The state of development of artificial intelligence in polish industry: opinions of employees

AI is the science of how to make computers do things that humans have so far been better at (Buckner & Shah, 1991).

The beginnings of AI date back to the 1950s. It was in 1956 at a conference in Dartmouth that John McCarthy coined the term „artificial intelligence.” However, it should be recalled that Charles Babbage is credited with the idea of creating a computational engine, which led to the concept of the computer taking shape at a later stage. Contributions from von Neumann and Alan Turing were also noticeable, resulting in the creation of modern computers. Both Turing and von Neumann realized the importance of a general problem-solving machine that could deal with any type of problem (Kumara & Lehtihet, 1989).

The first publications, even if they refer to AI, rather indicate the process of development of such intelligence, and not the very fact of its existence (cf., e.g., Conway et al,, 1959).

In the 1960s, research was carried out on heuristic procedures for balancing production lines and computer programs for the implementation of these procedures (Tonge, 1960).

In 1963, a work edited by E. Feigenbaum and J. Feldman entitled „Computers and Thought” appeared. A review of this paper published in Management Science said: „This work is notable.... points to the emergence of Artificial Intelligence as a mature field with well-established directions, methods, theoretical concepts” (Rapoport, 1964, p. 204). In 1965, it was thought that there were almost no businessmen left who would not think about the positive and negative sides of the use of computers. But the question arose: could they think, since this process consists in making judgments and creating concepts based on the senses of perception (Kernan, 1965).

The 1970s saw the development of decision support systems (Donovan & Jacoby, 1977; Keen, 1976; Simon, 1978); a period of application of AI in information and data management processes (Smith, 1976; Wong & Mylopoulos, 1977); in production (Pfefferkorn, 1975); and in composing music (Rader & Montgomery, 1974).

The 1980s brought further development of decision support systems (Bouwman, 1983; Keen, 1980; Lodish, 1982), information and data management (e.g., Veaner, 1983; Yannakoudakis & Fawthrop, 1983;), applications of AI in manufacturing (Rosenberg & Lawrence Livermore Laboratory, 1980; Wiig, 1986), development of expert systems (e.g., Biggs et al., 1988; Kastner & Hong, 1984; Kerr & Ebsary, 1988; Kumara & Lehtihet, 1989), as well as training and competence development (e.g., Gladwin, 1984; Harvey & SRI International, 1985).

In the 1990s, computers were already increasingly used to perform a wide range of diverse tasks. It was a period of development of various systems and applications in the sphere of education and science (cf. Husby, 1990; Pelton, 1990). Another area of application of modern technologies was production, with special attention being paid to integrated production systems (e.g., Babbar & Rai, 1990; Baker, 1990).

As in previous decades, many studies and innovative solutions were used in operations research (e.g., Byrd, 1993; Chen, 1994; Kathawala & Allen, 1993). The manufacturing industry saw further development of AI techniques, or rather techniques based on AI, such as the direct work management system (DLMS), HICLASS, and KNACK (Yazici & Benjamin, 1994).

AI began to cover an increasing number of new areas of management, and not only production processes (mainly production planning). In telecommunications, it was present in network management processes (Liebowitz & Prerau, 1995). Construction also became increasingly interested in AI (Rezgui et al., 1998; Yau & Yang, 1998). In marketing, expert systems were used (Winters, 1991). One of the dominant research topics—as it had been in previous years—was the development of decision support systems (Curry & Moutinho, 1994; Syam & Courtney, 1994; Te'eni & Ginzberg, 1991). Yet another area of application of AI was the financial sector (Egan, 1993). In the period analyzed, there was further progress in the use of AI in medicine (e.g., Hawkins & Pollock, 1996).

The first decade of the 21st century was a period in which researchers and commentators were using the term „the age of the economics of knowledge” with a growing frequency (Hauer, 2009).

AI tools continued to be used and developed in operations research—mainly to organize knowledge (e.g., Kobbacy et al., 2007; Murphy, 2005; Tan & Lim, 2006). How the concepts of AI and operations research could be combined to bring about its widespread use on an industrial scale was also studied (Powell, 2010).

Various applications and software in the area of production management were intensively developed, especially in planning (cf., e.g., Allen et al., 2005). The goal—as in previous studies and implementation programs—was to minimize the total cost of production (Lee, 2007; Munguía et al., 2010).

There was also a further development of the use of AI in network management (Qi et al., 2007).

Yet another field in which AI techniques were already used was construction (Arditi & Pulket, 2010; Bożejko & Wodecki, 2010; Kashiwagi & Byfield, 2002). Growing awareness of climate threats brought on by human activity led to AI finding applications is solving environmental problems (e.g., Oprea et al., 2005).

Expert systems were used in human resources management, including in areas such as planning, employee selection, and work analysis (Overholt, 2002). In addition to expert systems, neural networks were used, as they had been in previous years (Sommer et al., 2004). There was also a continued development of the use of AI in financial management processes (e.g., Abdelazim & Wahba, 2006; Dharmaraja, 2002; Sierra, 2007).

Following the example of the previous decade, AI techniques in marketing continued to be developed, for instance, to find and evaluate markets for future products and in the area of marketing research (Crunk & North, 2007; Daskou & Mangina, 2003). The second decade of the 21st century saw AI thrive, practically in all spheres of our lives. Driven by technological advances and public interest, AI came to be considered an unprecedented revolutionary technology that had the potential to transform humanity (Brock & Von Wangenheim, 2019).

The second decade of the 21st century is a period of interest in the AI of things. It combines AI technology with the Internet of Things (IoT) infrastructure to achieve more efficient operations, improve human-machine interaction, and streamline data management processes. Although this concept is still relatively new, many opportunities to improve businesses, as well as industrial and consumer products and services, are expected to emerge as AI develops. This intelligence can contribute to solving existing operational problems, such as the cost of effective human capital management or the complexity of supply chains and supply models (Dahiya & Sayyad, 2021).

Another field of AI development is medicine. AI is already used, for example, in studies to predict the risk of liver metastasis in colorectal cancer (cf., e.g., Han et al., 2022) and in cardiology (Bohm & Jajcay, 2022). The development of modern AI technologies will greatly affect the education and training of people in all professions related to knowledge. Therefore, already at the beginning of the second decade of the 21st century, attention was paid to the possible applications of AI in medical education (Henning et al., 2021).

Marketing has been one of the dominant areas of application of AI. Publications on this sphere of management indicate both the possibilities for applying AI (e.g., Davenport et al., 2020; Kumar et al., 2019), as well as numerous limitations or even errors in its use (Ascarza et al., 2021; Overgoor et al., 2019).

AI is beginning to be used in innovation management (Liu, 2022). It is the result of an innovation process, part of the innovation process, and influences or is influenced by contextual structures (João Correia & Matos, 2021; Navneet et al., 2020).

In the previous decade, a relatively large proportion of research and implementation concerned the use of AI for financial management. This trend has been maintained (Ghandour, 2021; Yubo, 2021). The pandemic has made the financial sector and the business world look even more closely at the opportunities ascribed to AI (Jayaram, 2021).

Research continues to focus on improving decision support systems (Gupta et al., 2021; Patalay & Bandlamudi, 2021).

AI is used in “big data” management processes, among others, in the integration of knowledge generated throughout the product life cycle. The goal is to ensure economical, practical and reliable production and improve product quality (Luo et al., 2021).

AI, including affective computing, is one of the most important and popular technologies used by educational institutions to process and analyze big data (Aljarrah et al., 2021).

One of the biggest beneficiaries of AI is industry, especially in the era of Industry 4.0. The main idea of the fourth industrial revolution (4IR or Industry 4.0) is the digitization and integration of all elements and processes in the company (Blazek, 2021; Yu et al., 2022). New technologies with high-performance computing potential enable the creation of complex AI systems (Oliveira et al., 2021). An intelligent manufacturing system is considered the key to gaining a competitive advantage (Chen et al., 2020). AI is used, for example, in the chemical (Oliveira et al., 2021), pharmaceutical (Kshirsagar et al., 2022), and food processing and handling industries (Kumar et al., 2021).

The use of AI has also had a beneficial effect on improving the state of the environment (Kshirsagar et al., 2022).

The theoretical part of this study involves a systematic review of the literature. The individual stages of the review included: (1) selection of keywords (employees, AI, industry); (2) search for works containing identified keywords; (3) familiarization with selected publications; (4) review of publications; (5) a mind map; (6) a summary of selected publications in terms of work goals; and (7) organizing the collected research material. The procedure used is consistent with the general methodology for conducting research in management sciences (Easterby-Smith et al., 2015).

The study was carried out using a survey questionnaire. The sample was taken with a specific aim in mind: first, 30 entities were selected for that purpose, while in the second stage the employees (managers and specialists) were chosen from among those. In total, 761 questionnaires were completed. Due to the non-random nature of the sample, the findings may not be treated as representative of the entire population of employees of industrial enterprises operating in Poland. They may, however, serve the purpose of a preliminary diagnosis of the studied problem, verification of initial hypotheses, and defining the direction of further research work. Numbers of respondents by organization are in tab.1.

The study mainly involved people aged 31–40 (34.95%) and over 40 years (53.17%). The seniority of the respondents ranged from 5 years (32.7), 6 to 15 years (30.9), to 16 to 25 years (19.58%).

The survey questionnaire used the following rating ranges from “very low” to “very high.” Respondents were given the possibility of expressing their lack of knowledge about AI (see Table 2).

The aim of the article is to show how employees of industrial organizations perceive the development of AI within businesses and to learn their opinions on what AI solutions are most commonly used in Polish industry.

On the basis of the literature review, the following research problems were formulated:

What is the general perception of AI development in businesses?

While the majority of respondents (summed up as “average,” “high,” “very high” opinions) state that the AI systems used in the company are safe and reliable, and that the implementation of AI systems has changed the way technologies are managed, it is puzzling that nearly one-third of employees do not have an opinion on this subject. This is due to the fact that some of the people involved do not deal with AI on a daily basis, i.e., they know about the use of these technologies, but they cannot clearly assess how safe and reliable their use is. There is also a clearly positive opinion about the use of AI. Nearly half of the respondents state that AI systems contribute to increasing the efficiency of processes and the speed of their implementation (27%, high; 20%, very high).

The majority of employees (46.9%, very high; 28.8%, high) are open to cooperation with AI systems that operate or will be implemented in the company.

Opinions on the extent to which companies take actions aimed at substantive support for employees in the field of acquiring knowledge about AI systems are very strongly divided (20.6%, lack of knowledge; 15.8%, very limited choices).

Employees are not afraid of losing their jobs due to the development of AI systems in the company (26.9%, high; 37.4%, very high).

A very interesting research thread is the opinions of employees regarding the degree to which they are willing to accept changes in the scope of their current professional duties. Certainly, these are mere declarations, yet it is worth noting that more than a half of the respondents report such readiness (in total, high and very high choices make up 63.9%) (see all results-tab. 4).

AI Solutions Used in My Enterprise

The research shows that AI is mainly used in production, in the sphere of IT, research and development, and to a lesser extent in marketing, finance, and procurement, as well as administration (see chart below).

Figure 1

Respondents had the opportunity to indicate examples of technological solutions which, in their opinion, meet the characteristics of AI. In their opinion, such examples include implemented robotic process automation (RPA) technologies (automating repetitive business processes using robots that simulate human work), the use of Cognex cameras using neural networks, and the use of predictive maintenance models using machine learning (ML) and data.

The research largely confirmed the previous results of the study. Among the companies using robotic process automation in Poland, robotization is most often used in the area of financial accounting. Robots are most often implemented either in processes usually characterized by a high volume of transactions (sales, purchases) or in processes determined mainly by internal data (closing the month process, reporting). Companies using RPA belong to the group of large enterprises with more than 1000 employees (Remlein et al., 2022).

RPA has already gained considerable popularity in industries such as financial services, telecommunications, public administration, and health care. Recently, the technology has also been flourishing in the areas of production and logistics (Pritchard, 2021). RPA uses AI (so-called digital employees) to free employees from repetitive work, allowing them to focus on activities with greater added value (Crijman, 2021). According to the surveyed employees, RPA is used in the energy and fuel, chemical, and food industries (14 indications).

Cognex cameras (named after Cognex Corporation) allows one to measure and control defects. With their enormous resolution, automatic tracking and lighting adjustment, such systems open up a full range of new applications in the packaging, electronics, automotive, and other industries, where online assembly and inspection operations are crucial for product quality control.

The research points to benefits of employing such systems in the process of machine production (7 indications).

ML is used in a variety of applications and has found numerous applications in various areas of our lives, including marketing, health, social media, science, politics, and even the arts. An array of innovative ideas for the use of ML in business and society is emerging (Padmanabhan et al., 2022). According to the employees surveyed, ML applications are mainly used in production processes; however, they are not widely distributed, being used in the chemical industry with maintenance applications. ML has been gaining in popularity, for example, in the production processes of biofuels (Rodríguez-Rángel et al., 2022; Xing et al., 2021), and in various production control systems (Zhao & Zhang, 2021).

The industry also uses distributed control systems responsible for control and visualization of particularly technological processes (14 indications), enterprise resource planning (ERP) systems (7 indications) and security information and event management (SIEM) (2 indications).

Within IT, regardless of the type of industry, voice assistants are used (5 indications). Various robots are quite popular in the electromechanical, light, and chemical industries (29 indications). However, it is difficult to say to what extent they have built-in elements of AI.

The primary purpose of distributed control systems (DSC) systems is to guarantee efficient and stable processes or the operation of various devices (cf., e.g., Xu et al., 2022). „Enterprise Resource Planning (ERP) systems are information systems in which all business processes in the supply chain, such as purchasing, procurement, production, quality control, shipping, finance, and accounting are managed in an enterprise that produces goods or services” (Aktürk, 2021, p. 70). Therefore, such systems ensure the integration of the processes being carried out.

Security information and event management (SIEM) systems are employed where the process of monitoring and analyzing security-related data plays an important role. In order to improve the effectiveness of SIEM, the use of ML, deep learning, and AI is increasingly postulated (PR Newswire, 2019).

Undoubtedly, the problem of AI and its applications has already raised many questions. Companies are still far from taking full advantage of the enormous opportunities offered by automation (Gotthardt et al., 2020). And when they do, automation of processes—which fosters their optimization—frequently gives rise to concerns among employees as lean corporate programs tend to reduce jobs. Therefore, questions arise as to how much these risks can be mitigated. Is the participation of employees in making decisions on the use of AI really the only way to partially overcome these threats?

While RPA contributes to customer satisfaction, very little is known regarding the impact of this technology on employee satisfaction. We do not know how effective it is in relieving employees from repetitive work, allowing them to focus on activities that generate a higher added value (Crijman, 2021).

Observers say that in the coming decades, robots and other advanced technologies will eliminate millions of jobs in wealthy democracies. However, are their respective populations aware of these forecasts, and if so, does it affect what significance they attribute to economic problems relative to other issues that governments could address? This question is already posed by many eminent scholars (cf., e.g., Heinrich & Witko, 2021).

With regard to the issues discussed herein, several crucial problems can be formulated requiring further literature studies and empirical research. First of all, does the mere fact of interacting with an industrial technical system on a daily basis mean that there is a greater acceptance for AI than in the service industry, which is dominated by software and applications that do not feature AI elements? Second, does the acceptance of AI depend on the forms of employment and employees' position within the structure: after all, questions are already being asked today if AI is able to replace supervisory board members. Third, how does the level of its development affect the level of public acceptance for AI? Where is the boundary that limits our ability to control technical systems?

The results of the research indicate that specific activities, supported by diagnostic tests, are necessary on what the current level of corporate awareness is regarding the need to adapt employee training programs, to what extent the assignment of AI-assisted tasks would give rise to claims of unfair treatment of employees (considering those who will be deprived of AI support). What matters is to determine to what extent the law itself, as well as the internal regulations within businesses, provide for the presence of AI.

The limitations of the research result primarily from the research methods used. Although the author reviewed articles from 1950–2021, the review was limited to the topics of research work. Some articles are opinions, not research results. Databases (so-called for-fee) do not contain all publications and therefore attempts were made to supplement the literature review with Polish publications listed in Google Scholar. Surveys are only employee opinions, often subjective. The scope of research did not cover all key factors, focusing—as planned—on employee opinions regarding the usefulness of AI. In addition, the number of study participants had been limited due to the COVID-19 pandemic.

For the most part, employees are not afraid of losing their jobs due to the development of AI systems in their place of employment. They are positive about the use of solutions that incorporate AI components. According to the opinion shared by a vast majority of respondents, modern technologies, including AI, help in work and facilitate it. More than 75% of employees believe that the development of AI is inevitable due to the number of benefits it entails (such as cost optimization and strengthening of competitive advantage), while almost 80% are fully convinced that the implementation of AI increases the efficiency of processes and the speed of their implementation.

The majority of respondents (the total of “average,” “high,” and “very high” opinions) state that the AI systems used in the enterprise are safe and reliable and that the implementation of AI systems has changed the way technologies are managed. It is clear that, on the one hand, employees are aware of the benefits of using AI, but on the other, some fear it would be accompanied by an increase in the level of surveillance. Just over a half of employees declare they would take on new tasks once AI is implemented (they are ready to accept such developments). At the same time, employees believe the responsibility for preparing and developing new skills should lie with the employer.

Currently, the most common industrial AI applications are: RPA technologies, Cognex cameras using neural networks, ML and data technologies, DSC, ERP systems, and SIEM. A detailed analysis of the results makes us realize that employees struggle to determine whether the solutions they are dealing with are genuine AI solutions. This points to a lack of knowledge, but, above all, a lack of appropriate awareness-raising activities on the part of the companies themselves.