ARTIFICIAL INTELLIGENCE – REALITY OR ILLUSION?

What if, in truth, AI companies are simply playing with words? What if the term “intelligence” is being used to designate something that ordinary people would otherwise describe quite differently?

Our Thesis

Ordinary people and scientists do not attach the same meaning to the word “intelligence”. Moreover, even among scientists there is no consensus: its definition shifts depending on whether one adopts a “human” or a “formal” profile. Companies producing advanced software, however, are not troubled by such subtleties; they simply declare their programs to be AI and proceed to celebrate them. Yet is this truly intelligence? And if so, what kind of intelligence are we speaking of? This is what we aim to decipher.

Our Categories of “Intelligence”

A disclaimer is necessary: we will not reproduce the standard dictionary definitions of intelligence nor the categories listed in scientific works. We assume full responsibility for departing from established conventions, as this is an opinion piece and not a scientific article. Indeed, we could hardly hope to compete with scientists who have, of late, become more skilled with words than the finest lawyers in Paris. Thus, our categories of intelligence are as follows: Mathematical or Logical-Mathematical Intelligence (MI): grounded in calculation, algorithms, probabilities—in short, mathematics. Human Intelligence (HI): comprising two essential dimensions: Sensory Intelligence, which employs information from the five senses to guide actions and reactions necessary for survival, protection, and reproduction of the species. Commonly referred to as the “survival instinct,” we regard it instead as a form of intelligence, since it is not purely mechanical or automatic but belongs to our nature as living beings. Emotional Intelligence, based on feelings, experiences, and thoughts. It shapes the rules of communal living—law, morality, etiquette—within societies, communities, and relationships. Together, these two dimensions constitute what we call Human Intelligence (HI).

What Kind of “Intelligence” Is AI?

Let us be clear: the game was rigged from the start. According to Wikipedia, John McCarthy coined the term artificial intelligence, defining it as akin to human intelligence but freed from biological limits. Marvin Lee Minsky, one of AI’s pioneers, was even more direct: “AI must eventually do everything that humans can do, only better.” From such premises we arrive at today’s situation, where AI is marketed as a magical tool capable of solving any problem. This is an illusion. In reality, AI is a form of Mathematical Intelligence (MI)—at best, a pale imitation of Human Intelligence (HI). It cannot do what HI does, and still less can it replace it, except superficially.

Why AI Cannot Replace Human Intelligence (HI)

Stripped of the advertising veneer, we see that AI is nothing more than mathematics. Emotions, instincts, and the vital dimension of life remain inaccessible to it, since it has neither body nor life. This constitutes the fundamental difference from HI: while AI can store immense quantities of knowledge, it does not know what to do with them independently. It requires human input first to learn how to use its resources and then for fine adjustments. As for genuine creation, this too is possible only after humans have already done most of the work.

Human Labour Behind AI

Behind AI’s increasingly “spectacular” achievements lies the labour of tens of thousands of poorly paid workers from impoverished countries—so-called “click workers.” Their work produces the standardized responses from which AI selects the one most suited to our queries. For more complex tasks, companies rely on computer science students who design algorithms to help AI recognize and simulate emotions in text, images, and other media. AI companies claim that once this “training” phase is complete, the systems evolve autonomously, generating increasingly refined content. But this is misleading: it is akin to a marathon runner who rides in a car until the 38th kilometre and then claims victory.

What AI Actually Does

Generative AI—those accessible to the public—operates by identifying the most probable “standard response” to a request and then adding a marginal plus, based on data collected from our online behaviour. Over time, it tailors responses more closely to individual users. Private AI, used by states and corporations, functions differently, being employed in standardized processes where unpredictability is minimal. There, AI excels through sheer computational power, but creation and invention remain beyond its reach.

The Limits of AI Compared to Human Intelligence

AI can reproduce typical human behaviour, provided it follows expected patterns. Yet it cannot anticipate or comprehend when a person suddenly changes behaviour, acts intuitively, or draws upon the so-called “sixth sense.” Being a mathematical intelligence, it cannot reproduce HI in its full complexity.

Human Intelligence

Each of us has a personal vision of what it means to be “intelligent,” often mixing mathematical abilities, emotional capacities, and even cunning—the skill of bending rules to one’s advantage. AI cannot replicate this amalgam, for it is neither predictable nor logical. Nevertheless, companies continue to present their products as if such replication were imminent.

Reality and Illusion

AI can replace humans only in domains requiring Mathematical Intelligence (MI). Its performance will expand massively in the coming years, thanks to stronger hardware, better software, and increasingly skilled human trainers. Within this framework, AI does meet the conditions for being considered a kind of intelligence. But beyond mathematical logic, its “intelligence” becomes an illusion—a rhetorical trick designed to make us dependent and willing to pay for it.

AI as an Instrument

AI should be understood as a tool: increasingly necessary, but dangerous if used against us. It is not infallible and can always be outwitted by human actions that defy routine or logic. Thus, the prudent stance is to use AI to our advantage without forgetting that the reverse is also possible.

And the Future?

Intense development of AI continues worldwide, led by the United States and, to a lesser extent, Europe. Rival states, however, exploit public fears to advocate slowing development in the West, while simultaneously advancing their own programs.

Most probable scenario (65–70%): Global South countries, leveraging powerful AI, surpass the West technologically and impose their dominance.

Less probable scenario (25–30%): These countries develop AI but without overtaking the West.

Highly improbable scenario (0–5%): Their efforts stall, and AI remains an unfinished project.