Artificial intelligence may predict our behavior, but it cannot comprehend the underground that makes us human.
The first time I taught a class on artificial intelligence, I expected questions about neural networks, algorithms, machine learning, optimization, or perhaps the standard science-fiction fear: robots taking over the world.
Then a student mentioned Notes from Underground, by Dostoevsky.
At that moment, the remark lodged itself in my mind like a splinter. I already knew the book, but I confess: only after that provocation did I truly read and reread it. And the more I returned to the text, the more it unsettled me. Not because Dostoevsky was too distant from our age of digital abundance and predictive models, but because he was far too close. Close to AI. Close to academia. Close to me.
Those who follow my writing may have noticed that I rarely write about artificial intelligence merely as technology. AI has become, for me, a way of investigating much older questions: consciousness, freedom, power, creativity, morality, and, above all, what it means to remain human when we begin building machines that reproduce abilities we once considered exclusively our own.
Thinking about this is not an innocent hobby. It is a sophisticated form of insomnia. The mind becomes a laboratory, a courtroom, and a bunker all at once. And the underground man lives exactly like that.
The Anatomy of the Underground
To understand the damage this character does to technological optimism, we must return to the work Dostoevsky published in 1864. Notes from Underground is divided into two parts: first, a corrosive philosophical monologue; then, memories from youth that reveal the formation of that wounded consciousness.
The underground man is often described as a failure, a resentful man, or a misanthrope. I find that reading incomplete. To me, he is one of the most honest characters in literature. Not because he has answers, nor because he is admirable, but because he refuses to lie about human complexity.
He dismantles an illusion that remains alive today: the idea that we are perfectly rational, predictable, and explainable creatures.
The underground man does not simply inhabit the world; he inhabits the analysis of the world. Before acting, he has already suspected his own intention. Before loving, he has already found a trap in love. Before accepting any form of happiness, he has already detected in it the possibility of humiliation.
His illness is not ignorance. It is an excess of lucidity.
Dostoevsky wrote this book as a brutal response to the scientism and utilitarianism of his time, especially to the belief that reason could perfectly reorganize society. It would be enough to know humanity’s true interests in order to predict choices, correct behavior, and build an efficient, harmonious, transparent civilization.
The underground man appears to spit on that Crystal Palace.
Not because he disagrees with mathematics. Two plus two is still four. What he rejects is the attempt to turn that certainty into a universal model of human experience. One thing is the logical necessity of arithmetic. Quite another is to imagine that inner life can obey the same clarity.
For the rationalists of his time, if reason showed the best path, everyone would naturally follow it. Dostoevsky considered this idea profoundly inhuman.
The underground man insists that something within us resists even what we know to be best for us. Not out of stupidity, but because reducing the human being to a calculation of interests turns freedom into mechanism. If all our decisions can be predicted, very little remains of what we call will.
It is precisely this refusal to accept that human existence can be fully explained by models, tables, or equations that makes Notes from Underground so surprisingly contemporary.
The Algorithm and the Digital Crystal Palace
Sometimes I have the impression that artificial intelligence represents the contemporary version of the Crystal Palace imagined by Dostoevsky. Not because AI is evil. Not because we should reject it. But because it makes seductive the fantasy that everything can be modeled, predicted, classified, and optimized.
AI loves patterns. It needs them.
It turns words into vectors, behavior into data, desire into probability, and preference into signal.
The algorithm looks at us and says:
“You usually want this. People like you tend to choose that.”
AI does not need to believe that we are rational. It is enough that we are predictable often enough. The more we repeat patterns, the more it learns. The more we become habit, the more it understands us. And perhaps this is precisely where Dostoevsky would begin to smile ironically.
Because the underground man would answer from the depths of his cellar:
“People like me may choose the worst possible option just to ruin your prediction.”
That is the brutal point of Notes from Underground. The human being is not merely poorly calibrated rationality waiting for an algorithmic patch. We are also pride, vanity, self-sabotage, the desire for recognition, and the need to preserve an intimate zone of freedom, even when that freedom manifests itself as error.
AI handles statistical error very well.
But it does not know what to do with the deliberate refusal to be reduced to a pattern. Even less with something more dangerous: bias intentionally inserted into the dataset itself.
Because the problem is not only that the algorithm may be wrong. The problem is that the algorithm may learn exactly what someone wanted it to learn. If the data already arrive contaminated by interests, manipulation, omissions, and calculated distortions, the machine does not reveal the truth: it merely automates the lie with the appearance of precision.
At that point, the underground man becomes even more contemporary. He is not only the noise that escapes the model. He is also the uncomfortable reminder that every model is born within a human struggle over meaning, power, and control.
Nietzsche and the Abyss
Those who spend their lives thinking, teaching, writing, and trying to build bridges between science, culture, technology, and human destiny eventually discover that this work has less to do with engineering than with philosophy. Consciousness is not only light. Sometimes it illuminates. Sometimes it burns.
To think deeply about the future, about the machines we create, and about the kind of humanity we are becoming can be fascinating. But it also has something underground about it.
You descend. At first, you believe you will find answers. Then you realize that each answer opens new questions. And when you least expect it, you are already looking into regions of the human condition that might have been more comfortable if they had remained in the dark.
Nietzsche wrote that when we look too long into the abyss, the abyss also looks back into us. Perhaps Nietzsche and Dostoevsky are describing opposite movements. Nietzsche looks into the abyss until he realizes that it returns the gaze. Dostoevsky decides to go down there and speak with whoever lives in that place.
It is precisely from this point that Notes from Underground seems to write: not from the illuminated surface of reason, but from that place where consciousness ceases to be merely a tool for understanding the world and begins also to reveal the shadows we carry within us.
Faust and the Artificial Infinite
This is where Russian literature crosses paths with another absolute masterpiece: Goethe’s Faust.
If the underground man is the figure who descends through excessive inner excavation, Faust is the man who wants to explode every external limit. He wants to know everything, experience everything, dominate nature, and touch the absolute. He cannot tolerate the mediocrity of the given world. He wants infinity, even if he must bargain with Mephistopheles to obtain it.
I also recognize myself in that tension.
There is something deeply Faustian in all research, in all writing, and in every intellectual passion. We want to go further. To understand more. To open doors. To force limits. Artificial intelligence is born from this impulse. It is one of the grandest expressions of the human will to expand the mind, accelerate discovery, and perhaps reorganize the very horizon of civilization.
But Dostoevsky reminds us of the other side.
It is not enough to expand intelligence if we do not understand the contradictory soul that will use it. It is not enough to build more capable machines if we still fail to understand our own abysses. AI may be Faustian in technique, but the human being remains underground in consciousness.
Perhaps the twenty-first century lives between these two characters.
Faust builds artificial intelligence.
The underground man is the one who will use it.
The great problem of AI is not that the machine may begin to think. The real problem is that it learns from us.
We are a chaotic training set. We are capable of composing Bach, formulating relativity, and building space telescopes — and, on the same day, falling into collective hysteria, moral lynchings, digital tribalism, and petty little acts of revenge.
The algorithm learns the pattern.
But what happens when the pattern we provide is our own shadow?
Notes from Underground should be required reading in every software engineering or data science course. Not to teach code, but to destroy the illusion that understanding the human being is merely a matter of collecting enough data.
A predictive system may anticipate which link I will click without understanding why that link wounds me. It may imitate my writing without carrying my memories. It may formulate elegant answers about ethics without ever having felt the weight of freedom of choice.
The technological horror we face is not that AI may become human.
It is that the human being may accept becoming a model.
It is that we may submit so completely to the optimized recommendations of screens that we begin to believe we are nothing more than that statistical, predictable, domesticated version of ourselves that the algorithm returns to us.
The Mirror in the Dark
Dostoevsky protects us from this delusion. The underground man is uncomfortable, wounded, and contradictory, but he reminds us that there is a region within us that resists the spreadsheet, the dashboard, the metric, and the cost function. A dark part, yes, but also desperately alive.
Between the Faustian ambition of our technology and the underground reality of our soul, we cross the twenty-first century.
We build brilliant machines in the hope that they will organize an existence we have never managed to organize by ourselves. But perhaps AI is only the most powerful mirror we have ever created.
And mirrors do not correct who we are.
They merely make it impossible to pretend we have not seen our own reflection.
Faust built the mirror.
Dostoevsky forced us to look into it.
Artificial intelligence merely increased the resolution.
Editorial transparency note: This article, as with all articles published on this site, was conceived, directed, written, and reviewed by Prof. Maurício Veloso Brant Pinheiro. Artificial intelligence was used as an assistant for editorial refinement, formatting, image generation, SEO metadata, and publication workflow.
Copyright 2026 AI-Talks.org
