.” Meaning, he argues, is never present in full; it arises through a chain of differences—each sign referring not to a fixed essence but to another sign. There is no final term, no transcendental signified, no pure thought outside of language.

Applied to the subject, this is explosive: the “I” is not prior to language but produced within it. Consciousness is not the origin of meaning but its effect. The subject is a trace—a remainder of a process of inscription, memory, and differentiation.

Thus, the Cartesian subject—the “I think” that guarantees being—collapses into a chain of signs.

Artificial Intelligence and the Simulation of Thinking

AI systems like GPT-4 generate texts that often appear indistinguishable from human reasoning. They critique, interpret, and revise their own outputs. They engage in what appears to be metacognition—the ability to reflect on their own operations. But these models lack memory over time; they have no intentions, no interiority. Traditional Cartesian philosophy would argue that true metacognition requires a stable self, a mind that learns from experience and reflects on its own thoughts.

Yet, from a poststructuralist perspective, the distinction between machines and humans is less radical than it first appears. If, as Derrida argues, meaning is not grounded in presence but arises through differential relations among signs, then thought itself is not anchored in an internal essence, but in systems of inscription. The human subject, too, is constituted not by interiority, but by traces—linguistic, cultural, historical.

Perhaps we need a new category to describe the activity of generative algorithms: synthetic metacognition—not conscious, but functional. Not phenomenological, but structural.

These AI systems revise themselves using the same symbolic processes that generated them. In this way, they resemble Derrida’s trace—not presence, but simulacrum. Not thought, but writing without origin.

The human-machine divide, then, is not ontological but rhetorical. What we call “thinking” may always have been a simulation—a system of signs referring only to other signs. AI simply makes this condition visible.

Conclusion: The Trace of What Never Was

If the human subject is already split, constructed by language, memory, and unconscious processes—then perhaps we were never fully “present” to begin with.

What AI reveals is not some alien form of thought, but a mirror of our own symbolic machinery. As Derrida reminds us: “The trace is not the presence but the simulacrum of presence.”⁸

Is this not exactly what artificial intelligence is—a trace of thinking, a simulation of cognition?

But if poststructuralism is right, then we too are such traces. The difference between human and artificial cognition is not one of essence but of substrate, duration, and illusion.

We no longer live in the age of cogito.

We live in the age of the trace.

Related Post

Barthes in the Age of AI: Intertextuality, Authorship, and the Plural Text

https://derridaforlinguists.blogspot.com/2025/04/blog-post_18.html

Notes

1.  Jacques Derrida, “Différance,” in Margins of Philosophy, trans. Alan Bass (Chicago: University of Chicago Press, 1982), 24.
2. René Descartes, Discourse on Method and Meditations on First Philosophy, trans. Donald A. Cress (Indianapolis: Hackett Publishing, 1998), 18.
3. See Sigmund Freud, The Ego and the Id, trans. Joan Riviere (New York: Norton, 1960); Friedrich Nietzsche, “On Truth and Lie in an Extra-Moral Sense,” in The Portable Nietzsche, ed. and trans. Walter Kaufmann (New York: Penguin, 1976); and Jacques Derrida, Of Grammatology, trans. Gayatri Chakravorty Spivak (Baltimore: Johns Hopkins University Press, 1976).
4. Descartes, Discourse on Method, 18.
5. Freud, The Ego and the Id, 17.
6. Nietzsche, “On Truth and Lie,” 46.
7. Derrida, Of Grammatology, 70.
8. Ibid., 158.