The Efficiency Paradox: Does AI Make Us More Stupid?

Irony of Automation: How AI Is Mentally De-Training Us

While the world celebrates the efficiency of artificial intelligence, a disturbing paradox emerges: AI isn't replacing us, it's de-training us. And this process of "cognitive offloading" is changing the way we think and remember.

The Mental GPS: When Efficiency Becomes the Enemy.

Remember when you could find your way around town? When you could recite your friends' phone numbers by heart? What happened to our sense of direction with GPS is now happening to our cognitive abilities with AI.

A 2020 study published in Nature Neuroscience by Louisa Dahmani of Massachusetts General Hospital showed that relying on GPS for navigation significantly reduces activity in the hippocampus, the brain region crucial for spatial memory and navigation.

The Google Effect: The Precedent That Explains Everything

The phenomenon has solid scientific roots. The "Google Effect" or digital amnesia was first documented in 2011 by psychologist Betsy Sparrow of Columbia University in a study published in Science.

Research has shown that people are less likely to remember information when they know they can easily retrieve it online. In one of the experiments, participants remembered better where to find information rather than the information itself.

The digital amnesia data seems concerning:

• According to a 2015 Kaspersky Lab study, 91% of people in the United States and Europe admitted to using the internet as an online extension of their memory
• Only 49% of participants could recall their spouse's phone number
• 71% did not remember their children's phone numbers

Microsoft-Carnegie Mellon Research: The First Data On AI.

A 2025 study conducted by researchers from Microsoft and Carnegie Mellon University analyzed 319 knowledge workers and their use of generative AI tools. The results show that:

• Workers report a "perception of enacting critical thinking" when relying on AI tools
• The use of AI produced "a less diverse set of outcomes for the same task" compared to people relying on their own cognitive abilities
• A tendency toward "cognitive offloading" is observed - delegating mental processes to external tools

But Wait: Not All "De-Workouts" Are the Same

Before continuing, let's take a critical look. This phenomenon is not new:

The Calculator

Who can still do long division by hand? The calculator has been "de-training" us in mental calculation for decades. Yet, mathematics is not dead - in fact, it has flourished. Freed from tedious calculations, mathematicians have focused on more complex and creative problems.

Scripture vs. Oral Memory

Socrates himself feared that writing would weaken memory. In Plato's dialogue Phaedrus (circa 370 BC), Socrates recounts the Egyptian myth of Theuth and Thamus, where Theuth presents writing as an invention that will improve wisdom and memory. But King Thamus objects: "This invention will produce forgetfulness in the souls of those who learn it: they will stop exercising memory because they will rely on writing, which is external".

He was right: the storytellers who recited the entire Iliad by heart are gone. But we have gained the ability to preserve and share complex ideas on a global scale.

The Press vs. the Calligraphy

Gutenberg's printing press (1440) made fine calligraphy obsolete. Before the printing press, in 14th-century Europe 80% of English adults couldn't even write their own name. By 1650, however, 47% of Europeans could read. By the mid-1800s, the number had risen to 62%.

We lost an art, but democratized knowledge. As historians have noted: "The net increase in literacy broke the literate elite's monopoly on education and learning and supported the emerging middle class".

The pattern is clear: every technological leap "de-trains" some abilities and enhances others.

So What IS the Difference with AI?

If every technology "de-trains" something, why should AI worry us more? The difference lies in three critical factors:

1. Speed and Pervasiveness

Electronic pocket calculators, marketed since 1971, replaced complex mental computation within about 15 to 20 years. AI is replacing critical thinking in less than 5 years.

‍We can no longer think in terms of generations as we did in the past - now we have to think in 5-year cycles, not 20-30.

Speed matters: the brain has less time to adapt and develop new compensatory skills. Human societies have traditionally evolved slowly, allowing institutions, education and culture to gradually adapt to technological change. But AI compresses this adaptation process by decades to decades, creating an unprecedented cultural and cognitive shock.

2. Breadth of Cognitive Offloading

• Calculator: replaces arithmetic calculations
• GPS: replaces spatial navigation
• AI: replaces reasoning, creativity, writing, analysis - cross-cutting skills we use in every field

3. Lack of Metacognition

With a calculator, you know you can't do long division. With AI, you often don't notice that you've stopped thinking critically. It's a silent, unconscious decline.

The Theory of AI-Induced Cognitive Atrophy.

The concept of "AI chatbot-induced cognitive atrophy" (AICICA), theorized in a 2024 study, is based on the brain development principle of "use it or lose it," arguing that excessive reliance on AI without simultaneously cultivating fundamental cognitive skills can lead to underuse of cognitive abilities.

A 2009 academic study published in Symbolae Osloenses had already drawn this parallel with the calculator: "The pocket calculator allows us to produce solutions to calculation problems, but does it enable us to know these solutions? It depends on what we mean by knowing here. If it means we should also be able to justify the solutions, explain why they are indeed the correct ones, then certainly not".

"It's Not a Bug, It's a Feature": Cognitive Dependence by Design.

But here's the twist: cognitive dependence might not be a side effect, but a design feature.

Crucial difference: the calculator didn't need you to become dependent to be profitable. AI does. The more you use it, the more data it generates, the more it improves, the more indispensable it becomes. It's a business model built on dependency.

It's a self-reinforcing cycle: the more effective AI is, the more dependent we become. The more dependent we are, the less we exercise our abilities. The less we exercise them, the more we need AI. It's like developing a tolerance to a substance: you need increasingly larger doses to get the same effect.

The Paradox of Cognitive Freedom: When Being Free Makes Us Prisoners

Medicine

A 2024 study published in Perspectives on Psychological Science warns that in radiology, where artificial intelligence is increasingly used, doctors risk gradually losing their intuitive diagnostic skills. But be careful: AI is freeing radiologists from the routine analysis of thousands of normal scans, allowing them to focus on complex and atypical cases. The risk is not that AI will replace diagnosis, but that doctors will stop training their "clinical eye" on routine cases - which often hide subtle details crucial for recognizing rare anomalies.

Programming

A 2025 study highlights an interesting phenomenon: developers who constantly rely on AI to write code develop a kind of cognitive dependency. AI excels at generating boilerplate code and standard functions - repetitive work that used to steal precious hours. The problem: freed from these tedious tasks, some programmers stop exercising algorithmic thinking even when they really need it. It's like a surgeon who uses robotic tools for routine operations but then struggles to operate manually in an emergency.

Education

As educator Trevor Muir explains: "I don't think teachers should use AI with students in writing until students have first mastered it." AI can correct grammar, suggest synonyms, even structure essays - all activities that used to require hours of manual revision. The hidden value: those errors and that seemingly "useless" struggle are actually a workout for the brain. It's like learning to drive with a manual transmission before an automatic: it seems harder, but it develops a control and understanding of the vehicle that automatic can't give.

It is like learning to drive: first you have to develop reflexes and road intuition through "inefficient" practice, then you can use cruise control safely.

As Socrates predicted in Phaedrus: "You will give your students the appearance of wisdom, not its reality. Your invention will allow them to hear many things without being properly instructed, and they will imagine they have come to know much while for the most part they will know nothing."

The Test of "Imaginary Substitution" (Revisited)

Instead of asking "Can AI do this?", try this updated thought experiment: "If everyone used AI for this tomorrow, what would we lose as a species? And what would we gain?"

• Writing: We would lose the ability to articulate complex thoughts → But would we gain time for deeper thoughts?
• Navigation: We would lose our sense of direction → But would we gain efficiency in getting around?
• Calculation: We have already lost mental arithmetic → But we have gained the ability to solve more complex problems

The real question: are we aware of the consequences of our choices?

The Strategy of Cognitive Resilience: How Not to Be Replaced by Your Assistant

1. Use AI to Amplify, Not to Forget

"Use AI to amplify your skills, not to forget them. Let it free you from strenuous work so you can focus on the creative and complex aspects -- but don't let those core skills atrophy through disuse."

2. Keep the "Cognitive Muscles" Trained

It's exactly like physical training: if you stop going to the gym for two months, you don't notice it when you look in the mirror-you look the same. But as soon as you try lifting a heavy weight or running stairs, you immediately feel the difference. Your muscles have silently weakened.

‍Cognitive atrophy is even more insidious: not only do you not notice it while it's happening, but often you don't even notice it when you need that ability - you simply delegate to AI without realizing that you once would have been able to manage on your own.

3. Practice the "First Without, Then With" rule.

To maintain our cognitive abilities, we must directly practice fundamental skills before delegating them to AI, and even after delegating them we must still continue to train them. It's not a question of "fundamental" vs "superfluous" skills, but of keeping the mind trained.

Just like a chess player who always uses the computer to analyze moves: he becomes technically accurate, but if he never reasons independently, he loses strategic intuition and the ability to 'feel' the position.

‍

The Future: AI As Collaborator, Not Crutch

The solution is not to reject AI, but to use it strategically. The professionals who will thrive will be those who combine their human intuition and experience with AI's superpowers - those who know when to delegate and when to think independently, always keeping control of the decision-making process.

Conclusion: It's a Feature, Not a Bug (But Which Feature?).

Cognitive atrophy caused by AI is not a flaw to be corrected - it is a design consequence that we must recognize and consciously manage.

But be careful: not all "de-trainings" are negative. The calculator freed us from tedious calculation, print freed us from oral memory, GPS freed us from the need to learn every street.

The real challenge is distinguishing:

• When de-training is liberating (it frees up cognitive resources for more important things)
• When it is impoverishing (it reduces abilities we need to think independently)

The question is not whether AI will replace us, but whether we will be aware enough to choose what to replace and what to keep trained. The future belongs to those who know when NOT to use AI.

FAQ: The Most Common Questions About AI and Cognitive Atrophy.

"Is AI making me stupid?"

No, it's not making you stupid. AI is making you cognitively lazy in some specific areas, just as GPS made you lazy in navigation. Your basic intelligence doesn't change, but you risk losing the habit of using it in certain contexts. Fortunately the process is reversible: you just need to start training again.

"Is it true that ChatGPT destroys the brain?"

Absolutely not. The sensationalist studies you've read about in the newspapers are often based on preliminary research with small samples. There is no scientific evidence that using AI causes brain damage. The problem is more subtle: it can reduce the motivation to think independently, not the ability to do so.

"Should I stop using AI?"

No, that would be counterproductive. AI is a powerful tool that can amplify your abilities. The key is to use it strategically: let it handle repetitive and tedious tasks, but keep critical skills active. It's like going to the gym: go ahead and use the machines, but don't forget bodyweight exercises.

"Will my children grow up less intelligent?"

Not necessarily. Children who grow up with AI might develop different skills than ours: greater ability to collaborate with intelligent systems, faster thinking in selecting information, creativity in combining multiple resources. The risk is that they skip fundamental formative steps.

‍But the real challenge will be the same for everyone - children and adults: learning to balance cognitive autonomy and collaboration with AI. Children might even be at an advantage, naturally growing up "bilingual" in both modes.

"Will AI completely replace human work?"

Not in the sense you think. AI does not entirely eliminate any "professional role," in fact, but transforms individual tasks within existing roles. And this is generating three simultaneous phenomena:

1. Layered automation: AI first replaces the most routine tasks, then increasingly complex ones. An accountant might see basic calculations automated first, then trend analysis, then even part of strategic consulting. The work transforms gradually, it doesn't disappear all at once.

2. Value polarization: A divide is forming between those who can collaborate effectively with AI (and become more productive) and those who can't (and become obsolete). It's no longer enough to be good at your field - you need to be good at your field + AI.

3. New bottlenecks: While AI handles analysis and routine, skills that once seemed "soft" become crucial: complex negotiation, leadership in ambiguous situations, creativity applied to problems never seen before. Paradoxically, the more capable AI becomes, the more valuable the more "human" skills become.

The real question isn't "Will my job disappear?" but "Which parts of my job can I delegate to AI today so I can focus on the ones only I can do?" And then, in six months, you'll have to ask yourself the same question again.

The paradox of mobile competence: the better you get at collaborating with AI, the faster you have to reinvent your role. Tomorrow's professionals won't have a fixed "core business" anymore, but a meta-skill: knowing how to quickly identify where to add human value in a landscape that changes every quarter.

"Is it normal that I can no longer write without AI?"

It's normal but not inevitable. If you've developed a dependency on AI for writing, you can "detox" gradually. Start with short texts without assistance, then progressively increase the complexity. It's like getting back in shape after a sedentary period: it's tiring at first, but your strength comes back quickly.

"Will AI make me lose my creativity?"

Only if you use it badly. AI can be a great creative partner if you use it for brainstorming, overcoming blocks, or exploring unexpected directions. The risk is using it as a substitute for your creativity instead of an amplifier. Golden rule: the idea must always start with you, AI can help you develop it.

"How do I know if I'm using AI too much?"

Try this test: try doing a task without AI that you normally delegate (writing an important email, solving a problem, doing a calculation). If you feel "lost" or much slower than usual, you're probably becoming too dependent on your digital assistant. Try working the old way every once in a while.

"Will AI make school useless?"

This is the hardest question. Traditional education is based on exercises (writing, calculations, research) that AI now does better than students. The dilemma: if you don't practice these skills because "AI does it anyway," how do you develop the critical thinking needed to evaluate when AI is wrong? But if you keep having students practice things AI does better, education seems anachronistic. ‍ A hybrid approach will likely be needed: developing basic skills through direct practice, then learning to orchestrate AI tools for complex goals."

"Is it just a passing fad?"

No, AI is here to stay. But like all technological revolutions, after the initial enthusiasm there will be a settling-in period in which we learn to use it better. "Cognitive offloading" is a real and lasting phenomenon, but we can manage it consciously instead of passively suffering it.

Remember: next time you're about to ask AI to write that email, stop and ask yourself - am I amplifying my abilities or am I atrophying them?