From “Playing with AI” to Structural Restructuring: What Is Happening in the World — and Why It Will Not Simply Blow Over

In the first material of the series, we recorded a thought that is uncomfortable for many: the question is not whether to introduce AI into Kazakhstan’s schools, but how to manage a process that is already under way. But before discussing “how,” it is worth answering an even more basic question that quietly sits in the mind of every principal, teacher, and parent: perhaps all of this is exaggerated? Perhaps, like previous “revolutions” — television in the classroom, interactive whiteboards, tablets — the current excitement around AI will subside in a couple of years, and school will return to its familiar life?

This question is not naive. It is the healthy caution of a person who has already seen more than one wave of hype. The most honest answer begins not with school, but with what is happening in the world beyond its walls. The main thing to understand about AI is this: school is changing not because someone decided to modernize it. It is changing because the world into which it sends children is changing. And this world, unlike previous technological fashions, is unfolding at such speed and on such a scale that there will be no return to familiar life.

To feel the temperature of the moment, it is enough to recall how 2026 began in the technology world. The first days of January brought a wave of statements that until recently would have sounded like science fiction: developers shared stories about new AI tools doing in an hour work that had previously taken a team of engineers a year, while the most prominent technology entrepreneurs seriously began to speak about humanity entering a “singularity” — a point beyond which technological progress becomes explosive and irreversible. One may treat such language skeptically; the technology environment likes grand formulations. But behind the rhetoric lies a measurable reality.

Let us start with the graph that looks the dullest, but is in substance the most important — the one economists show. It has two lines. The first is the cost of computing over the past decades. It is not falling smoothly, but collapsing: what once cost a fortune now costs pennies and continues to become cheaper. The second line is computing performance, the number of operations per second. It is doing the opposite: growing exponentially, year after year, with no signs of slowing down.

These two curves are the technical basis of the entire discussion. Previous “revolutions” in education ran into a physical ceiling: a television remained a television, a tablet remained a tablet, and their capabilities were fixed. AI is built differently. Beneath it lies an exponential curve that makes machines smarter and cheaper literally every year. That is why what seemed like fantasy in 2022 had become everyday reality by 2025, and what seems impossible today is highly likely to become ordinary by 2028.

This leads to a simple conclusion. If a technology follows an exponential curve, it will neither “blow over” nor “settle down.” It will only accelerate. A wait-and-see strategy — “let us wait until things calm down” — is a losing strategy, because things are not going to calm down. As one of the world’s most prominent AI experts, investor and former president of Google China Kai-Fu Lee, has noted, this technology will change the world more than anything in human history — more than electricity. One may debate the scale of the metaphor. But it captures the direction accurately.

Those who still think that talk about AI is about the future should look at the labor market in 2025. The future has already begun; it is simply unevenly distributed.

Throughout the year, global media carried stories that until recently would have sounded invented. A software engineer with a six-figure salary loses a job handed over to AI, sends out hundreds of resumes without response, and starts working as a courier. Senior executives with titles such as “AI director” — the very people who implemented automation — themselves fall under layoffs. The Russian-speaking labor market shows the same picture: the number of resumes rose sharply, the number of vacancies fell, and the two lines opened like scissors.

Notice who was hit first. Not low-skilled manual labor, as people became accustomed to thinking during previous waves of automation. The blow fell on “white-collar” workers — programmers, analysts, managers — precisely the professions for which children go to university after school. These are exactly the trajectories toward which the whole logic of secondary education is aimed today: study well, enter a prestigious field, and obtain stable intellectual work.

This is the first alarm bell for school. The familiar link “school — university — profession,” on which education has rested for the past half-century, is beginning to fall apart not in theory but before our eyes. School can no longer promise a child a predictable future by transmitting a fixed set of knowledge, because the professions for which that knowledge was sharpened are themselves becoming unpredictable.

What happens next? The honest answer is that no one knows for sure. But experts broadly agree that there are three main scenarios — and education needs to understand each of them, because it will have to prepare for all of them at once.

The first scenario is radical and rapid displacement. In this picture, automation advances swiftly. Something emerges that might be called “work under the API”: a person is still needed, but performs tasks set and controlled by an algorithm — effectively becoming an appendage to the system. Then even this layer collapses, because robots and programs begin to perform the tasks in full. The minority that creates and controls these systems — “work above the API” — comes out ahead; the majority is left without a place. This is the harshest scenario, and it cannot be dismissed.

The second scenario is an attack on mid-level competencies. Here the mechanism that economists call job polarization comes into play. The tasks under pressure are neither the simplest nor the most complex, but the middle ones — routine intellectual operations that can be easily described by an algorithm. Simple manual labor that requires physical dexterity in an unstructured environment may hold out for a while. Creative and strategic tasks on the upper floor may also hold out. But the middle, where most “normal” office employment is concentrated, is hollowed out. And what is especially important for us: studies show that young people — those just entering the labor market — face the highest risk of automation. In other words, today’s schoolchildren.

The third scenario is structural restructuring. This is the baseline scenario on which governments and investors tend to rely. It is neither catastrophic nor benign: some jobs disappear, others emerge, and a major redistribution of work between humans and machines takes place. According to the World Economic Forum’s Future of Jobs Report 2025, by 2030 around 170 million new jobs will appear worldwide while about 92 million will be displaced — a net increase of 78 million. Demand is rising for data analysts, AI specialists, cybersecurity experts, and digital transformation professionals. Demand is falling for those whose work is easiest to automate: data-entry operators, routine-level accountants, secretaries, and clerks.

The most important point about these three scenarios is not to guess the “right” one. The main conclusion is the opposite: it cannot be guessed, and there is no need to try. Scenario uncertainty is ceasing to be a temporary condition and is becoming a permanent feature of the environment. This means education cannot be built around one pre-known version of the future. Children cannot be prepared “for professions that will be in demand,” because we do not know which of them will survive the next ten years. They can be prepared only for one thing: the ability to act when the rules are changing in real time. This is a 180-degree turn away from the logic by which school has always operated.

Behind all three scenarios lies one common shift, deeper than the labor market. The very role of knowledge in society is changing.

For centuries, knowledge was a scarce and valuable resource. School provided access to it, a diploma certified it, and a career monetized it. The entire social machine — exams, certificates, competitions, interviews — was built around measuring, verifying, and rewarding someone’s knowledge. This worked as long as knowledge was difficult to obtain.

Artificial intelligence breaks this structure at its foundation. When any fact, any explanation, any literate text, or working code can be obtained in seconds, the ability to reproduce knowledge ceases to be rare — and therefore ceases to be a value for which society is willing to pay. Knowledge becomes something like cognitive infrastructure: everyone has it, like electricity from an outlet. What becomes valuable is what cannot be downloaded — the ability to ask the right question, choose between alternatives, assume responsibility, distinguish truth from plausible nonsense, and live through and make sense of experience. In the first material, we called this “wisdom,” and we will return to it more than once.

This idea was sharply formulated by one coach who works with senior executives of leading technology companies: answers can now be obtained from AI, but to use them wisely one needs wisdom; and wisdom is the ability to live, an alloy of lived mistakes and reflection that cannot be acquired hastily and cannot be copied. If knowledge becomes free and instant, then the scarce resource is precisely what a machine cannot provide: lived, felt, and interpreted human experience. This is a cornerstone of the future school.

For school, this means an earthquake in the foundations. If its main historical function — transmitting scarce knowledge — is being devalued, then a question arises that must be answered honestly: what is school for at all? It cannot and should not compete with an algorithm in providing information. Its value must therefore lie elsewhere.

It is useful here to distinguish two fundamentally different stages in how AI enters education — and to understand which stage we are in.

The first may be called the stage of “playing with AI,” and most of the world lived through it in 2023–2025. At this stage, AI is an “add-on” placed on top of familiar school. The same curriculum, the same lessons, the same exams — only here and there a chatbot was added, here and there automated checking, here and there attractive analytics. School does not change in substance; it is merely decorated with technology. This is a useful but superficial stage.

The second stage — the one the world is entering roughly now, in 2026–2028 — is “structural restructuring.” AI ceases to be decoration and becomes an end-to-end process that changes the very architecture of learning. This is not about a single tool, but about an entire range of directions unfolding simultaneously: adaptive educational trajectories tailored to each student; generation of learning content on the fly; predictive analytics that anticipates where a child will stumble before it happens; project-based learning in peer environments; educational simulators and game worlds; and, finally, neuroscience and neurotechnologies entering education from an entirely new direction.

The difference between the two stages is fundamental. At the first stage, AI could be treated as an optional addition — switch it on if desired, switch it off if desired. At the second stage, AI becomes an environment in which the economy, society, and school all exist. This environment can no longer be switched off; one can only learn to inhabit it meaningfully. That is why the wait-and-see strategy with which we began does not work: waiting out structural restructuring while remaining at the add-on stage means preparing children for a world that will no longer exist.

This turn has a consequence that many fear most: if AI can explain, check, and select assignments, does that mean the teacher is no longer needed?

The answer is paradoxical: the teacher becomes more needed — but in a different role. When knowledge was scarce, the teacher’s main asset was superiority in knowledge: the teacher knew what students did not, and authority rested on that. In a world where knowledge is available to everyone through an algorithm, that superiority is devalued. But space opens for three things that an algorithm does not do and should not do.

First is motivation. Making a child want to learn, igniting interest, supporting them when their hands drop — this is the work of a living person, not a program. Second is navigation. In a world where information is infinite and reliable beacons are scarce, the key ability is to chart a route: what matters and what is noise, where to move, what to learn next. Third is facilitation. Organizing live interaction among children, teaching them to argue, agree, work together, and form friendships — this cannot be learned from a chatbot.

What children themselves say is telling. In one study of children’s voices, schoolchildren formulated their request to teachers with striking simplicity: “We want teachers to respect us and become our friends.” Not “deliver knowledge” — respect and friendship. This is the human role that cannot be automated and that, in the age of AI, becomes not peripheral but central. The teacher ceases to be a source of knowledge and becomes a motivator, navigator, and facilitator. This is not a downgrading of the profession’s status; it is its rebirth.

There is another institution cracking at the seams: assessment. The entire school assessment system is built on checking knowledge — how well a student has reproduced what they were supposed to learn. But if knowledge can now be reproduced by an algorithm, and better than any schoolchild can, what is the point of checking precisely that?

There is an old cartoon that describes the problem mercilessly well. Animals — a monkey, an elephant, a fish in a bowl, a penguin, a seal — stand before an examiner, who announces: “For fairness, everyone will take the same exam: climb that tree.” It is clear who will fail: the fish and the elephant, however hard they try, even though they are excellent in their own environments. A single exam based on a single standard measures not a child’s abilities, but their conformity to one narrow norm.

Researchers have long shown that the “average person” for whom these standards are designed simply does not exist. If any of us is measured across many parameters, it turns out that we are “jagged profiles”: somewhere far above average, somewhere far below, and almost no one is “average” in everything at once. Standardized assessment, which levels everyone to one bar, loses meaning in this logic. Something else comes in its place: personal educational trajectories, portfolios instead of one-time exams, assessment of projects, argumentation, and the course of thought rather than a final answer that can in any case be generated. In the new world, what needs to be assessed is not what a child has memorized, but how they think, how they learn, and who they are becoming.

If all this is brought together, a picture emerges that can be frightening or inspiring, depending on how one looks at it.

The world is entering an era in which familiar beacons and reference points no longer exist. The falling cost of computing has launched a process that will not stop. The labor market is already shaking, and it is shaking first the professions for which children go to school and university. No one — neither governments, nor corporations, nor experts — knows for certain what the world will look like in ten years. In this sense, everyone today is in the same position: ministers, teachers, parents, and children themselves have all, as it were, “started first grade” together — learning to live in a world no one has yet seen.

And here lies the good news. Since no one has ready answers, the winner is not the one with more technology or money, but the one who more quickly and honestly understands what is happening and builds a meaningful trajectory. In this sense, Kazakhstan is not lagging behind — it is in the same conditions as the rest of the world. And the Year of Digitalization and Artificial Intelligence is a chance not to catch up with other people’s solutions, but to define our own from the start.

But to define our own, we need to know what those directly affected by this think: teachers, parents, principals, methodologists, and students themselves. We gathered precisely their voices in a large study, the results of which are discussed in the next material of the series. The global picture sets the horizon. How this horizon is seen from a Kazakhstani classroom is a separate and much more concrete conversation.

The material was prepared based on the results of foresight sessions held in autumn 2025 and winter 2026.