Artificial ‘Intelligence’: A Fallacy Since 1956

When Learning’ Lacks Agency: The Fatal Flaw in Machine Epistemology

The fathers of modern Western philosophy, Immanuel Kant and René Descartes, both emphasized the importance of reason and the mind in understanding the world. Descartes is best known for his “Cogito, ergo sum” (“I think, therefore I am”) and Kant for his “Critique of Pure Reason,” which grounds all knowledge in rationalism and empiricism — emphasizing the role of the conscious mind in shaping our human experience.

The promise of “artificial intelligence” is a grand illusion — a philosophical sleight of hand that conflates computation with cognition. Terms like neural networks and machine learning evoke biological and psychological legitimacy, yet they describe mere statistical tools devoid of consciousness, the very bedrock of intelligence.

Aristotle defined nous (intelligence: the highest form of intellect, understanding, or "mind,") as the soul’s pursuit of truth and virtue; Descartes proved that doubt — an act of conscious will — is the essence of thought. Machines do not doubt, desire, or interpret, and Kant showed that subjective minds mediate reality.

Machines process data, nothing more. To call this “intelligence” is to indulge an anthropomorphic fantasy, mistaking mimicry for meaning. The AI debate isn’t about technology’s potential — it’s about defending the irreducibility of human consciousness against reductionist hype.

Therefore, the continued kicking of the ball down the road with the many promises of “artificial intelligence” amounts to a grand illusion — a philosophical sleight of hand that conflates computation with cognition.

Kant famously stated, “All our knowledge begins with the senses, proceeds then to the understanding, and ends with reason,” highlighting the progression of knowledge from sensory experience to rational thought and expression.

Descartes argued that consciousness was the first and most certain truth of human intelligence, a foundation upon which we can build further knowledge. Hence, intelligence is a function of authentic reasoning based on our conscious understanding of the real world.

A Brief History of Artificial Intelligence

In the book, A Brief History of Artificial Intelligence, Oxford Professor Michael Wooldridge brings a dose of healthy reality and humility to the over-hyped world of AI, that the real story of AI elucidates a history of boom-and-bust cycles.

Wooldridge highlights that AI has “always attracted crackpots, charlatans, and snake oil salesmen as well as brilliant scientists.” Much of what is published about AI in the popular press is ill-informed or irrelevant, so the more we listen to the hype, the more misinformed and ignorant we can become.

The computational complexity surrounding AI has led many to turn to simplistic stories about things they don’t understand. These nice comforting stories have taken on a fervour of optimism and irrationality — making AI rhetoric into a new religion.

The terms “neural networks,” “machine learning,” and “artificial intelligence” are theoretical constructs masquerading as scientific inevitabilities. These frameworks lack empirical evidence for their foundational claim: that consciousness-free systems can replicate intelligence, a concept inseparable from subjective experience.

The phrase “artificial intelligence” is itself a contradiction — intelligence, as classically understood, requires consciousness, agency, and self-awareness, qualities no machine possesses.

“Artificial intelligence” is an oxymoron. Intelligence, as defined by philosophy’s greatest minds, requires consciousness (Descartes’ cogito), subjective agency (Kant’s transcendental ego), and the capacity to infuse raw data with meaning. AI’s proponents sidestep this by redefining intelligence as “problem-solving” — a rhetorical pivot that reduces millennia of human inquiry to an engineering challenge.

Even in ancient times, for Aristotle, intelligence is not merely problem-solving or pattern recognition — it is the soul’s active pursuit of truth and virtue, rooted in a living being’s purpose. He would critique AI’s mechanistic model as a parody of intelligence for three reasons:

1. No Telos, No Truth: Machines lack intrinsic goals. Aristotle saw intelligence as inseparable from telos — the purpose guiding a being’s actions. An LLM generates text, but it doesn’t seek truth or wisdom; it has no stake in why it exists.

2. Soulless Computation: Aristotle’s “soul” (psyche) is what animates and unifies a living being. Intelligence requires this holistic integration of body and purpose. A neural network, being inert matter arranged by humans, cannot possess the unity or intentionality of a soul.

3. Moral Blindness: Aristotelian phronesis (practical wisdom) demands moral reasoning — weighing actions against the “good life.” AI systems, devoid of values or self-awareness, optimize for external metrics (e.g., accuracy) but cannot judge whether those metrics align with human flourishing.

If the practical Aristotle were having a conversation with the perfectionist Plato, he would say: “calling machines ‘intelligent’ confuses calculation for cognition. True intelligence breathes, desires, and strives — it cannot be simulated, only lived.”

Therefore, AI’s core problem is that its mechanistic definition of intelligence collapses against reality: it reduces human cognition to computational models, ignoring the metaphysical dimensions of consciousness — like sensory qualia and subjective agency — that resist quantification. This oversight mirrors a centuries-old epistemic rift, one Descartes and Kant revealed as inseparable from authentic intelligence.

The applied intelligence counter about AI isn’t about technology’s potential — it’s about defending the irreducibility of human consciousness against reductionist hype — applied intelligence (ai)  involves applying critical and independent thinking to transform information into knowledge, fostering insight and rational strategic decision-making.

Historical Context

It can be argued that Galileo ushered in the Scientific Revolution, new concepts of physics and astronomy, marked by precise measurement and systematic observation. Galileo’s exclusion of subjectivity from measurable “primary qualities” found its philosophical echoes in the philosophies of Descartes and Kant.

Descartes’ cogito (“I think, therefore I am”) established consciousness as the indubitable foundation of intelligence: doubt, will, and imagination are acts of thinking, not computation. Descartes’ cogito and Kant’s transcendental unity of apperception rooted intelligence in conscious self-awareness. For Descartes, doubt — an act of consciousness — was the proof of existence; for Kant, the mind’s ability to synthesize experience defined reason. Both rejected mechanistic explanations of thinking.

Kant later argued that the mind actively structures reality through categories like time, place, and causality — a process that requires subjective agency. For both, intelligence is constituted by consciousness, not reducible to its outputs.

Galileo’s quantitative framework, while revolutionary, excluded consciousness from “science,” enabling later theorists to conflate computation (a tool) with cognition (a conscious act).

Modern AI plays fast and loose with science. AI inherits Galileo’s quantitative framework but stumbles when confronted with Descartes and Kant. LLMs process data but, unlike human conscious intelligence, cannot doubt their inputs, will a new perspective, or structure reality through subjective experience. They mimic mathematical reasoning (Descartes’ “mathesis universalis” or “universal science”) yet lack the why behind it — the human capacity to question axioms or redefine truth.

Kant’s noumenal realm — the “thing-in-itself” beyond empirical reach — hints at a deeper limitation: machines cannot grasp the unmeasurable meaning humans assign to the world.

Implications

• Ethical Risks: Systems that cannot doubt or reflect may enforce biases embedded in their training data, mistaking correlation (statistical patterns) for causation (human understanding).

• Epistemic Arrogance: Framing intelligence as computation ignores Kant’s insight: reality is interpreted, not processed. A machine’s “answer” lacks the moral and existential weight of a human’s response.

• Cultural Shift: Celebrating AI as “intelligent” risks eroding the value of human creativity — rooted in Descartes’ mind-body duality and Kant’s transcendental idealism.

AI’s Epistemic Shortfall

• Neural Networks: A metaphor, not a replication. Biological neurons operate within a conscious being; artificial “neurons” are mathematical abstractions with no capacity for qualia or intentionality.

• Machine Learning: A statistical process, not “learning” as humans experience it. It identifies patterns but cannot question their meaning or choose to reject them (as Descartes did with sensory data).

• Artificial Intelligence: An oxymoron. Intelligence, per its philosophical foundations, is embodied and teleological — it serves a conscious being’s goals. Machines have no goals, only programmed directives.

Aristotelian Insight

• No Telos, No Truth: Aristotle saw intelligence as inseparable from purpose. An LLM generates text, but it does not seek truth or wisdom; it has no stake in why it exists. AI’s “goals” are imposed externally, reducing intelligence to a tool, not a way of being.

• Soulless Computation: Neural networks borrow biological metaphors but lack Aristotle’s psyche — the holistic integration of body, mind, and purpose. A machine’s “intelligence” is inert matter arranged by humans, incapable of the unity or intentionality of a living soul.

• Moral Blindness: Aristotelian phronesis requires weighing actions against the “good life.” AI systems optimize for metrics (e.g., accuracy) but cannot judge whether those metrics align with human flourishing.

Implications for Public Discourse

• Semantic Deception: Using terms like “intelligence” for pattern-matching algorithms risks normalizing a stripped-down, mechanistic view of human cognition.

• Epistemic Overreach: Claims of “AI consciousness” ignore Kant’s distinction between phenomena (observable data) and noumena (the conscious self that interprets it). Machines process phenomena; they cannot be noumenal agents.

• Cultural Consequences: Accepting “artificial intelligence” as legitimate cedes ground to a worldview where consciousness is optional — a slippery slope toward devaluing human creativity, ethics, and existential reflection.

Implications: Cutting Through the Hype

• Ethical Risks: Systems labelled “intelligent” gain undue trust. Example: Medical AI might optimize for statistical outcomes while ignoring a patient’s lived experience of pain (a qualia that machines cannot comprehend).

• Epistemic Theft: Framing machines as “intelligent” erodes the language to describe human uniqueness. Creativity, doubt, and moral reasoning become “algorithms” — a theft of meaning.

• Cultural Surrender: Accepting AI’s narrative cedes authority to a paradigm that cannot explain why we care about truth, beauty, or justice — only how to replicate their shadows.

Conclusions

To call machines “intelligent” is to empty the term of its history. Descartes’ doubt to Kant’s categorical imperatives, intelligence has always implied a who, not a what. Until AI confronts this philosophical bedrock — acknowledging that consciousness is not an add-on but the essence — it will remain a useful fiction: artificial, yes; intelligent, no.

The AI debate is a mirror: it reflects our willingness to trade philosophical rigour for top-down mechanistic convenience — intellectual laziness. Intellectual virtue, doubt, and the mind’s structuring of reality are all foundational to authentic thinking and reasoning. These are not metaphors but arguments against the very possibility of “artificial” intelligence.

Until we demand that AI’s proponents defend their terms — not just their tech — the hype will persist, and the fallacy will deepen — distorting our very humanity.

Aristotle, Descartes, and Kant all converge on a truth AI cannot compute intelligence: that it is not a tool but a condition of being human. It requires a soul that doubts (Descartes), interprets (Kant), and strives toward purpose (Aristotle).

Machines, devoid of consciousness and telos, are a modern-day Icarus — flying high on mechanistic hubris, destined to fall when confronted with the irreducible depth of human experience. AI is a system that answers but never wonders — we humans do not think because we ‘process data,’ we think because we are alive.