AI Inequality: From Access to Authority

Miskola methodology · Essay · Updated September 2026

What is AI inequality? AI inequality is not only about who can pay for a subscription. It is the deeper gap in access to knowledge, time, automation and authority — who decides which tools exist at all, and who had the chance to learn to think before AI arrived. An essay by Gatis Šeršņevs (Miskola, Latvia).

When a particularly powerful and expensive AI model appears, the first reaction is often simple: “that is a tool for the rich”. True — but not enough. AI inequality is not only the difference between people who can and cannot pay for a subscription. It is a difference in access to knowledge, time, automation, and in the power to decide which tools exist at all.

But there is a deeper layer. I am writing this text together with AI. I do not know every sentence from A to Z. But I choose the topic, I provide the context, I read, criticise and reshape. The final decision about what stays and what goes is mine. That is not less than authority. It is a different kind of authorship.

And here the real question begins: what happens to children who have never learned to write, think and argue without AI? They become operators, not authors. So my instinct is: no AI until grade 5. But I have my doubts. Perhaps the grade is not what matters. Perhaps what matters more is whether the child has acquired the base competence before AI arrives.

And here another layer appears — language. Yes, AI models are trained mainly in English. But translation keeps getting better. It is a bridge, not a wall. A student in a Latvian school can read an English study or follow a video lecture from a university in another country. That reduces inequality rather than increasing it. But only if the student can already think in their own language.

And then there is one more factor — the age of teachers. The average teacher in Latvia is around 50. AI is developing at cosmic speed. The result is a huge gap between what AI can do and what teachers can actually use in class. This is not simply “everyone must learn AI”. It is a systemic challenge.

And here I arrive at my hypothesis: the country that invests in AI education now — in teachers, in students, in infrastructure — will move ahead of the others. It will not be only an economic advantage. It will be the ability to create, think and innovate. It will be the ability to keep authority over its own thinking.

AI is not going to slow down. And the sooner we start talking about it, the greater the chance that AI becomes a tool that amplifies people, rather than a system that replaces them.

But many questions still have no answers:

Who is the author when the text is generated by AI and edited by a human?
From what point is a child ready to use AI — not as a replacement, but as a tool?
How do we preserve linguistic and cultural diversity when AI mainly speaks the big languages?
Who owns AI-generated content — the user, the company, or is it a shared resource?
What remains uniquely human when AI can write, draw, compose and argue?
How should education respond — prepare people for the labour market, or raise people who keep authority over their own thinking?

The neurobiological layer: what does research show?

When we ask whether a child is “ready” to use AI, we are talking about neurobiology. And here research is starting to give the first answers — they are not comforting.

A European Parliament study (2026) warns that using AI at school creates four cognitive risks: over-reliance, erosion of basic skills, loss of metacognition, and attention problems. Children and adolescents are especially vulnerable, because working memory, executive functions and critical thinking are still developing through adolescence. A student who delegates cognitive work to AI during the school years does not merely “learn worse” — they may fail to build the cognitive architecture on which all later learning rests.

Neuroimaging studies show that higher screen time (ages 0–12) is associated with reduced cortical thickness, lower grey and white matter integrity, and weaker connectivity in language and attention networks. This is not just “reading less”. It is a question of brain structure.

A 2026 study on generative AI and childhood education speaks of “intellectual deskilling” — children lose the ability to think for themselves because AI “does the work” in their place. A Brookings report (2026) calls it a “doom loop”: a circular dependency in which students offload more and more of their thinking onto AI, and as a result knowledge, critical thinking and even creativity decline.

But there is counter-evidence too. A Finnish study (2026) found that children with more screen time in childhood showed better cognitive processing in adolescence — provided there was a balance between physical activity and screen content that supports learning and creativity. That means: not “screens are bad”, but “which content, and in what context”.

And here we reach an open question: is AI the “content” that supports thinking — or the “tool” that replaces it?

Live communication: the foundation of education, not an add-on

There is one more layer that research cannot measure — live communication. Class evenings, field trips, summer jobs, being together, face-to-face conversations, shared experience. That is not an “add-on” to education. That is education.

When a child experiences that a conversation with another person takes time, patience and empathy; that an answer does not appear in two seconds; that silence is also part of a conversation — that is the moment they learn something AI cannot teach. They learn to be a human among humans.

AI can help write a text, find information, organise ideas. But AI cannot replace what happens when a class solves a problem together, when students argue, agree, make mistakes, laugh, fall silent. That is not “content”. It is a process that shapes a person.

And here is my instinct: if we want children to grow into authors rather than operators, we must not forget live communication. Because AI can give an answer. But only a live conversation can teach why the question matters.

So my vision of education is not “AI or no AI”. It is: live communication as the foundation, AI as a tool. Not the other way round.

A personal note: when AI becomes a bridge, not a prosthesis

I have cerebral stroke — including clumsy fingers. In the past I made many grammatical mistakes in my writing and my thinking. Not because I could not think, but because the physical skill of writing stood between me and what I wanted to say.

Now? Uhhhh — look how far I can reach! AI helps me organise a text, find the right word, build sentences so that the thought sounds clear. But the thought is mine. The content is mine. The authority is mine.

And here is the key: AI has not replaced my thinking. It has freed my thinking from a physical obstacle.

But if a person with no physical barriers at all gets used, from the very start, to AI thinking in their place — are we not creating a new dependency? Are we not turning a bridge into a prosthesis?

So my instinct is: before AI — learn to think without it. Let AI become wings, not a wheelchair.

Conclusion: what is at stake?

If I pull all these layers together — access, authority, neurobiology, live communication, my own experience — one thought remains.

AI in itself is neither a friend nor an enemy. It is a mirror that magnifies whatever is already there. If a child already has the base competence to think, read and argue, AI becomes wings. If not, AI becomes a wheelchair that never lets the legs firm up.

So the inequality I began this text with is not only a question of who has a subscription and who does not. It is deeper: who had the chance to learn to think before AI — and who did not. And that inequality does not look only at a bank account. It looks at school, family, language, country.

So the country that invests first in teachers, in time for live conversation and in base skills, and only then in AI infrastructure — not the other way round — will be the one that preserves the next generation’s authority over its own thinking. A country that starts with the tool before the base risks raising a generation of operators.

To most of the questions I listed earlier I still have no answers. But I have one conviction: the answer is not “more AI” or “less AI”. The answer is — first the human, then the tool. First the live conversation, then the algorithm. First the ability to think without a helper, then the helper that lets you think further.

That is a process without an end date. But it is a process we can start to steer — instead of only watching from the side.

This text was created together with AI. I chose the topic, provided the context, criticised and reshaped it. The final decision was mine.