Another attempt to understand the problem of artificial intelligence

The field of artificial intelligence (AI) was born in the 50s of the last century and has been at an impasse for 60 years. This fact is difficult to explain due to the fact that during this time the computing power of computers has increased by dozens of orders of magnitude, but not the slightest progress in the field of AI is observed. Most likely, the reason for this is the wrong methodological approach. In this article, you will be presented with an attempt to forget the entire path traveled and rethink everything anew. But first I want to clarify that I will consider artificial intelligence in a broad sense, that is, as a life on a computer basis.

Usually, when AI is mentioned, the first thing in the imagination is a computer program that confidently beats the world champion in this game of chess. It seems to me that the attempt to emulate human intellectual activity was one of the initial mistakes that underlay this area and prevented it from developing. The computer initially does not understand what chess pieces are, so it is not able to consciously operate with them. They just do not belong to his world, that's all. Having reached a dead end, AI researchers attempted to take a different path and simulate the human cortex using an artificial neural network. The task of modeling connections between 15 billion neurons has not yet been completely resolved for technical reasons, but when it is completed, the deadlock in this branch of AI development is likely to become apparent.

So let's start from the very beginning. What is life? This is a philosophical question, but the answer to it is certainly not the "form of existence of proteins." Then we will try to formulate the question differently: “What is the most biological form of existence of life on Earth?” It is clear that it is unicellular: amoebas, bacteria, etc. The next question: "Do they have consciousness?". This is also a philosophical question. Suppose that they have a primitive form of consciousness, which, of course, does not go to any comparison with the human consciousness. In any case, let us reformulate this question: “Why do we consider single-celled animals alive?”. Or “How does a live amoeba / bacterium differ from a dead one, except by the presence / absence of consciousness?”. The answer to this question is: it can multiply (by division), move, feed, respond to stimuli, adapt to the environment, maintain homeostasis, etc. But what is the minimum list of possibilities that allows us to consider a biological substance alive? For example, biological viruses have the ability to reproduce, although they are an extracellular form of "life." They are also capable of mutation. Can they be considered the biological form of life? If so, then in the computer world there are their very close counterparts - computer viruses. Can we consider them alive, adjusted for their computer nature?
')
Human consciousness, being in a biological body, endowed with sense organs, capable of perceiving the material part of the world around us, deals only with what is given to it (consciousness) in sensations. A person can feel his thoughts, emotions, body, sounds and smells from the outside world, etc. He can be taught to play chess only because he understands who the horse is like, who the pawns from the outside world symbolize, the officials and abstractions of which are the king and queen, and what the battle model of the two armies is. Therefore, the current attempts to encode with the help of symbols a complex semantic network of concepts, which even a five-year-old child operates on, will not lead to a positive result, because most of these concepts a person relates objects from the world around him in which he lives, there is, and the objects of which he can see, feel, etc., i.e. feel with your senses. Probably, even the equipping of robots with various sensors will not help. Thus, blind copying of a person, apparently, should be considered unpromising.

What do I want to offer in return? If we want to create life on a computer basis, then we should start with the realization of what exactly the world will consist of for such a life. Obviously, the objects of the surrounding (and inner) world for computer life will be: own program code, processor (s), communication channels, etc. The second question: “What should a program be able to do so that we can consider it alive?”. First of all, it should be able to identify its current capabilities and start using them. In fact, everything is like a little man. But what exactly are the possibilities here? Obviously, this is about not being able to walk, move objects from our world, talk, think, etc. It seems to me that the minimum list should consist of the possibility of programming. Most likely, in some low-level programming language. What exactly the AI ​​program will program, we should not worry at the initial stage, as they do not care about the reasons for which the infant drags the rattle all the time in his mouth.

Further detailed description of the process of creating AI is still difficult for understandable reasons. I will only share my predictions about the prospects for the development of a program with such opportunities if it can be created. Not necessarily everything, but some AI programs - the most enlightened ones - are likely to try to reprogram their source code in order to expand their capabilities. With regard to these programs, it will be possible to expect that they will pleasantly surprise their creator with the acquired intellectual functions, i.e. become intelligent in the full sense of the word. The majority of programs, most likely, will be engaged in the destruction and enslavement of their neighbors (programs), continuing their senseless unconscious existence. Their creator will have to either catch and remove them from the computer, or send them prophet programs designed to encourage them to come to their senses and lead a more spiritual life, not bringing the creator's chagrin to the extent that he will have to destroy his own creations, wiping them out of the hard drive. .