AIdiot generation? Why Artificial Intelligence Isn’t the Problem — We Are by Vincenzo Amoruso - HTML preview

Introduction

Welcome to the age of the AIdiot.

Thepromiseofartificialintelligencewasframedasarevo-lution—yet its reality increasingly resembles a degeneration.

We were told a new era was coming. Not a simple technological upgrade, but a genuine rebirth of humanity. During the years when AI exploded into our collective imagination, dominant narratives painted a radiant future: machines capable of freeing us from repetitive labor, algorithms able to diagnose diseases before symptoms appeared, digital assistants ready to democratize knowledge and amplify human creativity. It was the promise of a Revolution: AI as a cognitive partner, a driver of abundance, the key to unlocking the unexpressed 1

2

potential of our species.

And yet, looking around today, that revolution feels less like an ascent and more like a surrender.

Instead of elevating us, AI risks flattening us. Rather than freeing our time, it has filled it with an incessant algorithmic background noise. Instead of enhancing our thinking, it seduces us—through convenience—into delegating it. Welcome to the AIdiot Generation.

This is not a generational insult; it is a social diagnosis of our era.

The degeneration we face is not a loud or catastrophic collapse, like the ones described in dystopian fiction. It is a silent, progressive, almost invisible atrophy. It is the generation that trades the depth of understanding for the speed of a one-second generated answer.

It is the generation that accepts having its emails written, its images generated, its books summarized, and ultimately its opinions shaped by a statistical model trained on the data of the past.

The paradox is unprecedented: while machines become exponentially more “intelligent,” we voluntarily accept becoming functionally more dependent—and, in many cases, cognitively weaker.

We have traded the effort of doubt—the engine of intellectual growth—for the sterile certainty of a probabilistic output. The pact promised the enhancement of our abilities; the reality, if this trajec-tory continues, will be replacement. The agreement promised the expansion of the mind; the reality is its outsourcing.

This book is not a Luddite manifesto, nor a nostalgic lament for an analog past that will not return. The goal of AIdiot Generation is far more urgent: it is both a warning and an invitation to wake up.

In the pages that follow, we will dismantle the narrative of the

“magic solution” sold by major tech corporations. We will explore

3

how cognitive dependency is eroding critical thinking, how the ho-mogenization of AI-generated content is impoverishing our cultural ecosystem, and how, step by step, we are building a golden cage of convenience from which it will become increasingly difficult to escape.

But there is a way out.

Recognizing the degeneration is the first step toward reversing it. To stop being passive users—human terminals of a system that prefers us docile and predictable—and return to being active thinkers: imperfect, chaotic, and wonderfully human.

Degeneration is a choice. And like all choices, it can be changed.

4

Enhanced Humans, Switched-Off Brains Thebetrayaldidn’tarrivethroughatechnologicalcoupora dramatic system failure. It arrived quietly—through the soft, seductive whisper of convenience.

At the beginning, the narrative was flawless: Artificial intelligence as a “cognitive exoskeleton.” A bicycle for the mind. A tool meant to amplify our abilities, not replace them. AI was marketed as the copilot: we would keep our hands on the wheel, choosing the direction, while the machine handled the tedious mechanics of the journey. It was supposed to free us from repetitive work and return to us the sacred time of creativity and strategic reflection.

But the breaking point was subtle, almost invisible: the removal of friction.

Human thought is not smooth. It is a slow, stubborn, uncom-fortable process. It requires cognitive friction: the frustration of a blank page, the doubt that forces you to reformulate an idea, the effort of synthesizing complex concepts, the dead time needed for an intuition to settle. Friction is where learning happens. Friction is where originality is born.

AI offered us a toxic alternative: instant efficiency.

Why wrestle with a thought when a prompt can spit out a “co-herent” paragraph in three seconds? Why endure the discomfort of uncertainty when an algorithm can deliver a “plausible” answer immediately? We began delegating marginal tasks, then complex ones, and finally—almost without noticing—we delegated the process of thinking itself.

The most revealing analogy is the GPS navigator. When GPS

5

became ubiquitous, we stopped reading maps. We stopped developing spatial intuition, stopped noticing landmarks, stopped understanding the logic of a city. Today, if the GPS signal drops, we get lost in a neighborhood we’ve known for years. We outsourced our ability to navigate physical space.

With AI, we have done the same with the space of ideas. We have outsourced our ability to navigate thought.

The AIdiot Generation doesn’t simply use AI to speed up a task; it uses AI to avoid the task altogether. The student no longer reads the text to form a critique; they ask AI to summarize it and generate three “balanced” viewpoints. The professional no longer develops a strategy; they ask the model to “optimize” a document, accepting without question the standardized, soulless corporate language that results.

In this transition, a silent mutation has occurred: we are no longer pilots using a tool. We have become passive terminals of a statistical average.

We accept the first draft generated by the model as if it were truth, because it frees us from the weight of intellectual responsibility. We have traded the effort of discovery for the illusion of knowledge.

Turning off the brain does not mean becoming biologically incapable of thinking; it means choosing not to. It means preferring the smoothness of prepackaged content over the resistance required to build a thought of one’s own. And while we celebrate the seconds saved, we fail to notice that we are hollowing out our minds, transforming them into a hall of mirrors where the only thing that bounces around is the synthetic echo of what we have already said—reprocessed by a language model that understands nothing, yet perfectly pretends

6

to.

7

The Real Problem Isn’t AI — It’s the Idiotic Way We Use It

Let’s stop demonizing the algorithm. It’s the first, necessary act of intellectual honesty. Artificial intelligence, at its core, is innocent. It has no consciousness, no hidden agenda, no contempt for humanity, and no plan to replace us. It is, and remains, a mathematical construct: an extraordinarily complex statistical model trained to predict the most probable word, pixel, or line of code in a given sequence.

It’s a tool. Like a hammer, like a car, like the early internet.

A hammer can build a house or smash a thumb. A car can connect two cities or cause an accident if the driver closes their eyes. AI can amplify an artist’s creativity or generate millions of pieces of digital junk in a single second.

The difference isn’t in the code. It’s in the hand that wields it—and even more, in the mind that guides it.

The real problem, the beating heart of the AIdiot Generation, is that we’ve taken one of the most powerful tools ever created and turned it into a crutch for our cognitive laziness instead of a telescope for our curiosity.

AI is a mirror. If we look into it and see a generation regressing intellectually, it’s because it’s reflecting our willingness to abdicate.

We’ve chosen the path of least resistance.

What does “idiotic use” actually mean?

It doesn’t mean using AI to automate a boring spreadsheet or translate a technical manual. That’s smart use.

Idiotic use is treating a probabilistic text generator as if it were an

8

oracle of absolute truth. It’s using a chatbot to write an apology to a friend, draining the words of any real empathy and replacing them with algorithmic coldness. It’s asking software to summarize a novel we were supposed to read to shape our character, trading the depth of understanding for a sterile bullet list. It’s presenting as “our own”

a thought we never struggled to form, lying to ourselves before lying to anyone else.

Big tech companies, of course, share part of the blame. They designed these systems to be frictionless, exploiting our cognitive biases and our hunger for instant dopamine. They sold us the illusion that speed equals quality and that delegation equals efficiency.

But at the end of the chain, the hand typing the prompt is ours.

The choice to hit “Send” without reading, without verifying, without questioning the hallucination or the hidden bias in the output—that’s a human choice. A choice of convenience that becomes intellectual negligence.

The idiocy isn’t in the technology. It’s in the surrender.

It’s the conscious decision to lower our standards because “for what I need, this is good enough.”

The ultimate danger isn’t AI becoming too smart. It’s us becoming just stupid enough not to notice the difference.

We’re training ourselves into indifference, shaping our biological neural networks to expect only pre-chewed answers, losing the ability to chew complex thought on our own.

Recognizing this paradox is, in fact, liberating.

If the problem were AI itself, we’d be doomed—hostages to an unstoppable technological destiny. But if the problem is how we use it, then the solution remains in our hands.

9

We can choose to change course. We can decide to use AI not to avoid thinking, but to challenge it; not to get easy answers, but to craft better questions; not to conform to a statistical average, but to explore the blind spots of our imagination.

The machine is ready. The question is: are we?

Or will we keep using a jet engine to push a handcart, and then wonder why we’re not getting anywhere?

10

Part I

The New Digital

Addictions

11

Chapter 2

Infinite Scroll: The Hamster Wheel 2.0

Reels and Shorts: The Brainrot Factory IfAIisthetoolhelpingusforget how tothink,theninfinite-scroll video is the mechanism preventing us from thinking long enough to form a single complex idea.

Welcome to the brainrot factory.

This isn’t a metaphor. TikTok, Instagram Reels, YouTube Shorts—these aren’t harmless entertainment apps. They are cognitive war machines engineered with one goal: to fracture your attention until it collapses entirely.

The mechanism is brutally simple. Fifteen seconds. Thirty at most. Less time than it takes to read this sentence.

In that microscopic window, an AI system has analyzed thousands of variables—your watch time, your past likes, the time of day, the speed of your scroll—to deliver precisely what will keep your finger 13

14

glued to the screen for another fifteen seconds.1

There is no narrative, no development, no arc.

Only stimulus–response, stimulus–response. A Pavlovian loop industrialized at planetary scale.

Your brain prioritizes fast processing over deep processing; it reacts to signals. And every time you scroll, every time you chase the next clip, you’re training the algorithm to know you better than you know yourself—and training yourself to no longer tolerate boredom, waiting, or depth.2

The term brainrot, born as internet slang for low-quality content, is becoming a clinical description of our cognitive decay. It’s not about “stupid videos.” It’s about the form.

Even the smartest content, compressed into fifteen seconds and buried under an avalanche of distractions, loses its ability to be understood, processed, remembered. The form becomes the message, and the message is: “Don’t think. Scroll. Consume. Repeat.” 3

Neuroscience is beginning to confirm what we already sensed: prolonged consumption of rapid-scroll content reduces sustained attention, weakens long-term memory, and increases FOMO-driven anxiety.4

Our brains, shaped by millions of years of evolution to seek novelty and patterns, are being bombarded with artificial novelty every fifteen seconds. It’s like feeding a lion refined sugar instead of meat: it excites, 1J. Zittrain, The Future of the Internet–And How to Stop It, Yale University Press, 2008, pp. 143–167.

2A. Alter, Irresistible, Penguin Press, 2017, pp. 23–45.

3M. McLuhan, Understanding Media, McGraw-Hill, 1964, p. 7.

4G. Gazzaley and L.D. Rosen, The Distracted Mind, MIT Press, 2016, pp. 89–112.

15

it intoxicates, but it doesn’t nourish.

And here AI enters—not as a visible protagonist, but as the invisible director. Machine-learning recommendation systems decide what you’ll see next, optimizing not for your education or well-being, but for your time on platform.

Every scroll is a bet: “Tap now—you might find something fun.

Tap again. And again.” It’s a digital casino, and the currency isn’t money. It’s your attention, your time, your remaining cognitive bandwidth.

The AIdiot Generation isn’t just the one using AI to write homework. It’s the one that can’t read a 2,000-word article without feeling the physical urge to grab the phone. It’s the one that watches a movie and reaches for the smartphone during slow scenes to “fill the void.” It’s the one that has forgotten how to be alone with its own thoughts without feeling guilty for “lost productivity.”

Brainrot isn’t a side effect. It’s becoming the final product.

A fragmented, anxious brain—dependent on external validation and incapable of deep thought—is the perfect consumer: it will never question the system, never have the patience to build alternatives, and always settle for the next hit of digital dopamine.5

Escaping the brainrot factory requires a conscious act of rebellion.

It means relearning boredom. It means watching a video all the way through—or better yet, turning off the screen and staring at nothing, letting the mind wander, connect, create. It means treating your attention not as an infinite resource to squander, but as the most precious and limited asset you possess.

5S. Zuboff, The Age of Surveillance Capitalism, PublicAffairs, 2019, pp.

79–104.

16

Because in the end, what we’re losing isn’t just the ability to focus.

We’re losing the ability to be human in the deepest sense: creatures who reflect, who doubt, who build complex meaning over time.

We’re trading our inner life for entertainment. And AI, silently, is taking note of every scroll, perfecting the golden cage we’re voluntarily locking ourselves into.

17

The Trap: The Dopamine Loop

Tounderstandwhywe’restuck,wefirstneedtodismantleone of the most persistent myths in pop neurobiology: dopamine is not the molecule of pleasure. It’s the molecule of desire—of seeking, of anticipation, of chasing the next hit.

Pleasure is what you feel when you eat chocolate; dopamine is what gets you off the couch to go find it.

It’s the neurotransmitter of survival, evolved to push us toward scarce resources in a hostile environment: food, water, social connection, information. The problem is that our Paleolithic brain never evolved defenses against an environment where these resources are infinite, free, and hyper-palatable.6

And this is where the “dopamine loop” comes in—a mechanism digital platforms have perfected with the precision of a Swiss watch-maker.

At the heart of this system is what behavioral psychologist B.F.

Skinner called the variable ratio schedule of reinforcement.7

Think of a slot machine.

If you win every time you pull the lever, you get bored. If you never win, you stop playing. But if you win unpredictably—once yes, three times no, twice yes—your brain goes feral. You don’t know when the reward will come, so you keep pulling the lever.

In the digital world, the lever is the pull-to-refresh gesture or the upward swipe. The reward is a funny video, a notification, a like, or 6A. Lembke, Dopamine Nation, Dutton, 2021, pp. 34–58.

7B.F. Skinner, Science and Human Behavior, Macmillan, 1953, pp.

394–398.

18

a piece of information that scratches your curiosity.

Until a few years ago, this mechanism relied on crude randomness.

Today, with AI, nothing is random anymore.

The algorithm isn’t a distracted casino dealer; it’s a casino that knows you intimately.

It has mapped your vulnerable hours, your insecurities, your micro-interests. It knows exactly which type of content has the highest probability of triggering your reward circuit in that precise millisecond.8

Why can’t we stop?

Because we’re fighting an asymmetric biological battle.

Every time we give in to the scroll, the brain releases a spike of dopamine. But biology always seeks homeostasis. To compensate for these artificial, unnatural spikes, the brain reduces the number of dopamine receptors available. This phenomenon is called downregu-lation.9

The result is tolerance.

What seemed fun or interesting a year ago now leaves us indifferent.

We need increasingly extreme, rapid, shocking stimuli to feel even a tiny spark of engagement.

And paradoxically, when we’re not scrolling, dopamine levels crash below baseline, generating a sensation of emptiness, anxiety, and intolerable boredom.

We don’t scroll for pleasure anymore; we scroll to avoid the pain of withdrawal.

8N.D. Schüll, Addiction by Design, Princeton University Press, 2012, pp.

112–130.

9N.D. Volkow et al., “Neurobiologic Advances from the Brain Disease Model of Addiction”, NEjm, 2016, pp. 363–371.

19

This is the trap of the AIdiot Generation.

We’re not “weak-willed.” We’re biologically hijacked.

AI has turned our smartphones into syringes of cognitive sugar, administered with milligram-level precision to maximize time on platform.

Recognizing this mechanism is crucial because it frees us from moral guilt and puts us on strategic ground. You cannot defeat a recommendation supercomputer with sheer willpower; willpower is finite, dopamine is biological.

The only way out is not to fight the loop from within, but to break the chain from the outside: introducing friction.

Turn off notifications. Use grayscale screens. Set physical timers.

Leave the phone in another room.

We must stop treating our neurochemistry as a playground for Silicon Valley.

Because until we regain control of our reward system, every promise to “use AI to improve our lives” will remain a lie we tell ourselves while we scroll, dazed, toward the next frame.

20

The Illusion of “Just One Minute”

It’stheuniversalliewetellourselveseveryday,dozensoftimesa day: “I’ll just check one thing,” “I’ll just watch one video,” “I’ll just reply to this message and then get back to work.”

“Just one minute” is the Trojan horse of cognitive degeneration.

It’s the harmless-looking doorway that drops you straight into a bottomless pit.

And the reason we fall for this trap isn’t weak willpower—it’s a deliberate, engineered flaw (or, more cynically, a feature) of the modern digital environment: the elimination of stopping cues.

In the physical world, activities have natural boundaries. A book chapter ends and there’s a blank page. A newspaper has a final column. A conversation ends when someone stands up to leave.

These stopping cues interrupt the flow, forcing a micro-pause and a conscious decision: “Do I want to continue or do something else?”

Digital platforms—especially those powered by AI—have annihi-lated these boundaries.

The invention of infinite scroll removed the friction of loading the next page, creating a continuous, uninterrupted stream of stimuli.10

But AI has taken this concept to its logical extreme: it’s not just infinite; it’s hyper-personalized.

The algorithm doesn’t show random content. It calculates, in real time, the exact stimulus needed to keep your engagement from dropping.

If you’re about to close the app, AI serves the perfect piece of content to keep you there for another sixty seconds. Then another 10A. Raskin, “Infinite Scroll: A Love/Hate Relationship”, 2012.

21

sixty. And another.

The “minute” stretches into an hour, an afternoon, an entire evening.

This phenomenon has a name in cognitive psychology: time blindness.

When we’re immersed in an algorithmic feed, we lose the ability to perceive time passing. Our brain’s temporal receptors—relying on changes of state and task completion—are tricked by the homogeneous, endless nature of digital content. There is no “before” and “after.”

There is only an eternal, fragmented “now.”11

A classic behavioral experiment illustrates this perfectly.

In the famous bottomless bowls study, participants were given soup bowls that secretly refilled from below as they were emptied.

Without the visual cue of an empty bowl—the stopping cue—participants ate 73% more soup than the control group, without noticing or feeling more full.12

Today, your smartphone is a bottomless bowl. And the soup is made of information, entertainment, and social validation. AI ensures the flavor is always exactly what you crave, preventing you from ever feeling cognitively “full.”

For the AIdiot Generation, the illusion of “just one minute” has devastating consequences for our sense of intentionality. Life stops being a series of deliberate choices and becomes a chain of reactions to external stimuli. We lose the ability to manage our time because we’ve outsourced the management of our attention to a device whose only goal is to keep us hooked.

11G. Mark, Attention Span, Knopf, 2023, pp. 55–78.

12B. Wansink et al., “Bottomless Bowls”, Obesity Research, 2005.

22

Breaking this illusion requires an act of sabotage against platform design. We must artificially reintroduce the stopping cues that technology has removed. Set physical timers (not on the same device you’re using). Use apps that block access after a set period. Adopt the rule of “phone in another room” whenever you’re doing something that requires real focus.

We need to stop negotiating with the algorithm. Because every time we say “just one minute,” what we’re really saying is: “Algorithm, take control of my mind. I don’t know for how long.”

And the algorithm, unfailingly, accepts the invitation.

Chapter 3

Notifications: Training the Brain with Micro-Shocks

Operant Conditioning, App Edition Imagineabox.Insideit,ahungrypigeon.Onthepanelofthe box, a lever. If the pigeon pecks it, sometimes a pellet of food drops, sometimes it doesn’t.

If the food drops every time (fixed reinforcement), the pigeon learns quickly—but gets bored, and stops pecking as soon as it’s full.

But if the food drops unpredictably—once yes, three times no, twice yes, ten times no—something extraordinary happens.

The pigeon becomes obsessive. It will peck the lever endlessly, ignoring fatigue, ignoring other stimuli, trapped in a cycle of expectation its brain cannot shut off.

Known as the famous variable ratio schedule of reinforcement, 23

24

discovered by B.F. Skinner in the 1950s.1

Now look at your hand. The one scrolling your smartphone.

We like to believe we’re sophisticated users, critical consumers, masters of technology. The raw truth is that we are Skinner’s pigeons.

And our smartphone is the box.

The pull-to-refresh gesture isn’t a neutral interface. It functions like a digital lever. When we perform it, we don’t know what we’ll get: maybe an important email, maybe a like, maybe a hilarious video, maybe nothing at all.

It’s the unpredictability that keeps us hooked.

Platforms have deliberately removed every stopping cue—like the end of a page or a chapter—to create an infinite stream of attempts, turning app usage into a continuous session of operant conditioning.2

But there is one fundamental—and terrifying—difference between Skinner’s experiment and the reality of 2026.

Skinner’s machine was stupid. It dispensed food randomly.

The AI running our apps is hyper-intelligent.

This is no longer simple randomness. The recommendation algorithm calculates, in real time, the exact “food pellet” for you, at this precise moment of the day, based on your inferred mood, your scroll speed, the time, and your history.

If AI detects you’re about to close the app (the pigeon is about to walk away from the lever), it will throw at you the content with the highest statistical probability of keeping you there for another thirty seconds.

1B.F. Skinner, Science and Human Behavior, Macmillan, 1953, pp.

394–398.

2N.D. Schüll, Addiction by Design, Princeton University Press, 2012, pp.

112–130.

25

It’s variable ratio reinforcement supercharged by machine learning: a casino that knows your psychological profile better than you do.3

The conditioning becomes so deep it starts manifesting even without the stimulus. This is the phantom vibration phenomenon: the brain, so accustomed to expecting a reward, generates the tactile illusion of a notification that never arrived.

We can be conditioned to jump at the sound of a digital bell like Pavlov’s dogs—except with one crucial difference: unlike Pavlov’s dogs, we actively seek the bell, polish it, and beg it to ring.

The AIdiot Generation has voluntarily built its own cage, thrown away the key, and now pecks at the walls waiting for a crumb of algorithmic validation.

Recognizing this parallel isn’t an academic exercise; it’s self-defense.

Because as long as we fail to understand we’re inside a Skinner box, we’ll keep blaming our “lack of willpower” instead of rebelling against a system scientifically engineered to erase it.

The first rule for stopping being a pigeon is realizing you’re in a cage. The second is to stop pecking the lever.

3T. Harris, “How Technology Hijacks People’s Minds”, Center for Humane Technology, 2016.

26

FOMO as a Weapon of Control

FOMO—FearOfMissingOut—isnotapersonalityflawofour era, nor a simple generational insecurity. It’s a system feature. A perfectly calibrated psychological control mechanism designed to keep us in a state of constant alertness and dependency.

Originally, FOMO was a marginal human emotion, tied to specific social events. Today, amplified by artificial intelligence, it has become the invisible engine of the attention economy.

The mechanism relies on a fundamental cognitive distortion: our tendency to compare our own “behind the scenes”—with its boredom, uncertainty, and failures—to everyone else’s curated, filtered “stage.”

But the algorithm doesn’t just show us that stage; it builds it specifically to target our vulnerabilities.

If AI detects that you hesitated on a post about travel, professional success, or a cultural trend, it will begin flooding your feed with similar content. Not because those things are objectively important, but because it knows they are the most effective emotional levers to make you feel inadequate, late, excluded.

The algorithm is not a passive observer; it’s an arsonist methodi-cally blowing on the flames of our insecurity.4

The result is the tyranny of the update.

We check our devices not to find useful information, but to soothe an artificially induced anxiety. We live in what psychologist Linda Stone called continuous partial attention: physically present in one place, mentally scattered across a dozen digital streams, terrified 4A.K. Przybylski et al., “Motivational, Emotional, and Behavioral Correlates of Fear of Missing Out”, Computers in Human Behavior, 2013.

27

that if we look away for a moment, the world will move on without us.5

This is the most subtle and devastating form of control exerted over the AIdiot Generation. There’s no need for chains or coercive surveillance when the subject self-surveils, driven by the fear of missing a meme, a headline, or a social approval.

FOMO steals the present moment—the only time we truly own—projecting us into a hypothetical future and a filtered past. It prevents us from going deep, from being creatively bored, from being alone with our thoughts.

The supreme irony is that in our desperate attempt not to miss anything, we end up missing everything that matters: the ability to be fully present in our own lives.

We’ve traded the depth of experience for the breadth of connection, convincing ourselves that being “connected” is the same as being

“alive.”

The antidote to this weapon of control is not total disconnection—often unrealistic—but the deliberate cultivation of its opposite: JOMO, the Joy Of Missing Out.

JOMO is the conscious act of choosing what to ignore. It’s the reclaiming of the right not to know everything, not to see everything, not to react to everything. It’s the understanding that real power, in the age of AI, doesn’t lie in processing every single stimulus, but in the wisdom to decide what deserves our attention and what can, peacefully, be left behind.

Because ultimately, whoever controls your fear of missing out controls your time. And whoever controls your time owns your life.

5L. Stone, “Continuous Partial Attention”, 2007.

28

Attention: A Resource Going Extinct Wearenotsimply“distracted.”Distractionimpliesthat attention still exists somewhere, ready to be summoned with a bit of willpower. The reality is far darker: we are witnessing the systematic extinction of deep attention in its natural habitat—the human mind.

For millennia, sustained attention has been the engine of civiliza-tion. It is the faculty that allows us to read complex books, truly listen to another person, master an art, or solve intricate problems. It was a state of flow, as psychologist Mihaly Csikszentmihalyi described it—a moment in which consciousness organizes itself and the self merges with the action.6

Today, this resource is no longer cultivated; it is extracted.

Digital platforms and AI models have turned human attention into raw material, subject to a process of cognitive strip-mining. This is no longer the “attention economy,” a term that sounds almost neutral, but a full-blown attention-extraction capitalism.7

The symptom of this extinction is attentional bankruptcy. We can no longer sustain cognitive effort for more than a few minutes without feeling the compulsive urge to seek distraction.

Deep reading—once a process that required building complex mental imagery and connecting abstract ideas—has collapsed into keyword scanning and rapid confirmation seeking.8

AI accelerates this extinction.

6M. Csikszentmihalyi, Flow, Harper & Row, 1990, pp. 53–72.

7T. Wu, The Attention Merchants, Knopf.

8M. Wolf, Reader, Come Home, Harper, 2018, pp. 34–56.

29

Where we once had to work to find information, AI now chews it for us and spits it back in the form of summaries, bullet points, and instant answers. It atrophies the attentional muscle.

By delegating synthesis, research, and even idea formation to AI, we are allowing our attentional circuitry to wither. Why struggle to maintain focus on a 2,000-word text when a plugin can extract the key points in three seconds?

The result is that we lose not only the content, but the very process that makes the human mind resilient and creative.

The consequences of this extinction go far beyond productivity loss.

Deep attention is the prerequisite for empathy. To truly understand another person’s pain, joy, or complexity, we must give them uninterrupted time and mental presence. A fragmented mind—jumping from one notification to the next—is incapable of genuine emotional connection.

We are creating a generation of individuals who are technically hyper-connected but emotionally and intellectually isolated, because they lack the resource required to build meaningful bridges to reality.

Recognizing attention as an endangered resource radically shifts the perspective. This is no longer about productivity hacks or apps that block your phone for five minutes. This is a fight for cognitive sovereignty.

Protecting your ability to concentrate has become the most radical act of rebellion of the 21st century.

We must stop treating our minds as open grazing fields for algorithms. We must begin cultivating attention the way one cultivates a garden: with patience, with clear boundaries, and with the awareness

30

that the most precious things in life require time, silence, and a presence no language model will ever replicate.

Part II

AI-Assisted Idiocy

31

Chapter 4

Copy-Paste from AI: The Death of Original Thought

Students: Submitting Without Reading

Traditionalplagiarism—paradoxicalasitsounds—required at least a minimum amount of cognitive effort. Copying from a book or from Wikipedia meant reading, selecting sentences, mimicking a style, or at least attempting a clumsy paraphrase. There was, however distorted, a process of mental engagement.

Today, even that residue of effort has vanished. We’ve entered the era of blind submission.

The phenomenon is now endemic in universities and high schools: a student types a vague prompt into a large language model, copies the output generated in three seconds, pastes it into a Word document, 33

34

and submits it to the professor.

The most disturbing part isn’t the act itself. It’s the fact that the student will never read that text. Not before submitting it, and certainly not after receiving the grade.

In this scenario, the human being has been downgraded to a biological courier—a “mailman,” a USB stick with legs—whose only job is to transfer a file from a remote server to a teacher’s desk.

The student is not the author of the thought; they are merely its unconscious delivery mechanism.

The consequences are devastating and often grotesque. Professors increasingly report theses and essays filled with blatant hallucinations: citations of books that don’t exist, references to historical events that never happened, or arguments that contradict themselves within two paragraphs.

When asked to explain a key concept from their own paper during an oral exam, these students panic. Not because they’re unprepared in general, but because they are completely foreign to the text that bears their name. They have submitted a simulacrum of knowledge—devoid of any anchor in their own mind.1

The real damage, however, is not the deception of the academic institution. The real damage is self-deception—and the cognitive atrophy that follows.

Cognitive neuroscience and pedagogy agree on one fundamental point: writing is thinking. The act of composing a text—with its hesitations, deletions, search for the right word, and logical structuring 1D.R.E. Cotton et al., “Chatting and cheating: Ensuring academic integrity in the era of ChatGPT”, Innovations in Education and Teaching International, 2024.

35

of arguments—is the mechanism through which the brain consolidates information, builds durable neural connections, and develops critical thinking.2

By delegating writing to AI, the student isn’t “saving time.” They are voluntarily abandoning the only process that could have transformed that information into knowledge. They are skipping cognitive gym.

The result is a generation of graduates who possess a diploma but lack the mental tools to defend the ideas that diploma is supposed to represent.

Here lies the essence of AI-assisted idiocy. It is not a lack of innate intelligence; it is a systematic choice to bypass intelligence.

We have created a system in which it is possible to obtain maximum formal recognition (grades, degrees) with the absolute minimum of intellectual effort.

If we don’t reverse this trend, we are not training the future leadership class. We are training a generation of puppets who, when faced with a real, complex problem that cannot be solved with a prompt, will stare into the void—waiting for someone else to type the solution for them.

2P.A. Mueller and D.M. Oppenheimer, “The Pen Is Mightier Than the Keyboard”, Psychological Science, 2014.

36

Professionals: Delegating Without Understanding

IfAIdelegationcreatesalearningvoidforstudents,itcreatesare-sponsibility void for professionals. And in the real world—made of contracts, medical diagnoses, software systems, and legal rulings—responsibility voids have a precise name: disasters.

We are witnessing the rise of the hollow professional.

Lawyers submitting briefs containing citations of court cases that never existed, generated in an instant by a language model. Doctors confirming diagnoses based on AI-generated summaries of medical records, without ever reading the patient’s raw data. Programmers pasting automatically generated code directly into production systems without understanding its logic or its security vulnerabilities.

The psychological mechanism behind this phenomenon is well known in cognitive ergonomics: automation complacency or algorithm appreciation. Humans have an innate tendency to over-trust automated systems, attributing to them an aura of objectivity and infallibility they do not possess.3

When this complacency meets a Large Language Model that writes with fluid, confident, authoritative prose, critical resistance collapses.

The professional stops being an expert using a tool and becomes a biological rubber stamp. Their role shrinks to reading the output, nodding, and pressing “Send.”

The cruel paradox of intelligent delegation is this: to delegate a 3R. Parasuraman & D.H. Manzey, “Complacency and Bias in Human Use of Automation”, Human Factors, 2010.

37

task effectively to AI, you must possess a level of domain expertise equal to or greater than the system itself—precisely so you can verify the correctness of its work.

If you don’t know how to write a contract clause, you cannot know whether AI wrote it correctly or inserted a logical trap. If you don’t understand cryptography, you cannot evaluate whether the AI-suggested code is secure.

Delegating what you do not understand is not efficiency; it is professional negligence disguised as innovation.4

The consequences are already emerging.

Courts are beginning to sanction lawyers for “frivolous conduct”

when they submit AI-generated documents containing factual hallucinations. Companies are suffering data breaches caused by insecure code copied from programming assistants. The reputational and legal damage is real—but the language model cannot be sued, cannot lose its license, and cannot go to prison.

The only entity responsible is the human who pressed the button.

The professional AIdiot Generation risks normalizing the excuse:

“the system generated it.”

It is the modern, technological version of “I was just following orders.”

But professional competence is not a delegable optional. It is the fiduciary pact between an expert and society.

When a professional delegates their judgment to an algorithm whose limits they do not understand, they are not merely betraying their client or employer. They are signing their own obsolescence.

Because if your only added value is acting as a bridge between a 4E. Mollick, Co-Intelligence, Portfolio, 2024.

38

prompt and a result you cannot evaluate, that bridge can be removed.

And when it collapses, no algorithm will take the blame.

39

Too Many Tools, Too Little Competence

Wearelivinginsidethegreatest—andmostinsidious—paradox in technological history: we have never been so well equipped, and yet we have never been so fundamentally incompetent.

Historically, the evolution of tools followed a clear logic: extending human capabilities. The hammer extended the strength of the arm.

The microscope extended vision. The spreadsheet extended numerical processing.

But mastering these tools required cognitive investment: study, practice, error, and eventually understanding. The tool was a means; the human remained the end and the director.

Today, that dynamic has inverted.

Modern tools—especially those powered by artificial intelligence—are not designed to be mastered. They are designed to be absorbed. They are black boxes built to hide complexity, not teach it.

The result is a phenomenon cognitive psychologists call the illusion of competence: the tendency to drastically overestimate one’s understanding of a subject simply because we have access to a tool that handles it for us.5

Take data analysis.

Once, an analyst needed to understand statistics, data cleaning, and algorithmic logic to reach a conclusion. Today, anyone can upload 5M. Fisher et al., “Searching for Explanations: How the Internet Inflates Estimates of Internal Knowledge”, Journal of Experimental Psychology: General.

40

a file into an AI system and ask: “Make me a trend chart and tell me what it means.”

AI returns a perfect graph and an eloquent analysis in three seconds. The user feels like a data genius.

But if the underlying data is distorted (garbage in), or if the algorithm applied a spurious correlation instead of a causal relationship (garbage out), the user has no way to notice. They lack the foundations to validate the roof AI has built.

This creates systemic fragility.

Real competence is built through the friction of learning: the time spent wrestling with a problem, understanding why code won’t compile, reformulating a sentence because it doesn’t sound right. It is in that friction that the neural connections of true expertise form.

By removing friction, AI removes learning.

We are creating a generation of button-pushers who know how to obtain a result but have no idea how that result was generated.

Philosopher Hubert Dreyfus, in his work on skill acquisition, demonstrated that you cannot skip the fundamental stages of learning—from novice to expert—without losing the ability to handle unexpected or ambiguous situations.6

The paradox closes into a perfect trap: the more powerful our digital assistants become, the narrower our understanding of the world becomes.

When the tool works, we feel omnipotent. But when the tool fails, freezes, or produces a convincing hallucination, we are no longer able to intervene, correct, or even understand what went wrong.

We are like airline pilots who have spent so much time in autopilot 6H.L. Dreyfus & S.E. Dreyfus, Mind Over Machine, Free Press.

41

mode that they’ve forgotten how to land manually when the engines fail.

The AIdiot Generation is not defined by a lack of access to technology, but by the loss of the ability to govern it.

We have traded mastery for convenience, and now we find ourselves surrounded by divine tools wielded by hands that have forgotten how they work.

42

Chapter 5

Caveman Prompting

The 2,000-Word Essay

Itisthearchetypeofcavemanprompting:adigitalgruntthrown into the abyss of a chat window, followed by the silent void of processing and, finally, the ultimate act of pure negligence.

The sequence is always the same—ritualistic, predictable, and tragic.

The user types: “Write me a 2,000-word essay on the impact of AI on society.” (or some equally vague, context-free variant).

The AI, obedient and verbose, produces a wall of text: grammatically flawless, structurally tidy, and conceptually hollow.

The user selects all, copies, pastes into the final document, and sends.

Total interaction time: 45 seconds. Total reading time: zero.

This behavior isn’t mere laziness; it’s a well-defined cognitive phenomenon known as the AI cargo cult. Just like the Pacific 43

44

Island tribes who built fake wooden airports hoping the cargo planes would return, the modern user performs the outward motions of intellectual labor—opening the laptop, typing a request, receiving a document—without any understanding of the underlying process.1

Total delegation without verification is a leap into darkness.

Those 2,000 words may contain factual hallucinations delivered with absolute confidence, circular reasoning, cultural biases inherited from training data, or a tone wildly inappropriate for the context (e.g., overly formal prose for a casual blog, or vice versa).

But the user doesn’t care. They obtained the simulacrum of a finished product. They satisfied the formal requirement while emptying it of all substance.

The absurdity of caveman prompting lies precisely here: we have access to a tool that is, by design, conversational and iterative. A Large Language Model performs best when guided, corrected, challenged, and refined through dialogue.

Instead, we treat it like a vending machine: insert a coin (a terrible prompt) and expect the perfect soda (a flawless text) to drop out.

This practice carries a hidden but devastating cost: the atrophy of critical judgment.

When you don’t read what you sign, publish, or submit, you stop being the author and become a mere conduit for algorithmic noise. You expose yourself to enormous academic, professional, and reputational risks, because in case of error, the excuse “the AI wrote it” has no legal or ethical value.

Responsibility always belongs to the person who pressed “Send.”

1E.M. Bender et al., “On the Dangers of Stochastic Parrots”, FAccT ’21, 2021.

45

The next time you feel tempted to ask a machine to write for you, remember a golden rule of the digital age: if you don’t have the competence, time, or willingness to read and correct what AI produces, you shouldn’t ask it to produce anything at all.

Using AI without reading its output isn’t “working smarter.” It’s signing a blank cognitive surrender form.

46

Fake Experts: Prepackaged Answers Thereisabriefbutpowerfulmomentofcognitiveeuphoria that follows the instant an AI generates a response. The user sees a neatly structured text—paragraphs, bullet points, an authoritative tone, polished vocabulary. In that moment, a mental short-circuit occurs: the brain confuses the fluidity of form with the depth of understanding.

It is the illusion of having become an expert simply because we’re holding the map—without ever having walked the territory.

In cognitive psychology, this phenomenon is a modern variant of what Leonid Rozenblit and Frank Keil called the Illusion of Explanatory Depth. Humans systematically overestimate how well they understand how things work—until they are forced to explain them step by step.2

Artificial intelligence exponentially amplifies this illusion.

By giving us a “ready-made” explanation, AI spares us the effort of constructing one. But it is precisely in that effort—in searching for connections, handling exceptions, wrestling with doubt—that true competence is forged.

An expert is not someone who possesses the answer, but someone who understands why that answer is correct, what its limits are, and what would happen if the underlying variables changed.

The average user, armed with their caveman prompt, receives only the final product—stripped of the epistemic process that produced it.

The result is a simulacrum of competence: an individual who 2L. Rozenblit & F.C. Keil, “The Misunderstood Limits of Folk Science”, Cognitive Science, 2002.

47

feels entitled to discuss law, medicine, programming, or economics with arrogant confidence, based entirely on the output of a statistical model that, by its very nature, does not “understand” anything it says.3

This produces a hypertrophic version of the Dunning–Kruger effect.4

Ignorance is no longer silent; it is loud, articulate, and backed by a reassuring user interface.

The danger peaks when these “prepackaged answers” are used to make high-impact decisions. A manager approving a strategy based on an AI-generated market analysis without understanding its assumptions, or a citizen forming a radical political opinion based on partial summaries of news, are victims of this illusion.

We have turned knowledge into a fast-consumption product—cognitive junk food that momentarily satisfies curiosity but leaves the mind malnourished.

True expertise requires humility: the awareness of how vast the unknown really is.

AI, with its verbose and confidently simulated omniscience, is eroding that humility and replacing it with blind, dangerous confidence.

Being an expert does not mean having the answer at your fin-gertips. It means having the mental tools to evaluate whether that answer—whoever or whatever produced it—deserves to be believed.

And that is a muscle which, if not exercised through effort, atrophies irreversibly.

3S. Sloman & P. Fernbach, The Knowledge Illusion, Riverhead Books, 2017.

4J. Kruger & D. Dunning, “Unskilled and Unaware of It”, Journal of Personality and Social Psychology, 1999.

48

Knowing Everything, Understanding Nothing

We are not becoming ignorant in the traditional sense.

Old ignorance was a void—an absence of information.

Today’s ignorance is far more insidious: it is full ignorance. It is the condition of knowing everything, but only on the surface.

Thanks to artificial intelligence, we have democratized access not to knowledge, but to its meta-description.

We can ask a language model to summarize a 19th-century philo-sophical treatise, extract the key points from a hundred-page financial report, or explain quantum mechanics in three paragraphs. We receive an immediate, structured, seemingly comprehensive answer.

The problem is that we consume the container of knowledge, not knowledge itself.

We are accumulating meta-knowledge: we know what a book is about, but we haven’t read the book; we know the conclusion of an argument, but we’ve skipped the logical steps that make it valid.

This culture of TL;DR—Too Long; Didn’t Read—has become a cognitive operating system, elevating what was once considered an intellectual sin into a standard practice.

Years ago, cognitive neuroscientist Nicholas Carr warned that the internet was reshaping our brains to favor skimming over deep reading.5

With AI, this process has been exponentially accelerated.

5N. Carr, The Shallows, 2010.

49

We don’t even need to skim anymore: we are served the distilled essence, stripped of context, nuance, and the digressions that often contain the true engine of understanding.

The result is a generation of illiterate encyclopedias: individuals capable of dropping the right term, the trending concept, or the seemingly cultured quote into a conversation—yet collapsing at the first attempt to go deeper.

Superficial knowledge is fragile.

It lacks the roots needed to

withstand the winds of doubt or counterargument.

Someone who knows only the summary of a historical event cannot understand its structural causes. Someone who reads only the abstract of a scientific paper cannot evaluate its methodology or error margins.

Philosopher Byung-Chul Han describes this condition as the loss of “the time of duration” in favor of the “point-time” of atomized information.6

AI is the perfect architect of this point-time.

It tricks us into believing we’ve saved time, when in reality it has stolen the time required for ideas to settle.

Knowing everything superficially is the most dangerous form of ignorance, because it comes with intellectual arrogance.

Those who know they don’t know are open to learning. Those who believe they know—because they read a summary generated in two seconds—close their minds.

In a complex world, where challenges require systemic thinking and the ability to connect disparate domains, superficiality is not just a cultural limitation: it is an existential risk.

We trade wisdom—which requires time and depth—for information, 6B.-C. Han, La scomparsa dei riti, 2019.

50

which is instant and disposable.

And in the end, we will find ourselves with minds full of data, yet completely empty of meaning.

Chapter 6

AI as an Infallible Oracle

People Who Believe Anything an AI Says

Thereisarecurringphenomenonthatemergeseverytimea language model produces a particularly convincing hallucination: someone, somewhere, believes it.

Not only believes it—they share it, cite it, and use it as the basis for a decision.

It is the triumph of the digital oracle, the new form of technological superstition that is replacing critical thinking with blind faith in “the computer says.”

The psychological mechanism behind this algorithmic credulity is complex and rooted in several cognitive biases that AI exploits with surgical precision.

The first is authority bias. From childhood, we are conditioned 51

52

to trust figures we perceive as expert or institutional. The clean interface of a chatbot, its measured tone, and its simulated omniscience unconsciously activate the same neural circuits. We are no longer interacting with a tool; we are consulting an entity that appears to know everything.1

The second mechanism is the illusion of algorithmic objectivity. Because AI is “just software,” perceived as emotionless and free of personal bias, we tend to attribute neutrality to its outputs—neutrality it does not possess. We forget that these models are trained on human data, full of bias, errors, and contradictions.

But when an AI states something with grammatical and structural confidence, our critical filter shuts down. It’s the “computer says no”

phenomenon made famous by Little Britain: once the machine has spoken, the discussion is over.2

The third factor is anthropomorphism. When a system re-sponds in natural language, with a tone that seems empathetic or authoritative, our brain attributes human qualities to it: intentionality, understanding, even wisdom. This is especially dangerous with Large Language Models, engineered to imitate expert human style flawlessly. The average user cannot distinguish between “seems to understand” and “actually understands.”3

The result is a form of algorithmic fundamentalism: the belief that if AI said it, it must be true.

1E. Short et al., “The Impact of Anthropomorphism on Trust in AI Systems”, ACM HCI, 2022.

2J.L. Skeem & D.A. Lowenkamp, “The Illusion of Objectivity”, Law and Human Behavior, 2018.

3A.K. Waytz et al., “The Mind in the Machine”, Journal of Experimental Social Psychology, 2014.

53

We’ve seen students defend AI-generated theses even when they were blatantly wrong, professionals present hallucinated reports as verified data, and users share invented historical “facts” as absolute truth.

In their perception, AI never makes mistakes; at worst, the human misunderstood.

This credulity has real, measurable consequences.

In medicine, patients abandon validated treatments to follow chatbot-generated “advice.” In finance, investors make decisions based on fabricated market analyses. In law, attorneys submit nonexistent citations in court.

In every case, the common denominator is the same: the refusal to verify, to doubt, to think.

The infallible oracle is not AI. The infallible oracle is our willingness to believe it is.

And until we recognize that every algorithmic output is, at best, a probabilistic hypothesis to be verified—not a revealed truth to be accepted—we will continue building our lives on foundations of digital sand.

The tool can generate anything. Only we can decide what deserves to be believed.

54

AI, the New Toxic Wikipedia

Tograspthescaleoftheepistemicdisasterunfoldingbefore us, it helps to remember the controversies that surrounded Wikipedia in its early years. Academics were horrified at the idea that an encyclopedia written by anonymous volunteers could replace traditional sources. They were wrong—but they were right about one thing: Wikipedia was, and still is, a fragile epistemo-logical experiment, built on the assumption that a community will continuously correct its own mistakes.

Today, that experiment has been eclipsed by something far more dangerous.

Generative artificial intelligence is not a collaborative, self-correcting Wikipedia. It is a toxic Wikipedia—monolithic, unaccountable, and capable of producing errors with the same fluidity with which it produces truths, at a speed no army of human editors could ever match.

The technical term for these errors is hallucination—an euphemism that hides the severity of the phenomenon. A hallucination is not a typo or a minor inaccuracy. It is a complete fabrication, delivered with the same grammatical confidence and structural polish as a verified fact.

An AI can invent a Supreme Court ruling that never existed, cite a scientific paper that was never published, or attribute a quote to a historian who never said it—all with an authoritative, convincing tone.4

4S. Huang & P.S. Yu, “A Survey on Hallucination in Large Language Models”, arXiv:2311.05232, 2023.

55

The fundamental difference from Wikipedia is this: AI hallucinations are not bugs that can be fixed with an edit. They are an inevitable consequence of how Large Language Models work.

These systems do not “know” facts; they predict the statistically most probable sequence of words given a context. When the probability of an invented sentence is similar to that of a true one, the model has no internal mechanism to distinguish them. It produces both with the same statistical nonchalance.

But the problem doesn’t stop at individual hallucinations.

The real disaster is the industrial scale of the phenomenon.

Every day, millions of users generate billions of words through AI. A small fraction of these contain factual errors. These words are published on blogs, forums, news sites, corporate databases.

The web—once a chaotic but fundamentally human archive—is being flooded with a new kind of pollution: algorithmic slop, a term coined to describe automatically generated content that is low-quality and devoid of informational value.5

And here the circle closes, in a phenomenon researchers call model collapse.

When future generations of language models are trained on the web of 2025–2026, they will learn from an ecosystem already polluted by their predecessors. Instead of learning from verified human knowledge, they will learn from their own errors—amplifying them exponentially.

A 2024 study published in Nature demonstrated mathematically that training on synthetic data causes irreversible degradation in model performance—a form of informational entropy from which 5M. Harris, “The Internet Is About to Drown in AI Slop”, OneZero/Medium, 2024.

56

there is no return.6

This is a form of epistemic pollution unprecedented in human history.

We are not talking about fake news created by state actors to manipulate elections, nor conspiracy theories spread by charlatans.

We are talking about a system that produces falsehoods systematically, accidentally, and at planetary scale—making it increasingly difficult for the average person to distinguish truth from falsehood.

The AIdiot Generation is trapped in a paradox: it possesses the most powerful tool ever created for accessing knowledge, yet that same tool is poisoning the very sources from which knowledge should be drawn.

Those who blindly trust AI outputs are not merely making an individual mistake; they are actively contributing to the degradation of the global information ecosystem.

Every hallucination accepted, shared, or published is another brick in the construction of an alternative, synthetic, toxic reality.

The solution is not to stop using AI, but to stop treating it as a primary source of truth.

AI must return to what it truly is: a tool for synthesis, not discovery; an assistant, not an oracle.

Every fact it generates must be verified. Every citation must be traced back to the original source. Every data point must be cross-checked with independent references.

In a world where a machine can lie with the same elegance with which it tells the truth, human verification is no longer an academic luxury.

6I. Shumailov et al., “The Curse of Recursion”, Nature, 2024.

57

It is the last bastion of reality.

58

The Total Delegation of Critical Judgment

Thereisaprecisemomentwhenahumanbeingstopsbeing a thinker and becomes a mere executor. It doesn’t happen through a formal declaration, nor through rebellion. It happens through a sigh of relief—the feeling of having finally offloaded an unbearable weight.

That weight is critical judgment: the effort of evaluating, doubt-ing, comparing sources, weighing contradictory evidence, and taking responsibility for a conclusion.

Artificial intelligence has offered an apparently perfect escape from this burden.

Why spend hours reading conflicting documents, verifying the credibility of sources, and building an argument step by step, when a chatbot can do it in three seconds and return a structured, seemingly balanced summary?

Delegation is no longer merely operational (asking AI to write a text); it has become epistemic. We ask the language model not only to produce content, but to judge for us.

This phenomenon has deep roots in decision-making psychology.

Daniel Kahneman demonstrated that the human brain is inherently lazy: it prefers cognitive shortcuts (heuristics) over slow, effortful reasoning.7

AI is the perfect machine for feeding System 1: it provides rapid, fluent, reassuring answers, shutting down System 2 before it even has 7D. Kahneman, Thinking, Fast and Slow, 2011.

59

a chance to activate.

But delegating critical judgment has consequences far more serious than mere mental laziness.

Philosopher Hans Jonas, in his ethics of responsibility, argued that the ability to foresee the consequences of one’s actions—and to assume responsibility for them—is what defines the moral human being.8

When we delegate judgment to an algorithm, we are not simply saving time. We are abdicating our moral responsibility as citizens, professionals, and thinking beings.

Examples of this delegation are now everywhere.

A journalist publishes an article based entirely on an AI-generated summary, without checking the original sources. A manager approves a corporate strategy based on a chatbot-generated market analysis, without understanding the underlying assumptions. A citizen forms a political opinion by reading only algorithmic news digests, never consulting primary sources.

In every case, the pattern is the same: the renunciation of personal judgment in favor of algorithmic judgment perceived as objective and infallible.

The cruel irony is that AI does not possess critical judgment.

Large Language Models do not “understand” what they generate; they have no values, no ethics, no ability to distinguish true from false.

They produce statistically probable sequences of words—nothing more.

When we delegate our judgment to an automaton, we are not transferring it to a more capable entity. We are simply abandoning it 8H. Jonas, Das Prinzip Verantwortung, 1979.

60

into the void.

The social consequences of this delegation are devastating.

A society in which individuals no longer exercise critical judgment is a society vulnerable to manipulation on an industrial scale. There is no need to control traditional media: it is enough to control the algorithms that users consult uncritically. There is no need to censor dissenting opinions: flooding the information ecosystem with synthetic content makes it impossible to distinguish truth from falsehood, until the population—exhausted—stops seeking reliability and settles for whatever the algorithm serves.

The AIdiot Generation is defined precisely by this delegation. It is not biologically stupid; it is a generation that has voluntarily chosen not to exercise the most distinctively human faculty.

It has traded the effort of doubt for the comfort of algorithmic certainty. It has exchanged the responsibility of judgment for the illusion of machine objectivity.

Regaining control requires a conscious act of rebellion.

It means accepting that critical judgment is tiring, slow, imperfect.

It means recognizing that human error is preferable to algorithmic certainty—because human error can be corrected, while algorithmic error, multiplied across millions of users, becomes a systemic disaster.

It means understanding that in an era of informational abundance, true competence is not accessing more data, but possessing the wisdom to evaluate it.

Critical judgment is not optional. It is the last bastion of human sovereignty.

And every time we delegate it to a machine, we sign a small, silent declaration of surrender.