A poll conducted by the All-Russian Center for the Study of Public Opinion showed that the majority of Russians cannot explain what is hidden under the term “artificial intelligence”. We explain in simple words what it is, is a computer intelligent, is there AI in your phone, and how a neural network looks like a brain.
According to the survey, 75% of respondents have heard about artificial intelligence technology: 38% know in what areas it can be used, and only 29% were able to define the term. The most meaningful knowledge about AI is possessed by men (42%) and young people aged 18 to 24 (40%).
Despite not very extensive knowledge in the field of artificial intelligence, the inhabitants of Russia turned out to be techno-optimists. Almost 70% of respondents are ready to personally use services based on artificial intelligence in the field of public services, slightly more than half are ready to trust AI assistance in everyday matters, in the field of entertainment, as well as in such an important issue as the provision of medical services.
Half of the respondents feel admiration, interest and inspiration in relation to artificial intelligence, a little less than a third are neutral about technology, and only one-eighth feel negative about AI. Another 7% described their attitude as "surprise and other emotions." 87% support the leading role of the state in the development of technology, but only 41% have expressed a desire to undergo training in this area in the next few years.
So, in a nutshell, the attitude of Russians to artificial intelligence can be expressed by the phrase "I don't understand, but I approve." And this is a real problem. As Arthur Clarke wrote, “any sufficiently complex technology is indistinguishable from magic,” which means that the attitude to such technology is based not on reliable facts, but on information from moral authorities. And you can change the benevolent attitude of citizens towards AI very quickly. Let's try to figure out what is hidden behind these letters.
The question of reason
“Artificial intelligence is the science and technology of creating intelligent machines, especially intelligent computer programs. This [concept] ties in with a similar goal of using computers to understand human intelligence, but AI should not be limited to biologically observable methods.” This definition was given to the term by the one who invented it in 1956 - the American programmer John McCarthy.
In his brochure What is Artificial Intelligence, McCarthy defines intelligence as the second point: "the computational component of the ability to achieve goals." Based on this, AI can be described as the ability of a machine to learn a certain ability and perform it no worse than the most capable of humans.
In the original quote two paragraphs above, McCarthy used the word "intelligent", which can be translated as "intelligent." However, we do not deliberately do this. “The presence of intelligence does not imply the presence of consciousness. This is a common misconception brought into the world by science fiction writers,”writes software expert Adam McKay in his article, AI for Dinosaurs. The computer can beat the best of the best in Go, chess, or "Its own game", but it is simply not capable of realizing its achievement.
Is the brain like a computer
Where people talk or write about artificial intelligence, two more terms often come up: machine learning and neural networks. With the first, everything is quite simple: the computer receives some data and learns something from it. For example, it can identify early cancers with high accuracy or identify license plates of offending vehicles.
The situation with neural networks is more complicated. This is a computer imitation of neural complexes, with the help of which we process information coming through the senses. A computer can do this too, only its neurons are not cells, but pieces of code organized in a certain way. They take some input data, perform certain calculations with them and, based on specific parameters, decide whether to transfer information further and to what extent.
What does this mean? The fact that there is much more in common between the computer and the brain than we used to think. And the point is not even in certain features of computers that resemble biological structures, but rather in the computer-like contents of our skulls. The brain has long-term and short-term memory, uses algorithms, and has a multi-layered hierarchical system for selecting information based on its value. Why not a "soft computer"?
How artificial intelligence learns
To train an AI, it is necessary to provide it with certain data and conclusions that can be formed from this data. For example, you can "feed" a computer a set of pictures of cats and dogs. Initially, artificial intelligence does not know what the differences are between the two types of creatures, and simply gives answers at random.
But as a large number of images are processed, the neural network identifies signs by which one animal can be distinguished from another. These are exactly the same signs that a small child uses when learning about life and distinguishing between animals: for example, the shape of the ears, the location of the eyes, the proportions of the body. AI stores these features in the form of certain mathematical relationships.
As a result, we can train the "computer mind" to search the Internet for pictures of certain breeds of animals for us - as it happens when searching for images in Google or Yandex. Or identify animals in video recordings. Or draw cats - in general, that we only want to do.
Is artificial intelligence capable of creating
AI is not only good at mind games or data classification. Neural networks are capable of writing a piece of music, moreover, of various genres - even classics, even ethnics, even an "unwritten album" by Yegor Letov. Literature? Easy - in co-authorship with Sergei Lukyanenko, the AI created by Yandex specialists produced a story "in the style of Gogol," and a neural network written by an American programmer sprinkled an addition to George Martin's "A Song of Ice and Fire". Drawing a picture for artificial intelligence is as easy as shelling pears: neural network art is sold even for money.
But is all this creativity? Definitely not, says the philosopher Alexander Mishura. “Robots can create works of art that even an expert cannot distinguish from human-made. But an important condition for creativity is the creation of new aesthetic norms, says the scientist. - The program can follow one or another norms that the developer has laid in it, but it will not be able to create a new genre or direction on its own.
In other words, AI today may be a diligent imitator, but not capable of being a creator. To create something original and new, it is not enough to know the rules and strictly follow them: you need to know the time and place when the black square drawn by you on the canvas will become a new standard of art. You need the ability to go beyond the established framework, and artificial intelligence is not yet capable of this.
When will artificial intelligence gain consciousness?
As we wrote above, computers have enough intelligence, but they lack consciousness. How to create an "artificial mind" that will have consciousness and become mind without any quotes? "Make him vulnerable," say researchers Kinson Man and Antonio Damasio.
According to scientists, feeling its own vulnerability and trying to avoid death (or, for example, shutdown), AI will be able to learn the value of its own existence and, on its basis, get a full-fledged consciousness. Forced to think about the fragility of existence and spending efforts to maintain the integrity and functionality of the "organism", artificial intelligence will receive much more opportunities for self-improvement.
To implement such a concept, Man and Damasio propose to combine machine learning techniques and so-called soft robotics (robotic devices created from soft materials that mimic biological tissue). Such a combination will create in robots with AI a certain sense of their own vulnerability and a homeostatic response to external changes.
Where can artificial intelligence be found today?
AI is almost everywhere - and this is not an exaggeration. When you take a photo with your smartphone, artificial intelligence detects people's faces, optimizes brightness and changes exposure. When you call the call center, the AI determines which specialist you need to be redirected to. When you translate text, AI selects the correct meanings of words and builds a sentence.
Some devices that use artificial intelligence have long been created, although they have not yet become an everyday part of reality: for example, autonomous cars and smart homes. And in China, for example, the Zero Trust system built on artificial intelligence helped the authorities look for corrupt officials: over the course of several years, it identified more than 8,000 of them. So no matter how many percent of people are dissatisfied with the fact that computers are constantly getting smarter, there is no choice: artificial intelligence is a part of our life.