Imagine a world where every doubt is resolved with a single question to artificial intelligence. You ask — and it instantly serves you a polished synthesis of knowledge, the essence of a topic delivered in a few crisp sentences. No need to wade through hefty volumes or dig through dozens of search results. You get an answer and move on. It's a seductive vision: knowledge at your fingertips, effortless, instant. A world that can be understood too easily.
No wonder so many see the latest AI technologies as the dawn of a new era — an era of cognitive convenience. People speak of a breakthrough on par with the invention of the printing press, or even the birth of artificial general intelligence (AGI) that will solve humanity's most pressing problems for us. If intelligent algorithms can instantly summarise books, analyse data from around the globe, and recommend personalised knowledge, doesn't that amount to a civilisational leap? Won't we become smarter, sharper, more aware? That is certainly the prevailing, enthusiastic tone — a narrative promising that AI will be our greatest teacher, adviser, and partner in understanding reality.
But every age of disruption breeds doubts. This essay is precisely a counterpoint to the unbridled optimism surrounding AI. Let's take a critical look at this vision of a world "understood too easily." Could readily available syntheses of knowledge conceal a hidden trap? What cognitive risk might arise when everything is handed to us on a silver platter — filtered through algorithms, stripped of context and first-hand experience? Let's start with what undeniably appeals: the comfort.
The Comfort of Summarisation, the Allure of Algorithmic Recommendation
The human mind is governed by the law of least effort. Our brains are inherently a bit lazy — they avoid unnecessary exertion in order to conserve energy. Easier is often subconsciously mistaken for better. When a quicker, simpler route appears, it's hard to resist. Technology exploits this tendency flawlessly.
Not long ago, Google was the symbol of cognitive convenience — instead of wandering through a library, you typed a few words and had an answer in seconds. Today, we've gone a step further. Thanks to language models like ChatGPT or Gemini, we can get not just a list of results, but a ready-made summary of any topic. "Ask a question, receive a polished synthesis, and move on — it looks like effortless learning," as one new-media researcher aptly put it. And indeed, that's exactly how it feels: like magic. Instead of reading a multi-hour report on climate change, we ask AI for a five-point summary. Instead of ploughing through a legal document, we get a clear explanation in plain English. The savings in time and energy are colossal.
A similar comfort is offered by algorithmic recommendations, which increasingly curate our informational everyday life. We no longer need to choose what to read, watch, or listen to — smart applications do it for us. Spotify assembles a playlist that perfectly matches our taste. Netflix suggests a series tailored to our previous choices. Facebook and X (formerly Twitter) ensure we see primarily the news and posts that will "engage" us. GPS navigation guides us along the most convenient route, eliminating the need to orient ourselves. In each of these situations, the algorithm takes on the effort that once rested on our shoulders: the effort of searching, selecting, deciding. All that's left for us is to consume what's been served.
It all sounds wonderful. We live in an age of numerous life simplifiers, as psychologists of new technology observe. We carry a smartphone in our pocket — a universal remote control for reality. Why burden your memory when any fact can be googled? Why learn from mistakes when an app will suggest the optimal solution anyway? Our great-grandparents had to painstakingly memorise facts, addresses, routes — we have a "brain in the cloud." A digital assistant always ready to answer or remind. On one hand, it's a dream come true: technology makes us faster and more efficient.
On the other hand, it's worth asking: does easier really mean better? In nature, nothing comes for free. If a body lies motionless for a long time, muscles atrophy. The same may be true of the mind: a brain that doesn't exert itself becomes like a body sunk into an overly comfortable armchair. At first, it feels blissful — but over time, getting up becomes increasingly difficult. New technologies undeniably make us somewhat cognitively lazy. They minimise the friction in acquiring information — and it was precisely that friction, that effort, which until now was essential for genuine understanding.
We're already observing the first side effects of living in informational comfort. Take the phenomenon of digital amnesia, also known as the Google effect. It refers to the fact that when we know a piece of information can be found online at any moment, we stop memorising it. Our brain says: "Why hold onto this when I can look it up on my phone in a second?" Research has confirmed that people today have greater difficulty recalling facts, numbers, and dates — but they're excellent at remembering where to find them. In other words, we're offloading the burden of memory onto devices. We remember the web address rather than the content. It seems like a minor shift — and yet it means that we're truly assimilating less and less information as our own. Knowledge becomes something we have access to, but no longer carry in our heads.
The same may apply to deeper thinking skills. A few years ago, researchers noted that younger generations, raised from childhood with the internet, were quite adept at finding information but struggled with critical synthesis and analysis. Now that AI offers a ready-made synthesis on any topic in seconds, even that last effort — independently processing the information found — may prove unnecessary. Why puzzle over drawing conclusions when an AI model will do it for us? This gives rise to the temptation of letting algorithms think on our behalf. As Dr Jakub Kuś, a psychologist of new technology, writes, we are witnessing the emergence of an app generation (AppGeneration) — young people dependent on digital solutions in every daily activity. Their brains are developing along a somewhat different trajectory than those of their predecessors — because they've never had to function without the constant support of digital tools. What does this mean in the long run? We don't fully know yet. But alarm signals are already appearing that something essential may be slipping away from us.
The Illusion of Understanding: When Knowledge Comes Too Easily
What's most alarming is not the convenience itself, but the illusion that may spring from it. Because when we receive a ready-made answer in a few seconds, we feel as though we've understood the topic. We can tell someone: "Yes, I know — I've read about that." But do we really know? Or have we merely skimmed a few sentences of a summary?
It's a bit like reading an abridged summary of a novel instead of the novel itself. We can summarise the plot of War and Peace in a few sentences — but does that mean we've grasped the depth of the work? Or like watching a film trailer and being convinced we already know the whole film. Such second-hand knowledge, in summary form, carries the risk of an illusion of competence. We feel familiar with a topic, yet our knowledge may be fragile, fragmentary, devoid of context. Worse — we may not even be aware of what we don't know, because we haven't gone deep enough to spot the gaps.
Scientists call this phenomenon the illusion of explanatory depth — people often believe they understand something far better than they actually do. A classic experiment: participants were asked whether they knew how an ordinary door lock or a toilet flush works. Most replied: of course, it's simple. Then they were asked to describe the mechanism in detail — and suddenly it turned out their understanding was riddled with holes. They had only an appearance of knowledge, an impression of grasping the principle, while their actual understanding was shallow. The mere fact of knowing the general idea (e.g., that we turn a key and the bolt slides) had been mistaken for a real understanding of the specifics.
In the age of the internet and AI, this illusion may intensify. A quick online look-up delivers instant cognitive satisfaction. We see a neat answer and everything seems clear. Moreover, psychological studies show that after using a search engine, people overestimate their knowledge in other, entirely unrelated domains. This happens because the mind treats the internet as an extension of its own memory. If I can always find something with one click, the boundary between what I know myself and what the web knows begins to blur. Great erudition without learning — warnings about this state of affairs appeared long ago.
It's worth recalling a scene from centuries past: when writing was invented in ancient Greece, some sages sounded the alarm then, too. In the Phaedrus, Plato puts words of criticism against writing into the mouth of Socrates — that it is supposedly an invention that will give people "a recipe for reminding, not for memory." Students, warned the mythical King Thamus, "will possess great erudition without learning and will fancy themselves knowing much, while for the most part they will know nothing." In other words: writing was to give an appearance of wisdom to those who did not truly possess it. Sound familiar? Today, the role of such "writing on steroids" may be played by the omniscient AI serving us knowledge in capsule form. We may feel we're becoming experts on everything — while in reality, we're only skimming the surface.
This illusion of understanding is very comfortable, but can prove dangerous in practice. For example, a pupil or student who, instead of reading the assigned texts or scholarly papers, relies exclusively on AI to generate summaries, may believe they've "got it." Until the moment they face a trickier question or a practical application of knowledge — and it turns out they lack the foundations to answer. The machine shortcut didn't teach them to think independently or solve problems. They saw the answer, but they didn't see the process of arriving at it. It's a bit like knowing the result of an equation without understanding the mathematics behind the solution.
Furthermore, there is a risk that after receiving a ready-made answer, we lose the motivation to dig deeper. Researchers at the University of Pennsylvania recently conducted a series of tests in which participants were asked to learn a new subject (e.g., how to start a vegetable garden) using two methods: one group used traditional web searching, while the other could ask ChatGPT and receive a ready-made explanation. The results were unequivocal. Those who relied on AI felt they had learned less, subsequently wrote less detailed and more generic explanations for others, and independent evaluators rated their knowledge as shallower and less useful. Even when the experimenters ensured both groups received an identical set of facts, those who received the facts from AI in the form of a synthesis had poorer knowledge than those who clicked through various sources themselves and synthesised everything on their own. Why? Because traditional information seeking, though laborious, engages our brain far more actively. You must read, select, compare, connect the dots — in short, do the mental work that builds genuine understanding. An answer from AI, by contrast, is served on a platter — we accept it passively, without breaking a sweat along the way. The result: less effort, less retained, less thought through.
In other words, nuance and context escape us. An algorithm often smooths out and simplifies complex matters to deliver a clear answer. This creates the illusion that things are simpler than they truly are. If an AI model can summarise the causes of the conflict in the Middle East in a single paragraph, we might believe the conflict can be easily resolved — while anyone who has delved into the region's history knows how complicated it is. Similarly, a recommendation-driven newsfeed on social media can give us a false sense of being excellently informed about world events, because we're constantly scrolling and something keeps catching our eye. But reading only a curated slice of news tailored to us, we may not see the full picture. We see what the algorithm selected based on our preferences — and it omits what might jolt us from our cognitive comfort zone. We live, then, in an information bubble, convinced that this is what the world looks like, because nothing else reaches us. This is also a form of cognitive illusion: we confuse the map (in this case, a personalised map of information) with the territory of reality. And yet the map is not the terrain. Even the most perfect map is a simplification, a symbol — it cannot capture the full complexity of the real landscape.
Content-recommending algorithms have yet another effect: they lock us inside a narrow circle of experiences. If you listen to one genre of music for a while, the app will keep feeding you similar tracks on repeat. If you read only a certain type of article, you'll soon see mostly that. This encourages the formation of informational tunnels, where everyone moves along their own habitual path. As a result, we can get stuck in intellectual monotony, never encountering anything new. The algorithmic gatekeeper ensures nothing "bores" or surprises us — but in doing so, it steals the chance to broaden our horizons. We stop confronting different perspectives, we lose our openness to the unexpected. And yet what is cognitively valuable often emerges precisely from stepping outside the bubble, from encountering something unforeseen, difficult, demanding reflection.
Interestingly, I recently conducted a rather simple but disturbingly effective experiment on myself. For a period of time, I began intensively browsing social media exclusively for content related to firearms — pistols, rifles, military equipment — and geopolitical topics: Poland, Ukraine, Greenland, Iran, Venezuela. The effect appeared faster than I expected. In a short time, nearly every platform I use (perhaps apart from LinkedIn — though even there, not entirely) began flooding me with content about pistols, assault rifles, tanks, drones, and armed conflicts. The algorithms unhesitatingly locked me inside a hermetic bubble that became a kind of informational prison. What's more, if an outsider looked solely at my feed, they might conclude I was a firearms fanatic — or someone far more unsettling. And yet it was merely momentary curiosity and a deliberate experiment. This situation painfully illustrates how easily an algorithm not only simplifies our picture of the world, but also constructs a simplified — and often false — picture of ourselves.
The long-term consequence of this state of affairs may be a general shallowing of our understanding of the world. Since everything is served in convenient, digestible portions, we less frequently reach the deeper layers of knowledge. Why consult the sources, read multi-hundred-page books, when we have a distilled summary? Fewer and fewer people will have the patience to read original philosophical or scientific texts — after all, AI will summarise the key theses for us. Except that in this way, we lose the flavour of the original. We lose contact with first-hand experience — with raw data, primary texts, with observation through our own eyes. We settle for meta-knowledge (knowledge about knowledge) instead of knowledge itself. And this leads to a kind of second-handedness in our cognition. We know everything at one remove, from reviews and digests — and fewer and fewer things directly.
Dr Jakub Kuś makes an astute observation: "Many of the conveniences we have access to only seemingly make our lives easier." He gives the example of social media: we fall into the illusion that we know what's going on with our friends because we see their posts on Facebook or Instagram. As a result, we call them and meet them less often — it seems to us that the contact has been maintained, when in fact we know only carefully curated snippets of their lives, those "pretty pictures" published online. True, deep human connection is supplanted by its algorithmic ersatz. The same may happen with our relationship to knowledge: instead of direct engagement with a subject (through personal experimentation, thorough study of the literature, conversations with experts), we have a convenient interface in the form of AI that simulates it all for us. We receive answers without bothering to ask the questions ourselves. Something, however, may be slipping away — something difficult to define but crucial: a sense of real depth and understanding.
Are We Heading Toward a Civilisation of Shallow Knowledge?
Let's now consider the consequences broader than the individual. If most people begin to rely on accelerated knowledge synthesis and algorithms in their everyday understanding of the world, what impact will this have on our cognitive culture? Will humanity as a whole become more intelligent thanks to AI — or, on the contrary, do we face the cognitive shallowing of civilisation?
History teaches us that every automation of certain skills carries the risk of humans losing proficiency in those very skills. When the calculator was invented, fewer and fewer people could perform complex calculations in their heads — why bother, when a machine does it faster and flawlessly? When GPS navigation became widespread, people stopped paying attention to maps and the lay of the land — because the app always points the way. Research has even shown that the brains of professional taxi drivers in the pre-GPS era had a more developed hippocampus (the structure responsible, among other things, for spatial orientation), because those drivers had to memorise the street grid from memory. Young drivers today, relying exclusively on maps in their smartphones, no longer develop these abilities — their brains operate differently, knowing that the task of external orientation is handled by a device.
Similar examples proliferate across various fields. Airline pilots increasingly warn that over-reliance on autopilot weakens their manual flying skills. In an emergency — when full control of the aircraft must be taken — it turns out they've fallen out of certain habits and reactions, because electronics handled most of the flight for them. Several crashes have been partly attributed to this factor: degradation of pilot alertness and competence as a result of excessive automation. In response, the American FAA began recommending that pilots deliberately disengage the autopilot more often and practise traditional flying so as not to rust in their skills.
In medicine, there are analogous warnings about the risk of diagnostic vigilance loss among doctors who begin to uncritically trust AI systems. Recent studies in Europe showed that gastroenterologists using AI for polyp detection during colonoscopy, over time, began detecting abnormalities less accurately when AI wasn't assisting them. This is referred to explicitly as the Google Maps effect — because just as a person accustomed to GPS gets lost without navigation, a specialist habituated to constant algorithmic support loses part of their own perceptiveness. Dependence weakens self-reliance. The doctors in that study admitted they felt less motivated and focused when relying on AI prompts — and when those prompts were suddenly removed, their performance dropped compared to their pre-AI baseline. As one commentator noted, "in effect, routine use of AI dulled human pattern recognition." What was meant to be merely an aid became a prosthesis — and natural faculties partially atrophied.
These examples teach us an important lesson: every tool that thinks for us carries the risk that we'll unlearn how to think for ourselves. Today, this concerns specific skills (arithmetic, navigation, piloting, medical diagnosis). Tomorrow, it may concern the general capacity for critical thinking and analysis, if we lean too heavily on intelligent knowledge synthesisers. For isn't it the case that knowledge acquired without effort breeds less understanding? When we stop questioning what an algorithm feeds us, it's easy to fall into the trap of misplaced certainty. Entire generations may learn to look up answers but not to verify them or understand the process of arriving at them. They'll know where to click to find a solution, but won't be sure what it truly means. And this threatens not just individual misunderstanding — it threatens the fragility of our entire civilisational knowledge.
Imagine a world where researchers stop delving into experiments because AI generates hypotheses and summarises the results of thousands of studies on its own. Science might seemingly accelerate — but who will guarantee that AI hasn't overlooked or flattened something? If we all trust synthetic accounts, there may be a shortage of those inquisitive souls who dig into the details, spot unusual cases, question established conclusions. Cognition will become more uniform, averaged out — because algorithms, fed on big data, gravitate toward generalisations and consensus. Yet progress is often born from deviations from the norm, from exceptions noticed by keen observers. If no one looks at the raw data anymore (because everyone reads only the synthesis), who will spot, say, a new, strange symptom in medicine that doesn't fit existing disease profiles? Or who will discover a novel interpretation of War and Peace when everyone uses only one summary-guide? We risk falling into cognitive uniformity and stagnation.
Moreover, who will control this omnipresent synthetic knowledge? Today, it's large technology corporations that possess the most advanced AI models and unimaginable volumes of data. By unreflectively using their "free" services, we effectively hand them power over our minds. That's a sharp formulation, but think about it: if we daily feed on content selected by a company's algorithms, if we trust the answers generated by their model — then that company indirectly shapes our views, our tastes, our knowledge of the world. And algorithms, as Dr Kuś notes, "feed on the emotions of internet users; the bigger the uproar, the higher the revenue." This is why extreme, controversial content is promoted on social media — it boosts engagement, even at the cost of social polarisation and the dumbing-down of public debate. The same may be true of algorithms delivering knowledge: they may emphasise what's attractive and easy to swallow, while avoiding what's complicated and ambiguous. Our collective wisdom may degrade if we succumb to this tendency. Wisdom requires the confrontation of different perspectives, stepping outside the bubble, debate — while an algorithm tends to reinforce what we already think (because that keeps us satisfied longer and inclined to stay on the platform).
As a result, we may become, to paraphrase the words of a certain contemporary thinker, pancake people — spread wide and thin across a vast surface of information. We know a little about everything, but nowhere do we go deeper than a millimetre. Thinly spread knowledge is the antithesis of the old vision of a wise person like an oak with deep-reaching roots. Is that shallow versatility what we want for future generations?
Perhaps it's a price worth paying for all the benefits of AI. After all, there's no denying that technology gives us real power. Thanks to algorithms, we optimise transport, discover new medicines, streamline millions of processes. Divine technology — as the late, eminent biologist Edward O. Wilson aptly put it — has fallen into the hands of beings with Stone Age minds and medieval institutions. This dissonance between our biological heritage and the might of our tools is a fact. Wilson called this combination "fatally dangerous." Our emotions and intuitions can't keep pace with the explosion of capabilities we've bestowed upon ourselves. In other words: we tend to use these tools in ways that exceed our grasp.
Will we fail to keep up in the case of omnipresent AI-driven knowledge synthesis, too? Convenience and time savings ensure we'll probably use these capabilities ever more widely — just as we massively adopted GPS, social media, and Google Translate. Perhaps soon it will be standard in schools for students, instead of writing essays themselves, to edit AI-generated answers. Perhaps in companies, preliminary analyses and reports will always be drafted by a machine, with a human merely glancing them over. These are scenarios tempting in their efficiency. But we must closely observe what happens to our cognitive potential. If we notice our intellectual muscles growing slack — memory weakening, concentration evaporating after a few seconds, the capacity for critical thinking fading — that will be an alarm signal. Perhaps we're already hearing it: numerous studies point to a deteriorating ability for deep reading and sustained attention in younger generations, shaped by the flickering stream of short-form online content. The pace of information consumption imposed by digital media has accustomed us to haste and superficiality. We grow impatient when something demands more than a minute of our attention. In such a climate, the proliferation of AI knowledge-pills is virtually certain — because they fit perfectly into our cognitive style. The only question is whether it won't deepen this trend even further.
Increasingly, calls are heard to consciously introduce healthy friction into the use of new technologies. To learn to use AI strategically: where it genuinely saves us routine work, while simultaneously not allowing it to take over our key mental processes. Perhaps the solution will lie in designing tools that, instead of hooking us on instant answers, nevertheless encourage a degree of interaction, reflection, checking of details. Some researchers are already experimenting with AI models that offer the user the sources and links they draw from — to encourage independent verification. But here, too, an interesting finding emerged: it turned out that when a person first receives a summary, even with sources at hand, they rarely look at them. In other words — once we feel sated by an answer, the motivation to delve deeper drops dramatically. As Dr Kuś put it, if we continue down the path of cognitive laziness, we'll become "a human appendage to an application" instead of an autonomous agent. Strong words, but they capture the essence of the threat: we'll voluntarily reduce ourselves to the role of executors of a machine's whispered prompts.
Does it have to be this way? Fortunately, there is also a noticeable grassroots resistance to total reliance on technology. In many countries, smartphone use is banned in classrooms so that children also learn analogue skills and focus. There's a revival of interest in simple ("dumb") phones, stripped of a thousand distractions. More and more people consciously disconnect for a while from the digital noise — to rediscover the taste of their own thoughts, untainted by an algorithmic feed. These phenomena show that we recognise the problem and are trying to regain control.
In closing, it's worth emphasising: the point is not to paint an exclusively bleak scenario nor to condemn artificial intelligence outright. I'm an AI enthusiast myself. But I believe that, like any powerful technology, AI is a tool — and it's up to us how we use it. It can wonderfully support us in solving complex problems, but it can also make us lazy and dull if we're unreflective. What seems crucial is maintaining a balance between comfort and challenge. A convenient summary doesn't replace your own analysis, and an algorithm's recommendation shouldn't be the final word — sometimes it's worth checking what lies beyond it. Collective wisdom grows when individuals ask questions, challenge the status quo, learn through experience and trial and error, rather than merely consuming ready-made answers. Even the finest synthesis cannot replace the intellectual experience — that moment of illumination when you fully grasp something on your own, after toil and inquiry.
Where, then, are we heading? Will our minds flourish thanks to AI in the years ahead, finally having time for creative thought while machines handle the grunt work of information selection? Or will our curiosity and inquisitiveness drown in a flood of easy answers, leading to a world of seemingly all-knowing ignoramuses? Will a civilisation that understands everything too easily lose the ability to truly understand anything at all? These are questions whose answers we're still shaping — every day, through seemingly trivial decisions: whether to click "learn the shortcut" or to delve into the topic ourselves. Whether we want to realise the full extent of our intellectual potential — or whether we're content to carry our brains in the cloud of a large corporation…
