The impact of artificial intelligence on hr business processes in russian companies

Размещено на http://www.allbest.ru/

FEDERAL STATE EDUCATIONAL INSTITUTION OF HIGHER EDUCATION NATIONAL RESEARCH UNIVERSITY HIGHER SCHOOL OF ECONOMICS

Saint Petersburg School of Economics and Management

The impact of artificial intelligence on hr business processes in Russian companies

Bachelor's thesis

Silaeva Maria Dmitrievna, Sokolova Varvara Alekseevna

Abstract

The impact of artificial intelligence on people's lives in the last two decades is constantly growing and is expected to become even more global in the near future (Koro N., Karpova S., Burukina O., Vyatkina N. & Pavlov S., 2019). The same impact is seen in companies' business processes and in HR sphere in particular. The goal of the study is to extract the correlation between opinions of HR experts and employees about automation of the hiring processes. The results can be useful for companies, which are going to implement artificial intelligence in HR business processes to clarify the stages from which the implementation can be started with the smallest resistance from the society. Considering the methodology, the authors will conduct 15 in-depth interviews with HR experts. The interviews are needed to find out the main stages of HR business processes in different companies. After the interviews, the survey will be offered to HR experts. At the same time, the authors will provide the same survey to the group of people, who are in search of work. The surveys are needed to evaluate the readiness of people to be assessed by artificial intelligence technologies and the readiness of HR experts to give some stages of hiring to artificial intelligence. The authors found that HR experts are more likely to accept the automation. Moreover, the authors identified the main stages of hiring process in companies, ranked them by the least resistance to the introduction of artificial intelligence from the society, and determined the attitude of HR experts and potential employees to the use of artificial intelligence technologies in HR business processes.

Key words: artificial intelligence, HR business processes, hiring process automation, artificial intelligence implementation.

Introduction

According to Lee V. (2019), the artificial intelligence is one of the most perspective tools for improving the productivity of companies. At the moment, artificial intelligence is at the stage of active development, and it has already been successfully used in various companies in many industries and fields of activity. This research will describe the possibility and prospects of the application of artificial intelligence in the field of HR in Russian companies.

In 2018, the consulting company Deloitte included artificial intelligence and robotics in the list of top HR trends. “E&Y” analysts also confirmed their' conclusions: according to their data, 93% of the recruiter's time is spent on typical tasks, and about 65% of them are quite amenable to automation.

1. Research gap and research purpose

artificial intelligence management

Artificial intelligence helps recruiters remove routine processes from their daily tasks and focus on communicating with candidates. Therefore, often the technology is used for mass hiring, when dozens of responses come to one vacancy at once.

For example, in the “MGTS” company, the robot searched for candidates for mass vacancies in client databases according to the specified criteria and called them to find out their interest in the work. In total, the artificial intelligence processed 1,788 resumes and the company managed to hire five necessary employees in one week.

“Neftmagistral” fuel company has implemented an automated recruitment system based on the domestic AI-platform Sever.AI and automated other part of hiring process. The robot recruiter in this company is able to synthesize text into voice and back thanks to Speech-to-text technologies, as well as recognize speech. They can also talk about the job, ask questions and answer the other person, including an incoming call, while simultaneously continuing to search for candidates. The robot independently finds out the name, age, place of birth and the vacancy that interests the caller. Then, after analyzing the information received, it makes an appointment with the appropriate candidate, calculating the most convenient gas station for the applicant by geolocation, explained CNews in Talenttech.

There are other companies, which has implemented the artificial intelligence technologies in HR business processes and the process of hiring in particular. Nevertheless, there are no Russian companies, which automated the hiring process entirely. What are the reasons for that? As the authors suppose, the possibility of full automatization of hiring process and trusting it to the artificial intelligence technologies lies in the gap between the HR-managers opinion and the attitude to it of the employees.

2. Previous research

In the last 5-7 years much research in the HR sphere was focused on the challenges while implementing the artificial intelligence and possible solutions to them. According to Bersin J. (2018), in recruiting many decisions are made based on the motive “I feel that way”. One study found that most recruiters make a decision about a candidate within the first 60 seconds of meeting them, often based on a look, handshake, clothing, or speech. Managers and HR professionals use estimates, tests, simulations, and games worth millions of dollars to hire people, but many say they still make mistakes in 30 to 40% of their candidates.

Algorithms based on artificial intelligence can select resumes, find good internal candidates, profile highly effective employees, and even decode video interviews and give information about those who are likely to succeed.

HeadHunter made research which is the most relevant to the study. Representatives of Russian companies took part in the study. Of these, 41% are from Moscow, 14% are from St. Petersburg, and 45% are from other regions of the Russian Federation. These were mostly representatives of medium-sized companies (20%) with a population of 100 to 250 people: HR managers (30%), HR Directors (19%) and recruitment managers (13%). General managers (9%) and Department managers (2%) also participated in the survey.

30% of respondents believe that by 2050, half of the professions will be fully robotic. Each sphere has its own risk coefficient. The largest (coefficient more than 6) banking, accounting and Finance, insurance, logistics and production activities are affected. A 50/50 chance for retail (5.1), marketing, advertising and PR (4.8), and law (4.5). Creative professions are least afraid of artificial intelligence: art (1.8), design (2.9), media and journalism (3.2). At the same time, 67% of respondents hope that new technologies will allow them to focus on creative and social activities.

In the case of recruitment, 33% of respondents believe that artificial intelligence will completely change the role of HR in companies. Experts also estimated the probability of what roles of a recruiter AI will capture in the next 3-5 years: most likely, robots will be fully engaged in searching for candidates (63%), attracting applicants (47%) and training current employees (47%). Artificial intelligence is already partially performing these functions. Employee motivation will be carried out without human involvement unlikely (21%).

Nevertheless, there is no results about the process of hiring as a process being entirely conducted via artificial intelligence technologies. Moreover, no employees opinion about this aspect. These two things influenced the authors to conduct the deficient part of the research by themselves. The results of the study are useful both for employees (to understand what percentage of automation of their hiring process should they get prepared), and for experts (to understand the usefulness of artificial intelligence as a way to get rid of routine work).

Based on all the above data, the very controversial and doubtful research questions have arisen: "To what extent do the opinions of experts and employees on the issue of fully automating the hiring process coincide?" “From which stages of HR processes the implementation of AI can be started with the smallest resistance?”

This study will examine the opinions of experts regarding the implementation of artificial intelligence in HR business processes. People in search of work will also be interviewed about their attitude to replacing HR managers with artificial intelligence.

The research purpose of the particular study is exploratory with elements of descriptive research. This is based on the fact that the authors have described the stages of HR business processes basing on the in-depth interviews with experts, then explored the correlation between opinions of experts and usual employees, and at last generated new hypotheses regarding the issues of employees behavior and changes in the labor market. Therefore, the research can be defined as exploratory with descriptive elements (with usage of in-depth interviews and cross-sectional surveys).

Here are the main stages of the hiring process, which are usually conducted by all HR managers. There can be some differences in the hiring process depending on the specifics of the vacancy and technological development of the company. For example, for some jobs (and mainly in large corporations) applicants may be asked to pass some tests before being invited to the interview. It is usually done in order to check the level of the person's intelligence and ability to suit the requirements of the job. The second difference is usually in the existence of artificial intelligence in the company. For example, AI can check some activity of the candidate in social networks after interview is being conducted. Here are the main stages of the hiring process.

1. The study of the labor market. The first thing the HR Department needs to do is to study the labor market. The information obtained will help to identify the main problems of employment in the labor market at that time, and provide data on its dynamics.

2. Creating a resume database. No recruitment selection and recruitment process is complete without creating a database of suitable resumes for the vacant position. The following data specified by the applicant is mostly important while creating a database: level of education, previous experience, knowledge of foreign languages (if this is required by the specifics of the work), marital status, ownership of a computer and specific software.

Already experienced HR managers can make certain conclusions about employees' abilities just by looking at how the applicant's questionnaire is compiled. Often, close attention is paid to the literacy of filling out a resume, to its formatting, and to the overall level of literacy. All this allows you to evaluate your knowledge of the Russian language and the degree of proficiency, for example, while applying for a text editor.

3. Selection of prospective applicants

No selection or recruitment process would be as effective if those candidates whose resumes made a good impression on the HR Department were not interviewed in advance over the phone. The goal of telephone conversations is to reduce the total number of people who will be invited to an interview in the future. So what can you find out from the applicant at the same time?

It is usually done in order to get a general idea of a person, especially the way they communicate.

After this stage, the artificial intelligence can send the invitations to the applicants.

4. Conducting an interview

Interviews are usually conducted using a variety of methods. It all depends on the company and the staff service. Often an interview includes the filling in of questionnaires, filling out questionnaires, passing psychological tests.

After analyzing all the responses and behavior of the applicant at the interview, certain conclusions are made about how the person can communicate and whether they will join the company's team.

5. Making a final decision

After the interview has been conducted and the applicant has completed certain tests, the HR service selects the resumes of the most suitable candidates for the vacant position. At this stage, the final decision is made - whether to accept a person for a position or reject their candidacy. It is provided that the previous stages are carried out efficiently and in accordance with all the rules, this decision will be quite easy to make.

Methods and strategy

As the authors are conducting the research to explore the new topic, the opinion of experts is needed. The qualitative part of the research consist of conducting fifteen in-depth interviews with HR managers from different spheres of business. The spheres are chosen according to the dominant skills of employees, which are being searched for. The skills are divided into hard and soft skills, in particular: technical skills (1), creative skills (2) and skills of communication (3). These spheres are needed in order to analyze HR business processes in all its manifestations: the hiring processes and attitude towards the necessity of AI implementation can differ radically due to the difference of skills, which are being searched for. For example, HR managers, searching for employees with joiner skills or wood processing skills will not be strongly interested in interviewing the candidates, comparing to HR managers, searching for project managers, journalists or designers. The authors are dividing these spheres by hard skills and soft skills of potential employees needed at the workplace. The interview protocol and particular questions will extract:

a) which stages of hiring process exist in their company;

b) which stages are already automated;

c) the amount of time needed for particular stages;

d) the amount of efforts needed for each stage of HR business processes;

e) what is the personal opinion of experts regarding the artificial intelligence in HR sphere.

After the interviews, the experts are given a small questionnaire, which provides the authors with some quantitative variables. These variables are used to extract the correlations between the experts' opinion and the employees' opinion, which is extracted in the following way: simultaneously with the interviews, the authors are offering the questionnaire to the employees. Therefore, the quantitative part of the research consists of small questionnaire given to the experts, which includes questions about their willingness to give business processes to artificial intelligence. In other words, the authors are exploring, what is the attitude of employees to being hired entirely by artificial intelligence technologies.

The relevant hypotheses for the research are:

1) (For RQ1) HR managers are more likely to accept the automatization of the HR business processes than usual employees.

2) (For RQ1) There are common elements of mistrust to artificial intelligence between employees.

3) (For RQ2) The stage of inviting people to the interviews is the optimal choice for the start of AI implementation.

These hypotheses arose as possible answers on the research question. Based on the study of HeadHunter, the authors suppose that the experts will be positive about the artificial intelligence technologies. Moreover, if there are the reasons for the distrust to these technologies, the elements for it can be similar. This can help artificial intelligence technologies providers to take some actions to decrease the distrust between their customers and society.

The result of this study is based on the analysis of dependent and independent variables, which are identified through surveys for experts and employees. In this study, the dependent variables are:

1) The degree of agreement of experts with statements about the possibility of implementing AI at certain stages of HR business processes.

2) The degree of agreement of employees with allegations of the possibility of implementing AI at certain stages of HR business processes.

3) The degree of implementation of artificial intelligence in the company where the HR expert works.

Independent variables, in turn, are as follows:

1) Type of skills, which are being searched for technical skills (1), creative skills (2) and skills of communication (3). The skills are divided in this way according to the “McKinsey's” research about 3 most distinct spheres of skills, which are being searched for.

2) Type of respondent: expert or employee.

The study gathers data at one-time point and creates a kind of “snapshot” of the research subject. The authors collect data once and simultaneously from the experts and the employees that is why the cross-sectional timeframe is chosen for the research with no longitudinal elements. The advantage of cross-sectional studies is a single data collection, which makes it easier to attract research participants and speeds up the processing of the collected data. It is also easier to ensure that the sample is representative in cross-sectional studies than in longitudinal studies (British Medical Journal, 1997). Therefore, the study is a cohort one, as the authors are exploring the particular group of subjects, which do not change over the course of study.

The target audience as well as the research itself are applied, as the authors are offering practical solution to a concrete problem and are addressing the specific needs of the practitioners.

The population covers two groups of people: one is the HR managers of the companies, of the business segments described below, while the second group consists of the employees. Practically, the second group consists of bachelor students of 3-4 courses and master students of 1-2 courses. From the population the authors are selecting the respondents, the sample units. The statistical error is supposed to be not more than 5-7%. These sample units comprise the sample, selected via the sample strategy. The chosen sample method is convenient sampling, as the target population is defined in terms of a very broad category and the sample units are easy to approach. The members of population who are available at the moment are asked to participate in the research, that is why this sampling strategy is easy and inexpensive and applicable for the research.

The authors are using the primary data and collect it themselves. As the research question is based on the relation between the opinions of two groups of the respondents, HR managers and the employees, the authors need to collect very fresh opinions of the experts and employees of particular business segments. The primary data is used because there is no such research or even close research to this, as it is proven below, that is why the data should be collected via self-constructed in-depth interviews and cross-sectional surveys.

Limitations

Despite the fact that this study includes reliable methods for collecting and analyzing data. There are a number of factors that may serve as limitations on the relevance of the current study. Firstly, the fact that artificial intelligence has not been fully created and studied can serve as a limiting factor for expert answers: this technology is being introduced gradually, and not all companies managed to be affected, and not in all its manifestations. Experts can answer questions based more on their internal feelings and forecasts for this technology than on real data and facts about the introduction of artificial intelligence in the company due to the lack thereof. Secondly, people's opinions can also be based on passages of knowledge that they already have at the moment about artificial intelligence, but this data is even more limited than that of experts. However, these answers are very important in order to understand the level of people's mistrust in relation to artificial intelligence, and to identify ways to reduce this mistrust in the future. Thirdly, this study analyzes only the market of Russian companies. In Russia, in the labor market, the recruitment processes may somehow differ in their algorithms from the labor market of foreign companies, which may make the result of the study not very applicable for companies abroad. In addition, the mentality of foreign people and their level of trust may differ from the mentality of Russians, which also affects the level of people's trust in artificial intelligence.

3. Theoretical foundation

The sources of human intelligence

Science studies the essential connections and general patterns that inevitably, though in different variations, manifest themselves in every particular case. They can lie on the surface and literally shout about the need for their research, and can be hidden under the thickness of the superstitions and stereotypes, lost in the labyrinths of everyday practices, according to Lobanov A. P..

Dunkan J. states, that there is no wonder that people have so exalted the mind. After all, it created human world with all its outstanding achievements: medicine, art, food industry, cozy and warm homes -- with all the products of the human mind and will, through which people transform their lives. But it also creates many great dangers for people and planet Earth: climate change, weapons of mass destruction, deep imbalances in the distribution of food and other goods, pollution and destruction of the ecosystem, pandemics arising from people behavior - all the result of free choice and human activity, and all this could be avoided if our minds functioned differently.

Moreover, Dunkan J. shows several approaches to the definition of intelligence. The essence of first of them is that any decision and behavior of people can be explained by the action of mind. Man is regarded as a rational agent. The choice is based on reasonable motives, and by explaining these motives, the author is explaining the nature of people actions. When human behavior is explained by rational motives, that treat people as subjects, not as objects. From this point of view, people are free agents, the cause of what happens in the environment, not the consequence of it. From the perspective of second aspect, humans are biological mechanisms with certain biological limitations. Of course, people think and reason, formulate desires and ideas, make plans and intentions. However, these rationalities are not formed in some abstract way -- they are formed by a certain mechanism. The understanding of this mechanism forms the definition of intelligence.

The word intelligence or intellect comes from Latin word “intellectus” (“perception”, “mind”). Definition has several variations, some of them are even controversial. According to Feuerstein R.: “The theory of Structural Cognitive Modifiability describes intelligence as "the unique propensity of human beings to change or modify the structure of their cognitive functioning to adapt to the changing demands of a life situation". American Psychological Association (APA) in 1995 stated in the report, that: “Individuals differ from one another in their ability to understand complex ideas, to adapt effectively to the environment, to learn from experience, to engage in various forms of reasoning, to overcome obstacles by taking thought. Although these individual differences can be substantial, they are never entirely consistent: a given person's intellectual performance will vary on different occasions, in different domains, as judged by different criteria. Concepts of "intelligence" are attempts to clarify and organize this complex set of phenomena. Although considerable clarity has been achieved in some areas, no such conceptualization has yet answered all the important questions, and none commands universal assent. Indeed, when two dozen prominent theorists were recently asked to define intelligence, they gave two dozen, somewhat different, definitions”.

Hegel G. W. F., German philosopher, in his main work “Enzyklopдdie der philosophischen Wissenschaften” (1817) suggested that the tasks of the intelligence are:

a) to produce knowledge about the world;

b) to remember them;

c) to transform them into various plans and projects for the reconstruction of the world.

The intellect and the activity of knowledge are one and the same. All the intellectual faculties of man - contemplation, representation, imagination, memory, thinking - have no independent significance outside the activity of knowledge and act only as its working moments. Accordingly, to understand what intelligence is, it is enough to consider the content side of the human process of cognition of the surrounding world. The process of cognition itself includes three successive stages of activity of the intellect:

a) contemplation,

b) representation,

c) thinking (Hegel G. W. F., 1817).

Nevertheless, the idea about the existence of general intelligence (g) arises not only in the work of the famous psychologists and philosophers, but of Charles Spearman as well, who helped to develop the factor analysis approach in statistics. According to Spearman Ch., general intelligence (g) is connected with many clusters that can be analyzed by factor analysis. After usage of an earlier approach to factor analysis, the author found that scores on all mental tests (regardless of the domain or how it was tested) tend to load on one major factor. Spearman suggested that these disparate scores are fueled by a common metaphorical “pool” of mental energy. He named this pool the general factor, or g (Spearman, 1904).

All the aspects and definitions the authors have presented, leads them to the general trend in intelligence definition: the ability or the mechanism in particular of the individual to deal efficiently with or to adapt to the environment and act rationally in complexity of conditions. This trend was as well marked by Wechsler D. (1944) and by Gottfredson L. (1998).

The sources of artificial intelligence

The history of artificial intelligence began with Alan Turing's question: "can machines think?". Alan Turing was one of the people who worked to create the Enigma cipher machine that helped resolve the outcome of World War II in favor of Allied Forces (History of artificial intelligence is clearly stated in 2.3).

To begin with, to answer this question, it was necessary to give a clear definition of the two fundamental terms of this question: "machines" and "think". Only after these two concepts were defined the work on creating machines that could think like humans has begun. In other words, it was the moment when the era of artificial intelligence began. The main limitation in defining AI as simply "creating intelligent machines" is that it doesn't actually explain what artificial intelligence is? What makes a machine intelligent?

Authors Stuart Russell and Peter Norvig has written a textbook: “Artificial intelligence: a modern approach”, where they consider this issue through combining their work around the topic of intelligent agents in machines. Overall, in accordance with the words of Russell and Norvig, AI is "the study of agents who get recipes from the environment and perform actions". These authors consider four fundamental aspects of the above-stated fact: thinking humanly, thinking rationally, acting humanly and acting rationally. In other words, they have established a clear algorithm of analyzing the surroundings and, basing on the results of analysis, preforming actions. These processes are related to reasoning and behavior.

What is interesting is that in English, the phrase artificial intelligence does not have the anthropomorphic coloring that it acquired in the traditional Russian translation: the word intelligence in the context used rather means "the ability to reason intelligently", and not "intelligence" (for which there is an English equivalent of “intellect”).

Moreover, the problem in defining the term of “intellectual actions”. John McCarthy, American computer scientist, author of the term "artificial intelligence», claims that “we can't yet generally determine which computational procedures we want to call intelligent. We understand some of the mechanisms of intelligence and do not understand the rest. Therefore intelligence within this science is understood only as the computational component of the ability to achieve goals in the world”.

DataRobot CEO Jeremy Achin spoke at a conference in Japan AI Experience in 2017 offered the following definition of how AI is used today: "AI is a computer system able to perform tasks that ordinarily require human intelligence... Many of these artificial intelligence systems are powered by machine learning, some of them are powered by deep learning and some of them are powered by very boring things like rules."

In general, artificial intelligence can be divided into two categories: narrowly focused intelligence and artificial intelligence in the most general sense of the word. A narrow concept involves performing one specific task, the work of artificial intelligence involves analyzing the situation and choosing a solution, however, this process is much more limited than simple human intelligence. Examples of intelligence in the narrow sense of the word are services such as Google and Yandex search engines, as well as built-in image recognition systems; this is the Siri app for IOS; self-driving cars etc.

Artificial intelligence in the broad sense of the word is a much more interesting and exciting phenomenon. These are robots and machines that can apply intelligence to an incredibly wide range of problems, surpassing the ability and speed of thinking of the human brain. The strong AI test is the famous Turing test. If when communicating with a computer through an anonymous communication channel, you cannot understand who is on the other end of the wire, a person or a machine, then we can assume that such a computer-interlocutor really thinks. The essence of this test is that how many do not remember the answers that people give to certain questions and how many do not accumulate phrases that are appropriate at certain moments, there is always a situation where a "mechanical" answer will be impossible (The Thuring test is described more accurately in paragraph 2.3).

However, what is the difference between the artificial intelligence and human intellect? One of the positions about these differences is that artificial intelligence does not have the ability to reflect and motivation. As long as it does not have a certain motivation for action, he will remain an instrument in the hands of a human. Even now, autopilot, implemented, for example, in Tesla, requires a "hands on the wheel" mode. That is, he can drive, but requires constant monitoring by the driver.

The history of artificial intelligence

This chapter is based on the book of English professor, George F. Luger. He has two master degrees in math and doctor degree in Pennsylvania University in 1973. His book “Artificial Intelligence: Structures and Strategies for Complex Problem Solving” was republished 6 times.

Luger G. sets the logical point of the beginning of possibility for the history of artificial intelligence to exist - Aristotle works and his “Logic”. Aristotle combined the intuitive understanding, mysteries, and forebodings of the early Greek tradition with the careful analysis and rigorous thinking that was destined to become the standard for modern science. The closest approach to artificial intelligence is Aristotle's epistemology, or the science of knowledge discussed in his “Logic”: questions of truth of judgments based on their interrelation with other true statements are considered.

The ideas of the Renaissance, based on the Greek tradition, gave impetus to the development of a different, powerful idea of humanity and its role in nature. Mysticism as a means of explaining the universe was replaced by empiricism: the rhythm of life is no more defined by the nature, but by the work of factories. Copernicus and Galileo led science to the new level, but simultaneously expanded the “gap”, which became the origin of the modern definition of “mind”: introspection became an important motive in literature, philosophers began to study epistemology and mathematics, and the systematic application of the scientific method began to rival the senses as tools of knowledge of the world.

In XVII and XVII centuries, the main works about mind and intelligence belongs to Descartes: he attempted to find the basis of reality exclusively by methods of cognitive introspection. Rejecting the information coming from the senses as unreliable, Descartes was forced to question even the existence of the physical world and was left alone with the reality of thought, according to Nadler S. (2013). For the artificial intelligence, the sense of Descartes logic states the following: by separating the mind and the physical world, Descartes and his followers established that the structure of ideas about the world does not necessarily correspond to the subject studied. mental processes exist by themselves, obey their own laws, and can be studied by themselves. Moreover, separation of the human body and the intelligence is mortal for the human, but not for the intelligence. This is the basis of the methodology of artificial intelligence, as states Luger D..

As thinking came to be seen as a form of computation, the next steps in its study were formalization and final mechanization. In the XVIII century Gottfried Wilhelm von Leibniz in the work "Calculus Philosophicus" presented the first system of formal logic, and built a machine to automate its calculations as well (Leibniz, 1887).

“As one of the founders of the science of operations research, as well as the developer of the first programmable mechanical computing devices, the nineteenth-century mathematician Charles Babbage may also be considered one of the first practitioners of artificial intelligence” (Morrison and Morrison, 1961). Babbage's” difference machine “was a specialized device for calculating the values of certain polynomial functions and was the predecessor of his “analytical machine". The analytical engine, designed but not built during Babbage's lifetime, was a universal programmable computing device that anticipated many of the architectural positions of modern computers.

The goal of creating a formal language to describe thinking was also set by George Boole, a mathematician of the XIX century, the founder of Boolean algebra (Boole, 1847, 1854). Although Boole contributed to many areas of mathematics, his most famous discovery was the mathematical formalization of the laws of logic -- an accomplishment that formed the very core of modern computer science. Fully formal (implemented by technical means) approach to mathematical reasoning has provided an essential basis for its automation in real computers. The logical syntax and formal inference rules developed by Russell and Whitehead (1927) constitute the theoretical foundations of artificial intelligence.

Alfred Tarski is another mathematician whose work played a fundamental role in the formation of artificial intelligence. Tarski created the theory of reference, according to which correctly well-formed formulae of Frege or Russell-Whitehead in a certain way refer to objects of the real world. This concept underlies most theories of formal semantics, widely used in artificial intelligence.

Mirkin B. states that talks about real artificial intelligence started at the same time with the creation of computers - in the early 40s of the 20th century. In addition to the purely professional things - modeling of neurons, memory, calculation, everyone was interested in the question: “Can a machine think?”. One of the first works on the question of machine intelligence in relation to modern digital computers, “Computers and intelligence" was written in 1950 by the British mathematician Alan Turing and published in the journal "Mind". The Turing test compares the abilities of a supposedly intelligent machine with those of a human being -- the best and only standard of intelligent behavior. In a test that Turing called a "simulation game", the machine and its human opponent (the investigator) are placed in separate rooms, separated from the room in which the “simulator” is located. The investigator should not see them or speak to them directly -- he communicates with them exclusively using a text device, such as a computer terminal. The investigator must distinguish between a computer and a person solely based on their answers to questions asked through this device. If the investigator cannot tell the difference between a car and a person, then, says Turing, the car can be considered reasonable. This test has some important advantages, which made it a base for many modern artificial intelligence programs:

a) Gives an objective concept of intelligence, i.e. the reaction of a known intelligent being to a certain set of questions;

b) Eliminates bias in favor of living beings.

Mirkin B. shows the classification of history of artificial intelligence as well:

1) “Romantic” period of artificial intelligence: Turing test and machine translation (1940-1960).

2) Deductive stage: output automation (1960 - 1990). Mirkin B. (2010): “In this regard, it is striking that research on artificial intelligence initially mostly went on an intensive path-roughly speaking, reduced to the automation of logical inference. For a long time, perhaps until the early 90's, when leading figures simply walked off the stage, leaving no any interesting results, except that the concept of an expert system and the language of the PROLOGUE, intended for the implementation of formal constructions. In this regard, I recall that in 1974 an international conference on Artificial intelligence was organized in Tbilisi, almost for the first time in the USSR”.

3) Inductive stage - data analysis, data mining, knowledge discovery, clusters (1990 - 2005). The extensional pathway is being intensively developed now, writes Mirkin B.. Discipline “Data mining and knowledge discovery” - activity to identify "interesting" images or patterns in the observed data, helped people to develop logic of artificial intelligence in terms of Theory of Sets.

4) Initial synthesis stage: ontologies (2007-…). Ontology is a system of sets of concepts, with explicitly defined relations between concepts both within sets (usually hierarchies) and between them, as well as statements and facts about these concepts.

The history of artificial intelligence is not finished: moreover, nowadays it goes closer and closer to its pick. Smith C.: “Over the years we have learned that having a massive knowledge base isn't enough, nor is a million logical inferences a second. The hard problem in the field of AI is finding a way to teach a machine to think, but in order to articulate `thought' in a way current computers can understand we must first understand thinking and intelligence ourselves”.

The problems, solved by artificial intelligence and why to implement it

In general, the goal of creating artificial intelligence is to simplify people's lives, as well as a chance for humanity to reach a new technological level and open new horizons. However, artificial intelligence has ceased to be something magical and sky-high, and is now used in a large number of companies in different sectors (such as agriculture, education, medicine, mining, etc.). Basically, the creation of artificial intelligence was an urgent need in those areas of life in which a person has not yet managed, or will not be able to succeed at all. Further examples of the use of artificial intelligence will be analyzed, which will clearly show why it should be used.

The use of artificial intelligence in the medical sector increases the chances of patients for the earliest diagnosis of diseases, for high-quality treatment and recovery with minimal costs. Sarah Knapton argues that machine learning is applied mainly in two cases. First, it is everything related to the recognition and prediction of diseases, recommendations of medicines and tips for treatment. Second, it is the use of strong AI in the field of operations. At this stage, the introduction of artificial intelligence in medicine, creating sufficiently reliable robots that can perform operations that a person is not physically capable of. The movement of the robot perfected to the millimeters, which allows to avoid the possibility of an unexpected mistake.

Artificial intelligence is also used in the financial sector: it allows you to make the most accurate forecasts in the shortest possible time. A human being is capable of such calculations in principle if he has the necessary skills, but it will take a few seconds for a machine and a few hours or weeks for a human being. In this case, the problem of lack of time is solved: the person frees himself from calculations and time for more important things finally appears. According to Ernst and Young report on `The future of underwriting', machine learning will provide an opportunity for continual assessments of data in order to detect and analyze the anomalies and nuances to improve the precision of models and rules.

This section will describe the ability of artificial intelligence to recognize faces and images. In this case, we are talking about the fact that machine learning solves what people are not capable of in principle. Computer vision produces numerical or symbolic information from images and high-dimensional data. According to Masayoshi Yamada, it involves machine learning, data mining, database, knowledge discovery and pattern recognition. Image recognition technology is found in healthcare, automobiles, driverless cars, marketing campaigns, etc. Marketing campaigns of Makeup Genius by L'oreal are based on image recognition and stimulate social sharing and user engagement, which is vital for the developing technology of artificial intelligence.

The following example reflects a phenomenon that haunts us almost everywhere: targeting and product recommendation. No one person would be able to show their product or service to every person who could become a potential buyer. However, artificial intelligence copes with this task perfectly. According to Gregory Linden, the algorithm is quite simple. This is unsupervised learning, during which the machine collects demographic data about users, takes into account search queries and purchase history, on the basis of which it makes a conclusion about whether a person could potentially buy a product or service or not. The use of artificial intelligence allows you not to spend huge amounts of budget on marketing and advertising products, but to do point and giving good results targeting.

Spheres of use of artificial intelligence

According to Solntceva O. G., people tend to call as artificial intelligence any machine that more or less copies human behavior. However, this concept includes much more than just copying: it involves the ability to think logically in order to make rational decisions and evaluate possible developments using algorithms. The process of artificial mind creation is based on the gradual cultivation of the machine. Moreover, the training of the machine, as well as its subsequent use, depends entirely on the person.

The authors are considering several spheres, where artificial intelligence is implemented. This part is based on the work of Solntceva O. G., published in the end of 2018. The author explores the spheres and provides the analysis of the attitude to the artificial intelligence.

a) Agriculture

V. Dharmaraj and C. Vijayanand state: “In agriculture there is a quick adaptation to AI in its various farming techniques <…> to harvest benefits in the field by catching up with the recent advancements in farming sector, the farmers can be offered solutions via platforms like chatterbot.

At present in India, Microsoft Corporation is working in the state of Andhra Pradesh with 175 farmers rendering services and solutions for land preparation, sowing, addition of fertilizers and other nutrient supplements for crop. On an average, a 30% increase in crop yield per ha has already been witnessed in comparison to the previous harvests”.

The capabilities of artificial intelligence also allow people to improve the work of agriculture in these aspects:

1) Disease detection;

2) Identify the readiness of the crop;

3) Field management;

4) Identification of optimal mix for agronomic products;

5) Crop health monitoring;

6) Automation techniques in irrigation and enabling farmers.

And many others.

In the agricultural industry in the future, technologies will be used in vertical farms, that is, in farms where all the necessary conditions for growing crops are artificially created in greenhouses (parameters such as light, temperature, humidity); artificial intelligence will be able to control these processes, maintaining them at the necessary level, states Solntceva O. G..

b) Public service (police and firefighters)

According to “National aeronautics and space administration”, in August 2016, NASA announced that it was working on an intelligent assistant for firefighters "Audrey" (Audrey). This program can monitor a group of firefighters, send useful information to each team member, and give recommendations on how to work together. By observing the firefighters, the assistant can predict the development of the situation in the near future.

In May 2016, researchers at the University of Rochester, in collaboration with the New-York attorney General's office, published a study that says that using AI, you can track down drug traffickers on the social network Instagram. The algorithms embedded in this program analyze the accounts of potential criminals by hashtags, keywords, number of subscribers and "transaction data". This program identifies the perpetrator more accurately than professional experts (Yang, X., & Luo, J., 2016).

There are other examples, but anyway such programs should help people, but not replace them.

c) Household

Since January 2016, M. Zuckerberg has been working on creating a system to help manage the house. The artificial assistant will turn on the light, monitor the temperature in the house, turn on the music, be responsible for the operation of certain devices, including opening and closing the garage, gates, and much more. M. Zuckerberg also plans to teach him to recognize faces, so that the system itself could let friends or family into the house. In May of the same year, Google introduced a similar device Google Home.

Artificial intelligence is gradually infused into our daily lives. "Smart" houses no longer seem to be something fantastic and incredible; soon they will be able to become indispensable assistants to a person at home. In addition, they will be able to reliably provide the inhabitants of the house with the necessary minimum comfort, and will also help to predict emergency situations, because of which you can remain without electricity or heating.

d) Education

Many schools and universities are already using artificial intelligence technology for educational purposes. Most of them use artificial intelligence to keep track of whether students go to classes and perform the tasks given to them. Thanks to the introduction of artificial intelligence, there are intelligent learning systems that can check the student's tasks, his level of knowledge, analyze their responses, and make personal training plans. For example, the AutoTutor system teaches programming language, physics, and critical thinking. Online platforms such as Udacity, EdX, evaluate writing tests and essays. There are also platforms that help with learning foreign languages or improving the native language. By analyzing the natural speech of the student, the system detects errors in pronunciation and offers correction options.

e) Banking

The use of artificial intelligence systems in the banking sector is relevant today, but not many banks can implement projects related to such expensive high-tech technologies, according to Butenko E. D..

Nevertheless, in Russia is considered the trend of growing popularity of using artificial intelligence in banking sector. Many of the electronic payment system use a program that calculates the suspicious activity of a user. The artificial intelligence system is also used to analyze the level of taxes and income in order to demonstrate to the user what his financial condition will be in the near future. An application Wallet.ai based on the data about the person can manage finances, prompting in which cases it is better to save.

f) Transport

Since 2012, Google has been actively testing its self-driving cars on city roads. Google plans to launch them in production by 2020 companies such as General Motors, Tesla, BMW and Ford are also interested in producing unmanned vehicles. In their opinion, such machines are the future of humanity. The autopilot system will take control, ensuring the safety of the driver throughout the journey, and in critical situations-to transfer the control system to a person, states Rafique A..

g) Manufacturing

“Industrial AI is a systematic discipline, which focuses on developing, validating and deploying various machine learning algorithms for industrial applications with sustainable performance. It acts as a systematic methodology and discipline to provide solutions for industrial applications and function as a bridge connecting academic research outcomes in AI to industry practitioners” (Lee, J., Davari, H., Singh, J., & Pandhare, V. (2018)).

h) Medicine

Artificial intelligence it is already actively used for the development of the medical field.

AIME company, for example, is looking for ways to prevent diseases with technology. The approach of this company is based on the analysis of various problems, the study of difficulties and previous actions taken. This has already helped to find solutions for some diseases. For example, Zika virus and Chikungunya fever. Thanks to the use of AI, it is now possible to predict in advance the place where a new disease may break out and predict the course of its further development.