Artificial Intelligence and the Brain: Two Ways of Thinking

We live in an era when machines begin to think. No, they do not feel or experience, but they can write poems, diagnose diseases, drive cars, and even conduct a dialogue that is almost indistinguishable from human. Artificial intelligence has burst into our lives and made us ponder: what makes us human? What distinguishes our brain from a neural network? And is there anything common between them, besides the word \"neural\"? World Brain Day is the perfect occasion to delve deeper into this issue and try to understand where biology ends and code begins.

Architecture: biological chaos versus mathematical order

The first and main distinction is how both \"processors\" are structured. The human brain is the result of millions of years of evolution. It is not designed but grows like a living organism. Its neural networks are not perfect: they are noisy, slow, subject to fatigue, injury, and aging. But it is precisely this imperfection that makes it flexible. The brain can learn from a single example, it is capable of generalizations, and it knows how to transfer skills from one area to another. It is a living system that constantly restructures under the influence of experience.

On the other hand, artificial intelligence is created by engineers. Its neural networks are mathematical models operating on digital carriers. They are accurate, fast, and predictable. They do not get tired and do not get sick. But they cannot go beyond the data on which they have been trained. They do not understand the context unless it was encoded in the training. Their \"flexibility\" is merely the ability to iterate through billions of combinations but not create new principles of thinking.

The comparison here resembles the difference between a living tree and its 3D model. The model is beautiful and accurate, but it does not grow and does not bear fruit. The tree is chaotic, unpredictable, but it lives.

Learning: experience versus data

Humans learn through interaction with the world. A baby does not receive annotated data - it stabs, tries, falls, cries, and builds models of the world based on this chaos. Its learning is continuous, without a teacher, in conditions of uncertainty. The brain learns all its life, and each new experience changes its structure. It does not require billions of examples to recognize a cat - it is enough to see it a few times in different angles.

Artificial intelligence learns on massive amounts of data. To teach a neural network to distinguish a cat from a dog, it needs thousands, sometimes even millions, of labeled images. It does not \"understand\" what a cat is - it simply finds statistical patterns in pixels. Its learning is the optimization of the error function, not the formation of an internal model of the world. It does not know that a cat meows and catches mice - it knows only that there is a certain correlation between the shape of the ears and the label \"cat\".

Moreover, AI does not transfer knowledge from one field to another as naturally as a human. A neural network trained to play chess cannot play Go without retraining. A person, on the other hand, can apply chess logic to planning a route or to life strategy. This property is called \"generalization,\" and it remains a biological privilege so far.

Consciousness and emotions: where the spark

This is the main distinction that cannot be overcome. Humans not only process information but also experience it. They have feelings, intentions, desires, fears. They can get bored, be happy, sad. They are able to be aware of themselves, ask questions about the meaning of life, worry about the future. This is called phenomenal consciousness or qualia. We do not know how it arises from neural activity, but we know that it is not present in AI.

Artificial intelligence is an algorithm. It can imitate emotions, respond in a rhetoric that seems empathetic, but inside it there are no experiences or subjective experiences. It does not know what pain, sadness, or joy are. It does not choose where to direct attention - it reacts to a request. Its \"curiosity\" is just the search for information based on given criteria. Its \"creativity\" is just the combinatorics of known elements.

Consciousness makes us vulnerable, but it also makes us human. It is precisely what allows us to love, doubt, dream. And as long as we do not know how to recreate this in silicon, we remain the only creatures capable of asking questions about the meaning of our existence.

Similarities: common principles of operation

Despite all the differences, the brain and AI have important similarities. Both are information processing systems. Both use parallel data processing: neurons in the brain work simultaneously, as well as layers of neural networks. Both learn through reinforcement and error correction. The principle of backpropagation of error in AI was inspired by ideas about how the brain regulates its connections. And in both cases, information is transmitted through excitation and inhibition (in the brain - chemical, in AI - numerical).

Moreover, both the brain and neural networks are effective in recognizing images. They can find patterns in noise, classify objects, predict sequences. And both systems can \"remember\" information, although the mechanisms of memory are fundamentally different (synaptic plasticity versus weight coefficients). Both systems can make mistakes and both need \"rest\" - the brain in sleep, AI in breaks for retraining.

It is also important that both the brain and neural networks are built from a multitude of simple elements working together. In this sense, they are examples of \"emergent\" intelligence, where complex behavior arises from the interaction of simple parts. This similarity has given a boost to the development of the entire neurosciences, because AI has become not only a tool but also a model for understanding the brain.

Boundary: what will remain behind man

Today AI surpasses us in solving narrow tasks: it calculates faster, plays chess better, translates texts more accurately. But it cannot make decisions in uncertain conditions without data. It cannot adapt to a completely new situation without retraining. It does not have intuition, which in humans is based on many years of experience and subconscious signals from the body.

The boundary between man and machine runs not along the level of intelligence, but along the way of existence. We live, we suffer, we create meanings. Artificial intelligence is a tool. Powerful, useful, sometimes frightening, but a tool. And the best we can do is to use it to expand our capabilities, but not to forget that true wisdom, creativity, and freedom remain with us. World Brain Day is not a day to fight AI, but a day to understand ourselves.
© elib.ng

Permanent link to this publication: