# When technology meets people: the interplay of artificial intelligence and human resource management

Yusra Qamar (Indian Institute of Management Kashipur, Kashipur, India)
Rakesh Kumar Agrawal (Indian Institute of Management Kashipur, Kashipur, India)
Charbel Jose Chiappetta Jabbour (Global Chair in Management, University of Lincoln–Brayford Campus, Lincoln, UK)

ISSN: 1741-0398

Article publication date: 27 July 2021

Issue publication date: 9 November 2021

4590

## Abstract

### Purpose

An original systematic review of the academic literature on applications of artificial intelligence (AI) in the human resource management (HRM) domain is carried out to capture the current state-of-the-art and prepare an original research agenda for future studies.

### Design/methodology/approach

Fifty-nine journal articles are selected based on a holistic search and quality evaluation criteria. By using content analysis and structural concept analysis, this study elucidates the extent and impact of AI application in HRM functions, which is followed by synthesizing a concept map that illustrates how the usage of various AI techniques aids HRM decision-making.

### Findings

A comprehensive review of the AI-HRM domain’s existing literature is presented. A concept map is synthesized to present a taxonomical overview of the AI applications in HRM.

### Research implications/limitations

An original research agenda comprising relevant research questions is put forward to assist further developments in the AI-HRM domain. An indicative preliminary framework to help transition toward ethical AI is also presented.

### Originality/value

This study contributes to the literature through a holistic discussion on the current state of the domain, the extent of AI application in HRM, and its current and perceived future impact on HRM functions. A preliminary ethical framework and an extensive future research agenda are developed to open new research avenues.

## Citation

Qamar, Y., Agrawal, R.K., Samad, T.A. and Chiappetta Jabbour, C.J. (2021), "When technology meets people: the interplay of artificial intelligence and human resource management", Journal of Enterprise Information Management, Vol. 34 No. 5, pp. 1339-1370. https://doi.org/10.1108/JEIM-11-2020-0436

## Publisher

:

Emerald Publishing Limited

## 1. Introduction

The momentous rise of artificial intelligence (AI) is a global revolution that is expected to persist in the future. Seventy-seven percent of the consumers are knowingly or unknowingly using AI technologies ranging from interactive chatbots and smart wearables to personal digital assistants (Seal, 2019). According to recent projections, the total global revenue for AI software is expected to grow from $9.5 billion in 2019 to as much as$118.6 billion in 2025, a remarkable growth expectation of more than 1,100 percent (Seal, 2019). AI, also termed as machine intelligence, was introduced to develop “thinking machines” that mimic human capabilities and intellectual behavior and are capable of supplanting human intelligence (Jia et al., 2018; Min, 2010). AI has been (and continues to be) promoted as a proficient system that accurately deciphers external data and learns from it to attain explicit objectives through workable modifications to its prior learnings (Kaplan and Haenlein, 2019). Since the late 1970s, AI has demonstrated tremendous potential in improving human decision-making processes and in the ensuing efficiency in various business settings (Jia et al., 2018; Min, 2010; Partridge and Hussain, 1992; Sturman et al., 1996). With AI’s increasing acceptance as a decision-aid tool, it is set to become an integral part of nearly all the functional areas of an organization (Fountaine et al., 2019; Jia et al., 2018; Lawler and Elliot, 1996). Many AI tools (such as genetic algorithms, fuzzy sets and artificial neural networks, to name a few) are now being used in various functional areas of organizations (Arrieta et al., 2020; Holland, 1992; Rajpurohit et al., 2020).

One management area that has begun to leverage AI applications and has presented a diverse set of AI usage implications is Human Resource Management (HRM) (Strohmeier and Piazza, 2015). AI has been successfully applied in various HRM functions such as human resource (HR) performance evaluation (Huang et al., 2006; Moon et al., 2010; Zhang et al., 2012), employee selection (Chien and Chen, 2008; Hooper et al., 1998; Oswald et al., 2020; Shahhosseini and Sebt, 2011), employee turnover (Sexton et al., 2005), prediction of the level of employees’ emotional involvement (Lucia-Casademunt et al., 2013), and employee assignment (Hajiha et al., 2006; Karatop et al., 2015). However, prior studies suggest that the research domain of AI in HRM is relatively nascent (Strohmeier and Piazza, 2013) and underdeveloped compared to other fields (Tambe et al., 2019).

Due to the expanding and diverse AI applications in HRM, the area’s academic literature is currently scattered. To the best of the authors’ knowledge, three reviews (namely Abdeldayem and Aldulaimi, 2020; Berhil et al., 2020; Strohmeier and Piazza, 2013) exist in the AI – HRM interface. While Strohmeier and Piazza (2013) have carried out a technique-specific literature review to map the advancements in HR data mining, Berhil et al. (2020) reviewed the proposed information technology (IT) solutions aimed to address the various HR issues. Finally, Abdeldayem and Aldulaimi (2020) carried out a country-specific study to understand AI’s HR prospects. While these reviews enhance the existing body of knowledge, they have primarily focused on a specific technique or country’s context, thus making the scope relatively narrow. Consequently, owing to the expanding breadth of AI–HRM research, there is a growing need for a holistic discussion on current state-of-the-art to encapsulate existing knowledge and understand the extent of AI application in HRM and its current and perceived future impact on HRM activities.

The present study originally addresses the above-mentioned need by conducting a systematic literature review (SLR) on AI applications in HRM. SLRs analyze the literature in a structured way (Fahimnia et al., 2015) and present the results in an objective, reproducible and transparent manner (Hochrein and Glock, 2012). SLRs can provide an informative characterization of a research field’s existing knowledge (Aznoli and Navimipour, 2017) and uncover critical insights for investigating research trends in newer and emerging areas, thereby assisting in mapping future research potentials (Secundo et al., 2020). The present study contributes to the academic literature on AI in HRM by preparing a profile of the existing literature, followed by synthesis and assimilation of the domain’s intellectual capital using structural concept analysis. Based on the SLR’s findings, this study also chalks out potential research topics and the related research questions that demand scholarly attention for advancing the current body of knowledge. The following three research questions (RQs) are addressed to make the above-mentioned contributions:

RQ1.

What is the state-of-the-art research profile for AI applications in the domain of HRM?

RQ2.

What are the primary HRM functions wherein AI has found applications, and how it affects the HRM outcomes (e.g. employees’ emotional involvement level, absenteeism)?

RQ3.

What are the future avenues in the HRM domain that might benefit from the AI applications?

The remaining paper is structured as follows. Section 2 presents the methodological details of the current study. The analysis and findings are provided in Section 3, followed by the discussion and future research agenda in Section 4. Finally, Section 5 presents the implications and limitations of the study.

## 2. Methodology design

A literature review is a tool that helps map and evaluate the extant literature to highlight the boundaries of knowledge and identify potential knowledge gaps (Tranfield et al., 2003). An SLR approach is used to address the RQs of the present study. We decided to adhere to the SLR approach because, unlike traditional reviews, the SLR approach uses article selection methods that are replicable, specific and free from a priori presumptions on the relevance of the selected literature (Pickering and Byrne, 2014). In the view of Mallett et al. (2012), the essence of an SLR lies in its principles of rigor, transparency and replicability. When these principles are applied cautiously, SLRs provide a significant benefit over traditional literature reviews in terms of enhanced quality of findings through improved “transparency, greater breadth of studies included, greater objectivity and reduction of implicit researcher bias, and by encouraging researchers to engage more critically with the quality of evidence” (Mallett et al., 2012, p. 448). The present study consists of three major phases: planning, execution, and knowledge synthesis (a snapshot of the methodology is presented in Figure 1). The planning phase deals with identifying the need for carrying out the study in light of current knowledge gaps. Based on the study’s scope and the RQs, a search query is designed using the relevant keywords to search for relevant literature. In the execution phase, the chosen database is searched using the search query, and the articles are scanned, curated and selected as per the articulated exclusion, inclusion and quality evaluation criteria. In the knowledge synthesis phase, the finalized collection is studied and analyzed to extract relevant information and insights to address the RQs. The subsequent sections present a step-by-step description of the research process adopted in the study.

### 2.1 Designing the search query

For ensuring sufficient coverage of literature on the topic of study, three academicians with adequate subject matter knowledge were consulted to identify the relevant search keywords. Additionally, previous reviews (see Abdeldayem and Aldulaimi, 2020; Berhil et al., 2020; Strohmeier and Piazza, 2013) were also considered to ensure all relevant keywords are included in our search query. The final search query consists of two parts; the first part deals with the keywords related to “artificial intelligence,” and the second part deals with the various keywords used to capture the “human resource management” aspect. We used both “AND” and “OR” operators to design a holistic search query. The final search query was as follows: (“Artificial intelligence” OR “AI” OR “Machine learning” OR “artificial neural networks” OR “ANN” OR “Neural Network” OR “Support Vector Machines” OR “Markov Decision Process” OR “Natural Language Processing” OR “Fuzzy cluster analysis” OR “evolutionary genetic algorithms” OR “genetic optimization” OR “fuzzy inference analysis” OR “genetic algorithm” OR “data mining” OR “workflow automation” OR “fuzzy logic” OR “expert systems” OR “fuzzy logic systems” OR “genetic algorithms” OR “particle swarm optimization” OR “colony optimization” OR “simulated annealing” OR “evolutionary computing” OR “semantic modeling” OR “human performance modeling” OR “robotics” OR “agent-based systems”) AND (“Human Resource Management” OR “Talent Management” OR “Workforce Management” OR “People Management”).

The articles were retrieved from the Scopus database as it is a copious abstract and citation database of peer-reviewed journals encompassing diverse fields (Fahimnia et al., 2015) and is more exhaustive than the Web-of-Science database (Yong-Hak, 2013). We searched the query in the “title, abstract, keywords” of the articles. The initial search resulted in 308 results.

### 2.2 Screening and refinement of the collected literature

As the scientific articles appearing in peer-reviewed journals form “certified knowledge” (Ramos-Rodríguez and Ruíz-Navarro, 2004), therefore, we only consider the articles appearing in peer-reviewed journals for this review. Additionally, the articles yielded from the initial search were also screened to eliminate documents that were not in the English language and were inaccessible or were without full text. These refinements resulted in the reduction of articles to 117.

The next step comprised the manual screening of all the 117 identified articles to remove duplicate articles and the articles that were not relevant to the study’s scope. Considering the study’s purpose and RQs, all the 117 articles were scanned to ensure that only relevant articles are chosen for the study. Sixty-four (64) articles passed the manual screening stage and advanced to the quality evaluation stage.

### 2.3 Article quality evaluation and final selection

To ensure that the current SLR’s outcomes presented unbiased, insightful and transparent results, we evaluated each selected article’s quality in the final sample. Based on the recommendations of Behera et al. (2019), the quality evaluation (QE) followed in the present study consists of 5 rigorous criteria (see Table 1) on which each article was evaluated and scored by two independent researchers. The QE scores for each article are presented in Appendix 1.

Finally, the articles that could not achieve the minimum threshold score were not considered for further analyses. The minimum threshold value (4.5) was ascertained to be 50 percent of the maximum score achieved by an evaluated article (see details in Appendix 1). At this stage, five articles were excluded and the final sample comprised 59 articles. The next step is to extract and assimilate the relevant information to address the research questions formulated for the study.

### 2.4 Content analysis and data synthesis

Content analysis (Krippendorff, 2018) was used to accumulate, analyze, decipher and present various insights on the applications of AI in HRM conferred in the academic literature. Content analysis is primarily employed to comprehend and illustrate the context concealed in an extensive compilation of text data (Hsieh and Shannon, 2005). The unit of analysis for this research is the selected article that has been found eligible for review following a systematic search methodology and quality assessment. We jointly developed a coding template that aimed at extracting insights about (1) the type of research done, (2) the problem scope addressed in the paper, (3) the implementation status of the developed models, (4) the methodology adopted, (5) the key findings and (6) the vital contributions made by the article to the literature of AI – HRM interface (see Appendix 2 for a summary). Based on the reviewed articles’ content analysis, we provide an appraisal of AI’s current usage to aid decision-making in various HRM functions (see Section 3.2). Finally, the knowledge acquired from the reviewed articles’ content analysis is synthesized and presented in the form of a concept map. The presented map provides a taxonomical overview of the AI applications in HRM decision-making based on the type of HR problem (structured or unstructured) it tries to solve using AI sub-fields (see Section 3.3 for details).

While synthesizing the knowledge so acquired and preparing the concept map, the researchers use their respective judgment–driven by logical interpretations of their observations–to arrive at the learnings presented in the study. Nevertheless, adequate caution was exercised to ensure the reliability and validity of the research. Two of the researchers independently conducted the quality evaluation and the content analysis of the articles. To achieve an acceptable interjudge reliability level, we first randomly chose five articles and conducted a pilot test where the two researchers first jointly scored and assessed them. Moreover, the researchers discussed their respective experiences during and after the pilot test, which helped reduce disagreements and achieve a sufficient consensus on the findings. Whenever the researchers faced divergences during the process, the issues were debated to reach an agreement. The analyses were thus triangulated between the researchers to ensure that the findings are reliable and reproducible.

## 3. Analysis and findings

### 3.1 Research profile

The bibliographic information, such as the paper title, journal name, authors’ name(s) and affiliation(s), cited references and keywords for all the 59 articles, were stored and analyzed. These 59 articles were published across 46 journals and accounted for 168 different authors. Figure 2 shows the articles’ publication trend starting from the first article in 1994 until the year 2020 (including papers ahead of print publication, as of the last iteration of data collection activity conducted in July 2020 for acquiring relevant literature).

Table 2 shows the top ten influential articles (based on citation analysis) responsible for carving the research in the domain of AI in HRM. This study, to account for the possibility that a relatively older article may have received more citations just because it was published earlier, employs the total citation counts and citations per year as indicators to ascertain the most influential articles. The article with the highest total citations and average citation per year is Chien and Chen (2008), published in the journal named Expert Systems with Applications, and presents a data mining framework-based case study for personnel selection.

Within the studied literature, 59 articles were published in 46 journals. Around 87 percent of the journals published only one article on AI in HRM. This trend is an indication that the AI – HRM interface is yet to become a mature field of study as mature fields’ research tends to be condensed in a smaller number of dedicated journals, resulting in less scattered knowledge (Von Krogh et al., 2012). Figure 3 shows the ten most productive journals that have published articles on AI in HRM, along with journals’ H index and the number of citations earned across these articles. Expert Systems with Applications leads the list with seven articles, followed by three articles in Computers and Industrial Engineering and the International Journal of Recent Technology and Engineering. It should also be noted that the majority of journals in Figure 3 do not belong to the core domain of HRM, and the field seems to be an interdisciplinary research area that attracts scholars from various research domains.

A journal that has been in publication for 20 years or more and has an H index of more than 20 is considered a successful journal (Hirsch, 2005). Figure 3 suggests that successful journals of differentiable aims and scope do have acceptability for articles focusing on AI applications in HRM. The most influential journal that has published articles on AI in HRM is the European Journal of Operational Research, with an H index of 243.

Figure 4 presents the ten most contributing countries based on the first authors’ affiliation. The first author of the 59 articles belonged to 23 countries; however, six countries, namely USA (n = 10), Iran (n = 6), India (n = 6), South Korea (n = 5), China (n = 4) and Taiwan (n = 4), cumulatively accounted for 59% of the studied articles. Researchers participating from developing and developed economies suggest an international recognition of potential offered by AI applications in the HRM domain. Furthermore, Figure 5 presents the most frequently used authors’ keywords. The different font sizes represent the frequency of occurrence of the corresponding keywords. A bigger font is allotted to the more frequent keywords. “Human resource management,” “artificial intelligence,” “data mining” and “talent management” were observed to be the most used keywords.

### 3.2 The synthesis of the literature on AI applications in HRM functions

The reviewed literature presents a diverse set of recommendations on how specific HR tasks could be handled using various AI techniques. For example, the use of fuzzy logic for selection (Karam et al., 2020), artificial neural networks for performance management (Stavrou et al., 2007), and data mining for job allocation (Chen and Chien, 2011). Undoubtedly, such contributions bring in distinct, meaningful and essential insights into the application of individual AI subfields for different HR tasks. However, these contributions are currently scattered and diverse. To organize how and when these subfields appeared in the HR domain, we first present the timeline of AI application history in the field of HRM in Figure 6. Within the studied literature, we observe that the expert system is the first AI subfield that appeared in HRM literature in 1994, and the introduction of fuzzy logic occurred in 2000. The third subfield, artificial neural network, appeared for the first time in 2001, followed by Data mining in 2006, the Genetic algorithm in 2008 and finally, machine learning in 2011.

Secondly, to consolidate our learnings from the studied literature on the AI techniques employed to handle different HR functions, we created a cross-tabulation with the AI technique used and the HRM function studied across the two dimensions (see Table 3). In Table 3, each article has been appropriately mapped to the intersection of the AI technique used and the HRM function studied in the article. Our classification considers staffing, compensation, performance management and training and development as the primary four HRM functions (Devanna et al., 1982). Some articles that could not be classified into one of the four major categories, due to their usage of a more generalized approach toward HRM functions or nonspecification of function they attempt to address in the article, are classified into a broad category named “other functions.” The most prevalent AI techniques in the reviewed literature are fuzzy logic (n = 13), data mining (n = 12) and expert systems (n = 9). The row of “other techniques” includes the articles that do not deal with one AI technique in particular but discuss the global influence AI has on HRM.

#### 3.2.1 AI in staffing

Tabulating the functions and subfields shows detailed research attention on staffing (n = 23) – a subfield of HRM that is liable for arranging the needed quantity and quality of employees, recruiting (i.e. inviting and selecting) employees, cross-training and assigning employees to suitable jobs, and transferring or firing employees (whenever necessary) (Devanna et al., 1982; Wright and Snell, 1991). The first and the largest subcategory of research contributions concerns employees’ selection (n = 15). This subcategory addressed selection either based on competency (Hajiha et al., 2006; Karatop et al., 2015; Mavi and Mavi, 2014; Shahhosseini and Sebt, 2011; Wi et al., 2009), or a specific employee segment (such as army personnel) (Hooper et al., 1998), or with the focus on reduction of cost and time involved in the selection process (Karam et al., 2020; Maree et al., 2019; Shahhosseini and Sebt, 2011), or based on performance ranking (Canós and Liern, 2004; Fowler et al., 2008), or predicting work behaviors (Chen and Chien, 2011; Chien and Chen, 2008). Another subcategory discussed automation in the staffing of employees (n = 4) (Gupta et al., 2018; Nawaz, 2019; Richter et al., 2008; Vinichenko et al., 2019), followed by problems of the likelihood of absenteeism by the employee (n = 1) (Varalakshmi and Dhivya, 2019), allocation of employees to specific jobs (n = 1) (Lin et al., 2020), employee transfer problem (n = 1) (Acharyya and Datta, 2020) and finally a study on understanding the importance of AI in talent acquisition (n = 1).

#### 3.2.2 AI in compensation

Compensation focuses on the employees’ remuneration discretionally complemented by equity sharing, profit, or both (Devanna et al., 1982). As a subfunction that could facilitate extensive numeric data-driven decision-making to design an efficient AI-based compensation system, it is surprising that only one article addressed this subfunction of HRM (Escolar-Jimenez et al., 2019). As the subfunction deals with an enormous amount of numeric data, AI could potentially design a robust compensation system.

#### 3.2.3 AI in performance management

Managing performance refers to systematically specifying the performance goals and subsequently examining these goals’ realization (Devanna et al., 1982). Being the next largest subcategory after staffing (n = 12), it mainly discussed employee performance prediction (Huang et al., 2006; Jantan et al., 2009; Lamarca and Ambat, 2018; Lopes et al., 2018; Nazri et al., 2019). Apart from the prediction of employee performance, performance evaluation was also the subject of study in some of the articles, which was based either on comparative ranking (Moon et al., 2010; Paladini, 2009) or positive or negative factors influencing the performance (Akhondzadeh-Noughabi et al., 2016; Zhang et al., 2012). A significant number of studies related to employee segments, namely, lawyers (Lopes et al., 2018), bank employees (Zhang et al., 2012), call center employees (Akhondzadeh-Noughabi et al., 2016) and academicians (Lamarca and Ambat, 2018; Nazri et al., 2019). Discussions on performance from an organizational perspective were meant to forecast how a group of HRM practices affects future organizational performance (Manafi and Subramaniam, 2015; Stavrou et al., 2007).

#### 3.2.4 AI in training and development

The training and development subcategory mainly focused on the employees’ competency assessment, as well as basic and advanced training and career planning (Devanna et al., 1982). Of the articles studied in this paper, only five articles discussed this HR function through AI application. Studies included competency assessment based on visualization of the competency gaps (Bohlouli et al., 2017) or past data (Danping and Jin, 2011), either for evaluation or training. Research contributions also refer to the forecasting of future competencies and job profiles needed for organizational transformations (Jerman et al., 2020), career matching based on skill preferences (Lee and Ahn, 2020) and measuring psychological capital for recruitment and appraisal (Alola and Atsa'am, 2020).

#### 3.2.5 AI in other HRM functions

In addition to these function-based researches, a diverse set of themes were addressed in some articles, added as “other” work in this paper (n = 18). The overall influence of AI application was a frequently studied aspect, which includes AI’s ability to improve decision-making (Han, 2016; Masum et al., 2018; Oswald et al., 2020; Ranjan et al., 2008; Reddy et al., 2019), presenting AI as an essential approach (Abdeldayem and Aldulaimi, 2020), and the widespread influence AI has on HRM (Chakraborty et al., 2020). Some peculiar subjects like exploring emotional involvement (Lucia-Casademunt et al., 2013), developing an employee suggestion system (Marksberry et al., 2014), perception of service robots (Xu et al., 2020) were also addressed. Specific discussions on applications were also attempted by some authors in this subcategory (Byun and Suh, 1994, 1996; Lawler and Elliot, 1996; Sturman et al., 1996).

### 3.3 The AI – HRM concept map

Deane et al. (2008) suggested that knowledge mapping is essential as it allows meaningful learnings to occur. Developing a concept map is a way to organize the knowledge emerging from a large amount of qualitative data as it helps in structuring, visualizing and analyzing complex data – which might otherwise be challenging to comprehend – in a systematic way (Markham et al., 1994; Roth and Roychoudhury, 1993; Wallace and Mintzes, 1990). Concept maps can be of various shapes, ranging from nonhierarchical to hierarchical structures and even data-driven maps where the input determines the map’s shape (Davies, 2011). The usage of concept maps can help avoid problems related to learning disorientations and information overload because a concept map helps the learner comprehend and handle the complete picture of a domain’s knowledge in a relatively easy and concise way (Chen et al., 2008). A concept map’s hierarchical levels, the number of relationships, cross-links and branchings are considered to be an estimate of cohesion or integration in the knowledge base (Wallace and Mintzes, 1990).

From the content analysis, we were able to deduce that AI application in HRM is made to facilitate or support HR decision-making. Based on this, we synthesized the reviewed literature in the form of a concept map to comprehend how each subfield is applied to HRM by understanding the kind of HR problem the AI technique tries to solve. Figure 7 presents a reversed tree-like structural concept map for synthesizing and presenting the research focus area’s taxonomical overview. Based on Dulebohn and Johnson's (2013) HR decision-making framework, the map has been classified into two parts – namely, unstructured/semi-structured and structured. The figure reveals the type of AI technique that can be applied depending upon the degree of the HR problem structure. The decision problems that are agreeable to mathematical models such as linear programming or other statistical techniques are termed structured decision problems. There exist standard solutions for such kinds of decision problems, and the methods required for attaining these solutions are known (Niu et al., 2009). Decisions, such as employee performance ranking by prediction (Lopes et al., 2018), would be considered a very structured decision.

On the contrary, the decision problems that are not so clear-cut as they do not have standard solutions and require human judgment to solve are termed as unstructured decision problems. Examples of unstructured decision problems can be: what are the required job training for employees; what are the effective salary parameters; what level of employee benefits to propose (Saidi Mehrabad and Fathian Brojeny, 2007). The decision problems falling amidst structured and unstructured decision problems and requiring a mix of standard solutions and judgment are termed semi-structured decision problems. Researchers posit that more productive outcomes can be achieved when hybrid AI techniques (such as fuzzy artificial neural network (FANN), adaptive-network-based fuzzy inference systems (ANFIS), fuzzy transaction data-mining algorithm (MFTDA)) are utilized for unstructured, semi-structured, or indistinct decision-making problems (Jantan et al., 2009; Masum et al., 2018).

In Figure 7, starting from the top (i.e. HRM decision-making), the map is branched into six AI techniques used within the reviewed literature. Below is a brief discussion on all the six techniques applied in the HRM literature.

#### 3.3.1 Expert systems (ES) in HRM

The program devised to configure experts’ knowledge into logical structures, decipher their heuristics into orderly rules, and utilize these rules to grant eminent expert resolutions to users is termed as an expert system (Al-Attar, 1990; Sturman et al., 1996). The advancement of ES in HRM can be seen in studied literature as early as 1994, where Byun and Suh (1994) discussed that ES could be used for knowledge representation in the form of semantic networks or semantic nets in major HRM activities such as HR planning, compensation, recruitment and labor-management relations. Additionally, ES can improve decision-making by providing the decision-makers, in effect, online access to proficiency that might be arduous to generate and is not readily available (Lawler and Elliot, 1996; Sturman et al., 1996). ES development in the HR domain helps solve unstructured HRM problems and contributes to developing complete human resource information systems (HRIS) (Byun and Suh, 1994). Evidence of the development of a mathematical model for competence assessment (Bohlouli et al., 2017), basic rule-based ES for selection called BOARDEX (Hooper et al., 1998), using ES as a decision aid (Sturman et al., 1996), and using ES for efficacious assignment and selection of the job seekers (Saidi Mehrabad and Fathian Brojeny, 2007) can be found in the literature.

#### 3.3.2 Fuzzy logic in HRM

Fuzzy logic is based on the fuzzy set theory, which was proposed by Zadeh (1965). In this theory, there are membership levels defined in a set whose value varies within 0 and 1. 0 indicates no belonging, whereas 1 shows absolute belonging to the particular set. For any other element having a value between 0 and 1, this value shows the level of its belongingness to the set (Karatop et al., 2015). With these sets, fuzzy logic can quantify the data’s uncertainty and forecast future scenarios, which further facilitates decision-making (Lamarca and Ambat, 2018).

Fuzzy rule-based systems deal with semi-structured problems and are equipped with deciphering human judgment better than techniques that utilize adequate input data but disregard the critical interactions and logical possibilities within the data (Selden et al., 2000). In AI – HRM applications, it can be used for personnel selection and optimal staff design (Canós and Liern, 2004), differentiating between a suitable and nonsuitable job applicant (Karam et al., 2020; Shahhosseini and Sebt, 2011), eliminating inaccuracy in evaluating the proportionate significance of traits and the performance gradings of the choices (Mavi and Mavi, 2014). Furthermore, expert judgment can act as an input for the fuzzy logic implementation, which can further train artificial neural networks. When used in combination with ES, fuzzy logic can augment its reasoning capability, thus improving decisions’ quality (Masum et al., 2018).

#### 3.3.3 Data mining in HRM

According to Han et al. (2011), data mining is a process to draw out valuable but concealed information from large data sources. It additionally alludes to the significant process of recognizing potentially valuable, novel, and valid patterns in the data (Strohmeier and Piazza, 2013). By implementing data mining methods, organizations can transform useful information and patterns to achieve a competitive advantage (Ranjan et al., 2008). Moreover, data mining can be utilized for knowledge discovery as the extracted knowledge can be presented in patterns. The identified patterns can be used to represent the relationship in the form of a decision tree, which further supports decision making. Decision Trees (a collection of these is referred to as a random forest) are usually used for classification and prediction tasks. They are more suitable for predicting categorical outcomes (Chien and Chen, 2008) as they offer the advantage of simple understanding and interpretation for the decision-makers to contrast with their knowledge for justification and validation of their decisions (Chen and Chien, 2011). Data mining is one of the best ways to examine documents in databases as the outcomes can be utilized for prediction and future planning (Chien and Chen, 2008; Jantan et al., 2009).

The data mining techniques applied to HRM are association rules, rough set theory and cluster analysis. Association rules are utilized to explain the models that actively correlate data attributes, and the pattern is primarily found in the form of connotation rules (Danping and Jin, 2011). Rough Set Theory is a data mining procedure in which, within the presence of vagueness and uncertainty, the clarification and investigation of how a decision is being made can be done with straightforward, reasonable and valuable rules (Jantan et al., 2009). Clustering analysis has also been broadly utilized in image processing, pattern recognition, etc. (Han et al., 2011; Ranjan et al., 2008).

Data mining has been used for employee selection (Chien and Chen, 2008; Ranjan et al., 2008), performance evaluation (Akhondzadeh-Noughabi et al., 2016; Lamarca and Ambat, 2018; Nazri et al., 2019), competency evaluation (Danping and Jin, 2011), talent management (Jantan et al., 2009) and various other HR functions.

#### 3.3.4 Artificial neural networks (ANN) in HRM

ANNs are affiliated to the learning-by-example paradigm family in which actual examples are used to automatically generate problem-solving knowledge (Huang et al., 2006). ANN is a simplified model that has been developed to imitate the function of a brain (Hajiha et al., 2006). It is designed using a simple structure comprising of a processing element, layer, and network to reenact the human learning process (Huang et al., 2001). ANN is the most popular intelligent technique for prediction (Masum et al., 2018), which can help in solving models developed for predicting the HR functions like selection (Hajiha et al., 2006), recruitment (Huang et al., 2001) and performance (Lopes et al., 2018; Stavrou et al., 2007; Zhang et al., 2012).

An expert system can be replaced by using a neural network to solve HR problems (Maree et al., 2019). The integration of the neural network with AHP and fuzzy inputs is equipped for forecasting the effects of HRM practices on organizational performance (Manafi and Subramaniam, 2015). A neural network can be trained to perform a specific function by modifying the values of the connections (weights) between elements (Hajiha et al., 2006). If some effective (and ineffective) recruiting data is provided to the neural network, it can set up an entire talent acquisition system on its own (Huang et al., 2001).

A type of ANN called Self-Organizing Map (SOM) has also been applied in HRM. The idea of SOM is one of the most elegant instances of unaided learning, where ANN endeavors to extricate stable highlights or models (Stavrou et al., 2007). This unique characteristic makes it conceivable to train a network on a delegate set of input/target pairs and get excellent outcomes without training the network on all possible pairs. Backpropagation neural networks (BPNN) can also be used to train ANN (Zhang et al., 2012).

#### 3.3.5 Genetic algorithm (GA) in HRM

GAs are search techniques that include discerning experimentation, seeking to determine a global optimal (Masum et al., 2018). They use strategies found in nature, for example, replication, mutation and gene crossover, to arrive at optimal solutions for mathematical problems (Fowler et al., 2008; Masum et al., 2018). GAs have been relatively less used in HRM functions, with a few studies like Fowler et al. (2008) that implemented GA to solve the workforce planning problem and evaluate the performance. Acharyya and Datta (2020) used GA for problems related to the transfer of staff using real-life constraints, and the formation of a framework was done to analyze the knowledge of the candidates by Wi et al. (2009).

#### 3.3.6 Machine learning in HRM

Machine learning is the learning process in which machines can learn independently without being programmed to do the required work in a certain way (Rab-Kettler and Lehnervp, 2019; Varalakshmi and Dhivya, 2019). Studies have revealed that the adoption of machine learning in decision-making is beneficial (Diao and Shwartz, 2017; Varalakshmi and Dhivya, 2019). Data mining is likewise a specific kind of machine learning (learning from examples) (Huang et al., 2006). Machine learning techniques such as logistic regression and support vector machines are currently being utilized for the modeling of HR functions. A technique such as logistic regression is employed when the predicted dependent variables have a binary outcome, for example, accepted/rejected, passed/failed, high/low and so on (Hosmer et al., 2013). It has been used by Alola and Atsa'am (2020) for assessing the psychological capital of employees. Very few studies have used this technique for HR decision-making.

## 4. Discussion and future research directions

### 4.1 Ethical concerns around AI in HRM

Ethics, in general, are guidelines that help people tell apart right from wrong, which is beyond the law (Chang, 2021). Robust ethical frameworks and theories guide people and organizations to make reasoned, logical and justified decisions (Chang, 2021). The use of AI to analyze and visualize complex data from the entire workforce or individual teams, employees and divisions for providing actionable insights can result in ethical concerns and posit risks for employees’ autonomy and privacy (Tursunbayeva et al., 2018, 2021). Employing AI in activities like analyzing complex performance data, developing personalized training recommendations, predicting future performance, inferring employee satisfaction can get prone to unethical practices like biases and unfairness. For example, the expert system in selecting job seekers can set in biasedness indicated by the experts’ knowledge, which can further result in providing preferences to a particular gender, some specific skills, backgrounds, ethnic groups, etc. Since expert judgment also acts as an input for fuzzy logic implementation, it will be opening avenues for bias in HR activities like screening potential employees, grading performances, etc. Moreover, ANN can be trained using fuzzy data, which can continue the chain of human biases in HR practices.

Further, AI subfields, like data mining which uses data to look for valid patterns, may also hold potential ethical threats as the data is concerned with the individuals (Chang, 2021) and can result in opportunities for a data-miner to take advantage of data subject’s vulnerabilities (Dean et al., 2016). The manipulation of such sensitive people data can also be used for training algorithms to modify or “shape” employees’ behavior in and beyond the workplace. Further, if AI’s involvement is extended to monitoring employees’ social media activity, personal emails, usage of digital devices and apps, it can be a breach of privacy (Tursunbayeva et al., 2021).

Another AI subfield, GA, uses replication and mutation strategies, thus magnifying any wrongly fed characteristic to the system. Also, ML has complex inner processing, thus limiting people’s technological understanding, leading to critical information asymmetries among AI users and experts, which further cripples human trust in AI (Ryan and Stahl, 2021).

### 4.2 Toward a framework for ethical AI in HRM

Most organizations have values, policies and codes of ethics that aim to create an ethical culture (Ferrell and Ferrell, 2021). To work in line with these codes, the organization must address the ethical challenges for HR practices that AI implementation would be raising. Several mechanisms can be devised to handle and minimize ethical concerns at various levels. For handling biases in AI decisions, close attention can be paid to any training data fed into the system. Managers are (or can easily be) aware of potential consequences that their decisions may have, so they need to evaluate decisions from an ethical perspective. If the AI output indicates unfair bias, the manager should have the authority to overcome that decision. The expert system, where the expert’s knowledge is taken as a base for decisions, has to continuously update their commands based on the ethical guidelines.

While handling people’s data, special care must be taken as to which data is to be recorded, who should be in charge of the recording action and who can access the records and data. The concerned employees should have the right to choose what they wish to share with the organization and ask for the usage policy concerning the collected data. The clarity in communication has to be there to ensure that the data has not been deceived, manipulated, or coerced by AI. Furthermore, while handling sensitive data, AI’s architecture should be built and tested for security before implementation.

Moreover, finding patterns and making connections based on data should also be checked for any undesirable results. For example, if the input data for hiring is skewed with a disproportionate number of people from a specific race/gender and the AI will make predictions based on the fed data. The result may indicate that particular gender and/or race has the highest productivity and should be given hiring preference. Whereas the whole variable relationship is based on limited data sets and thus gives biased results. So organizations should be aware of the variable relations made based on data and what should be considered in decision making.

Understanding the processes within AI subsystems like machine learning is complex (Siau and Wang, 2020); thus, ensuring transparency in such systems can be troublesome. There will be a need to provide oversight to monitor and access outcomes associated with them. Here, the differences in the quality and quantity of training data can also mean that the outcomes can vary widely, emphasizing why we consider that the quality of being self-learning is one central distinguishing characteristic of AI in organizations (Brendel et al., 2021). The replication based on training data will need its own codes of conduct to address new engagements and ethical issues not previously programmed into the existing portfolio of alternatives and solutions. The programming is possible because machines will have a narrow focus or activity in most cases, and it will be possible to anticipate risks, alternatives and consequences that can be used in decisions. Thus, while implementing AI, there always has to be human controls that will have to adjust ethical outcomes.

Figure 8 shows an indicative summary of the ethical concerns observed in the studied literature that require appropriate mechanisms to cope up and transition toward effective as well as ethical AI applications in HRM.

### 4.3 Closing remarks on the state of AI – HRM domain

The emergence of AI has fundamentally transformed many organizations (Bankins and Formosa, 2020; Barro and Davenport, 2019; Kaplan and Haenlein, 2019; Jia et al., 2018). The crucial role of AI techniques in the workplace is to support complex HR managerial decision-making by enhancing the quality and pace of the decision-making process (Escolar-Jimenez et al., 2019; Reddy et al., 2019; Ranjan et al., 2008; Saidi Mehrabad and Fathian Brojeny, 2007; Sturman et al., 1996). The sophistication of AI-controlled frameworks has lately expanded to such a degree that no human intercession is required for their structure and deployment (Arrieta et al., 2020). The impact of the sophistication can be seen from the employment of AI techniques in the hiring and recruitment process (such as shortlisting of CVs from career sites, direct and break down video interviews) (Gupta et al., 2018; Nawaz, 2019; Rab-Kettler and Lehnervp, 2019; Wan Chik and Arokiasamy, 2019), predicting performance (Chen and Chien, 2011; Lopes et al., 2018; Nazri et al., 2019), and automation of tasks (Escolar-Jimenez et al., 2019; Gupta et al., 2018). Researchers posit that hybrid AI techniques (such as fuzzy artificial neural network (FANN), adaptive-network-based fuzzy inference systems (ANFIS), fuzzy transaction data-mining algorithm (MFTDA)), when employed to solve various HR problems, can produce more effective results (Jantan et al., 2009; Masum et al., 2018).

It is observed that literature has mainly focused on applying the AI technique to selecting employees, with 25% of the studies addressing this issue. While selective attention was also paid to employee performance prediction, evaluating competencies, employee assignment and the overall impact AI has on HRM, the remaining research contributions form a somewhat spotty compilation of incongruous HR issues. This review communicates that there is considerable knowledge available on AI applications in HR functions, but the application scope is skewed, focusing on few HR issue areas. For example, despite being essential issues for organizations, issues like employee turnover (Khatri et al., 2001; Pereira et al., 2013) and training investment (Malik, 2009) are barely discussed in the studied literature. Strohmeier and Piazza (2015) highlighted that the available systematic knowledge of AI application in HRM is indeed arbitrary and incomplete with skewed representation. The work done in the area is usually generic, and the models are developed to optimize or predict the outcomes of a complete HR function. Furthermore, there was only one study that addressed the perception of humans on service robots. Hence, there is a need for more focused research on AI applications in HRM that can easily be implemented in real-life application areas, and more studies are required to capture the responses from the individuals affected by the AI application in HRM work areas.

### 4.4 Recommendations for future research

We now present research directions that would benefit from increased scholarly attention. The past literature has endorsed various avenues for future researchers, including extending the models developed in the study (e.g. Hooper et al., 1998; Selden et al., 2000; Stavrou et al., 2007), replicating the developed models for other HR problems (e.g. Chien and Chen, 2008; Hajiha et al., 2006) and developing different approaches to solving these models (e.g. Fowler et al., 2008; Masum et al., 2018) to name a few.

We formulated and grouped the proposed research directions into four categories (as presented in Table 4); the first category addresses employee perceptions, fears, and potential misuses; the second is directed explicitly at workforce management through AI applications. The third is focused on evaluating the effectiveness of HRM-focused AI applications, and the last relates to the perspective of leaders for implementing AI solutions. A brief discussion on the possible future research directions is as follows (see Table 4 for specific research questions).

#### 4.4.1 Research direction #1: employee perceptions (and fears) and potential misuses

The literature suggests that HRM scholars should pay increased heed to study the impact of usage of AI techniques rather than just focusing on its application in problem-solving. Sturman et al. (1996) assert that the effects of such aids on the decision-making process (and the decision-makers) have been virtually ignored. Some pressing needs that need to be addressed for the development of the domain include the need to understand the impact of users’ personal characteristics on psychological and performance outcomes (Lawler and Elliot, 1996), the need to integrate the DSS, ES and behavioral research literature (Sturman et al., 1996), and the need to understand employees’ willingness to work with robots (Xu et al., 2020). The employee’s fears of losing their jobs or not being able to acquire the required skills for using the technique(s) can also be an added barrier to AI implementation. Furthermore, scholars have also been worried about the potential misuse of AI applications (Strohmeier and Piazza, 2013; Sturman et al., 1996) and some questions (like, can the employee sabotage AI machines not to schedule maintenance in their shift or to distort others’ final performance?) remain to be answered.

#### 4.4.2 Research direction #2: workforce management

Although the majority of the literature is focused on developing models for some significant workforce management activities, few activities are seldom discussed. Critical issues like employee turnover (Khatri et al., 2001; Pereira et al., 2013) were barely addressed in the literature except for Chien and Chen (2008). Training investment (Malik, 2009), although being such an essential part of the HR function, was scantily considered. Although performance prediction was often discussed, variation in the input variables can be of excellent scope for future studies.

Currently, very few studies have implemented behavioral theory(ies) for creating HR function models. Future researchers should explore behavioral aspects related to AI applications from the lens of available behavioral theories such as behavioral intentions, reasoning, dissonance and others. Future research could focus on developing performance models based on theoretical underpinnings from psychological phenomena like the Pygmalion effect. Additionally, the models developed in the current literature are usually modeled for a generic HR function, and the future research can focus on developing specialized models according to the type of organization, level of hiring, size of the organization, type of jobs (e.g. gig workers), geographic areas, etc.

#### 4.4.3 Research direction #3: evaluating the effectiveness and return on investment (ROI)

Another significant area for future research can be the evaluation of the effectiveness of HRM-focused AI techniques as they would only be helpful if they are comparatively more effective and efficient, that is, they provide improved results and take relatively lesser implementation effort than the already established HR technique(s) (Strohmeier and Piazza, 2013, 2015). Apart from AI, with the emergence of data-based decisions in HRM, other systems like big data system can also be designed to provide functionality and services for archiving, data storage, data management, automated backup and data recovery (Chang, 2015), and methods like organizational sustainability modeling (Chang et al., 2016) can be used for understanding HR-based data analysis. Calculation of the Return on Investment (RoI) can also be a valuable tool to recognize the actual worth of implementation of these sophisticated techniques.

#### 4.4.4 Research direction #4: perspective of leaders

Leaders’ perspective toward implementing effective AI solutions and how they can be implemented can play an essential role in sustaining the introduced technique (Reddy et al., 2019). Jerman et al. (2020) and Xu et al. (2020) discussed leaders’ role in accepting robotics in organizations. In-depth insights about leaders’ perceptions and their support, or otherwise, for AI applications and their potential impact could help develop the domain.

## 5. Conclusions, implications and limitations of the study

This study carried out a systematic literature review of 59 articles in AI – HRM literature retrieved from the Scopus database (as of July 2020). The steady growth in the number of publications in recent years indicates that the academic community has increasingly expressed interest in exploring the prospects of AI-aided decision-making in the HRM domain. In the subsequent sections, we present the research and practical implications followed by the limitations that must be considered while interpreting the study’s findings.

### 5.1 Research implications

Research on AI applications in HRM functions is gaining popularity, particularly within the selection and recruitment domains. The present study offers three significant research implications. In response to the study’s first RQ, this study contributes to theory by presenting and discussing the overall research profile, allowing researchers to understand the evolution and the current state of the academic literature in the AI – HRM domain. For addressing the second RQ, the study discusses the potential AI techniques that can be employed in HRM research. The aim here is to highlight a broader range of well-known techniques that might become more influential in future research and substantiate the characteristics of the focused HR problem(s). The second RQ also addresses the practices in HRM that have already been improved using AI. The goal is to illustrate domain success, particularly compared to conventional HR methods, and to analyze the academic literature that can pave the way for future studies.

The study also formulated a unique AI-HRM concept map that may help the researchers to understand the HRM decision-making processes using AI techniques. Moreover, this study also summarizes the ethical issues concerning AI applications in HRM and provides an indicative preliminary framework for integrating ethical practices and strategies to help transition toward ethical AI. Thus, this could help researchers avoid bluntly using AI in sensitive HRM functions that deal with people and people data, resulting in potentially unintended consequences. Finally, we present a brief and nonexhaustive description of the research opportunities along with some specific research questions. Not only can this help set up the course for future research in the AI – HRM interface, but it can also act as structured guidance that may help reduce repetition and bias in conducting AI-HRM research.

### 5.2 Practical implications

The influence of information technology on the study and practice of HRM has significantly transformed routine and nonroutine HR activities, starting from an administrative employee records management function to the strategic management of people (Malik et al., 2020). With AI assistance, HR managers can now employ a range of emerging technologies that enable machines to perform tasks like humans by integrating several databases of knowledge, assisting managers in performing productive data analyses, and organizing their activities toward desired outcomes. By adopting AI in HRM, the job requirements and the HR manager’s overall labor market have also been vastly affected in terms of their skills and capabilities.

As AI agents can better deal with low-end predictable and routine queries from the employees or requests from potential job applicants, several HR processes are additionally prone to change. AI agents can better monitor conversations in real-time, decipher a representative’s manner of speaking and investigate inexplicable circumstances that may require prompt intercession. Such cases may require diverse HR manager intervention levels regarding the data provided, the degree of association proposed and planned alternatives. Therefore, managers should know the organization’s grievance redressal process and are advised to act accordingly. Robots or AI bots can also monitor and transcribe interview and audio data in real-time for organizations like research and development, media, hospitals, etc. Therefore, HR managers need to update their skills and competencies to bring effective technological change and buy into the new changes induced by AI-enabled HR services.

HR systems are additionally prone to require regular updates as indicated by the functionalities that compare with the HR task’s requirements, as not every AI method is appropriate in HR management, and not every HR task can be tackled by an AI method. Finally, the fit of method and task must be arduously explained on an individual premise; some concretizations can be made on the categorical level (Strohmeier and Piazza, 2015). For instance, the expert system can help HRM managers as decision support systems (DSSs), especially in managing work assignments and employee selection issues. Other planning-oriented AI methods, such as hierarchical planning and distributed problem-solving, can be applied in work assignments and workforce planning activities. Therefore, HR managers need to know both the HR systems and the AI systems to achieve the technique’s fit. This study provides an AI-HRM concept map that can help managers decide the type of AI technique that can be applied to the HR problem at hand.

Ethical and privacy issues can also be a significant challenge for HR managers. This study will help HR managers to understand a holistic view of privacy risks and ethical issues in AI implementation. A broader view of the ethical issues and their management will also help managers consider organizational and technological arrangements to manage these issues better. Organizations are the ultimate bearers of privacy, and ethical concerns and literature suggest that mainly people’s information can be misused on the organizational side. Therefore, organizations should go beyond the boundaries to create awareness and educate employees to minimize the risk of such issues.

### 5.3 Limitations of the study

The results of the present study should be interpreted with consideration of the following limitations of the study. First, although systematic and scientific, the methodology for data collection may have led us to miss some essential articles due to their absence from the Scopus database or their usage of keywords other than those chosen for this study. This study only considered peer-reviewed articles available in the selected database and has excluded other publication types such as book chapters and conference proceedings. Second, although a scientific approach was followed to design and execute the current research and adequate precautions were exercised to ensure acceptable validity and reliability of the work, the possibility of personal biases and mis(over)-interpretations cannot be entirely dismissed. Moreover, the current study is more focused on providing a broader overview of the domain to open avenues for future research, and future research should consider employing advanced techniques such as meta-analysis to provide specific insights into the effect sizes of AI influence on HRM.

## Figures

### Figure 1

Snapshot of the research methodology

### Figure 2

Year-wise publication trends

### Figure 3

Ten most productive journals

### Figure 4

Ten most contributing countries

### Figure 5

Most frequently used Authors’ keywords

### Figure 6

Timeline of AI application in HRM

### Figure 7

AI – HRM concept map

### Figure 8

An indicative framework to handle ethical concerns in AI-HRM

## Table 1

Quality evaluation criteria

QE#Evaluation question(s)Scores
+0+1+1.5+2
QE1Is there an explicit discussion around data analysis? What is the type of data analysis employed?No evidence––QualitativeQuantitative
QE2Are the challenges and advantages of the topic of interest discussed?No––PartiallyYes
QE3Are the reported findings valid and coordinated with the applied methods and the topic of interest?No––PartiallyYes
QE4Does the article possess peer recognition and source reliability?TC+H=0
1  TC + H  49
50 TC + H100
TC + H > 100
QE5Are the utilized methods comparable with methods popularly used in prior studies?NoYes––––

Note(s): TC stands for total citations; H stands for the H index of the journal in which the particular article appeared

The above-mentioned evaluation protocol is sourced from Behera et al. (2019) with the following adjustments:

QE1: The presence of an elaborated method (models, rules, frameworks), AND results from computational experiments or surveys is considered to be equivalent to quantitative data analysis; The evidence of (qualitative research) OR (minimal details on methods with minimal results of evaluation) is considered to be qualitative analysis

QE2: The discussion of advantages and challenges has been adapted to the present setting as Yes (detailed discussion on the applicability of results to HRM context), Partially (limited discussion on the applicability of results to HRM context), No (no discussion on the applicability of results to HRM context)

QE3: Partial alignment and validation refer to an unavailable or limited explanation for a utilized technique and the reported outcomes

QE5: The comparability of the utilized methods is adjudged based on employed methods of prior studies in the AI – HRM interface

## Table 2

Ten most influential articles

PaperTotal citations (TC)TC per year
Chien and Chen (2008)20115.46
Wi et al. (2009)15312.75
Fowler et al. (2008)695.30
Strohmeier and Piazza (2013)486
Huang et al. (2006)483.2
Shahhosseini and Sebt (2011)464.6
Ranjan et al. (2008)443.38
Saidi Mehrabad and Fathian Brojeny (2007)412.92
Stavrou et al. (2007)412.92
Hooper et al. (1998)361.56

Note(s): Total Citations as per Scopus database

## Table 3

HRM functions and AI subfields

AI subfieldsHRM functions
StaffingCompensationPerformance managementTraining and developmentOthersN
Expert systemHooper et al. (1998), Maree et al. (2019), Saidi Mehrabad and Fathian Brojeny (2007) Bohlouli et al. (2017)Byun and Suh (1994, 1996), Lawler and Elliot (1996), Sturman et al. (1996), Xu et al. (2020)9
Fuzzy logicCanós and Liern (2004), Karam et al. (2020), Karatop et al. (2015), Mavi and Mavi (2014), Moradinasab et al. (2016), Shahhosseini and Sebt (2011)Escolar-Jimenez et al. (2019)Huang et al. (2006), Manafi and Subramaniam (2015), Moon et al. (2010), Paladini (2009), Selden et al. (2000) Kimseng et al. (2020)13
Artificial neural networkHajiha et al. (2006), Huang et al. (2001), Lin et al. (2020) Lopes et al. (2018), Stavrou et al. (2007), Zhang et al. (2012) Lucia-Casademunt et al. (2013)7
Data miningChen and Chien (2011), Chien and Chen (2008) Akhondzadeh-Noughabi et al. (2016), Jantan et al. (2009), Lamarca and Ambat (2018), Nazri et al. (2019)Danping and Jin (2011)Han (2016), Marksberry et al. (2014), Masum et al. (2018), Ranjan et al. (2008), Strohmeier and Piazza (2013)12
Genetic algorithmAcharyya and Datta (2020), Fowler et al. (2008), Wi et al. (2009) Lee and Ahn (2020) 4
Machine learningVaralakshmi and Dhivya (2019) Alola and Atsa'am (2020)Diao and Shwartz (2017)3
othersBaakeel (2020), Gupta et al. (2018), Nawaz (2019), Richter et al. (2008), Vinichenko et al. (2019) Jerman et al. (2020)Abdeldayem and Aldulaimi (2020), Chakraborty et al. (2020), Oswald et al. (2020), Reddy et al. (2019), Tambe et al. (2019)11
N2311251859

Note(s): N stands for the number of articles

## Table 4

Some recommendations for future research

Research themesPossible research questions
Research direction #1: employee perceptions (and fears) and potential misusesEmployee perceptions of values, attitudes, and behaviors toward the adoption of HRM-focused AI-applications
1. How will the employees’ attitudes and behaviors be affected due to the usage of decision aids like ES, DSS, and their integration with other decision aids?

2. How are decision-makers’s cognitive processes affected by the usage of ES, DSS and other decision aids?

3. What are the perceptions of the critical stakeholders (such as recruiters, managers, frontline employees, and customers) in-service robots (like, in hotels/restaurants) adoption?

4. Based on employees’ perceived behaviors, how is “employees” willingness to work with robots handled by the organizations?

5. Are employees’ perceptions critically dependent on the usability (ease of use) of robots? (can be analyzed using technology acceptance theory)

6. Can we design models as a software module that computerizes employees’ psychological capital measurements?

7. What is the effect on employees’ emotions after the change in the nature of work due to the implementation of AI techniques?

8. Can the usage of robots reduce employee stress (related to performing repetitive tasks)?

For studying and alleviating employees’ fears related to AI adoption, Rogers’s (2003) five steps of the innovation decision process. Questions such as these can be asked-
1. Are the employees exposed to some AI-related innovation’s existence in the HR department?

2. Have they gained some understanding of how it functions?

3. Do they have an unfavorable or favorable attitude toward that innovation?

4. Do they engage in activities that lead to a choice to reject or adopt that innovation?

5. Do they put that innovation into use?

6. Do they seek reinforcement of an innovation decision that has already been made?

7. Do they reverse the previous decision if exposed to conflicting messages about the innovation?

Potential misuses caused by AI applications
1. Can the employee sabotage AI machines not to schedule maintenance in their shift or to distort others’ final performance?

2. Can the factors that are used to develop a model be customized as per the employer’s benefits? (Self-selection Bias, Collider Bias)

3. Can the employer design a decision aid to endorse an optimal plan to save costs only from the employer’s perspective, even at the expense of the employees?

Research direction #2: workforce managementWorkforce management through AI-applications
1. Can a job offer recommender module be developed using semantic nets, where resume analysis will help in sending automatic job recommendations to the suitable applicants, which further can be extended to recommend candidate resumes to employers?

2. Can ANN models be developed for personnel arrangements in accordance with the real-life scenario? (e.g. considering that the personnel’s efficiency of work completion is directly affected by his/her working ability)

3. Can genetic algorithm-based models be developed for specific organizational needs? (e.g. an organization that has a policy of staff transfer at regular intervals of time?

4. Can the data-mining technique-based theoretical models be developed to describe and predict effective employee suggestion strategies?

5. Can data mining techniques be designed for collecting actual turnover reasons and performing root-cause analysis to aid managers in devising effective employee retention policies?

6. Can fuzzy logic be used to develop a relatively exhaustive and smart workforce screening system (framework)? (the system could make use of criteria such as social skills, resource management capacity, personality traits, etc.)

7. How can fuzzy logic-based systems be developed to assist the HRM experts in appraising the current and prospective workforce?

Evaluating performance through AI-applications
1. How can a fuzzy logic HRM evaluation system incorporate longitudinal components in the assessment criteria? How can the evaluation of each criterion be done?

2. How to develop a competency evaluation-based system that can suggest some seminars, workshops, and on-the-job training that may enhance the required skills of the current employees?

3. Can AI be applied to nudge a person in the direction to improve performance? (e.g. use software to schedule resting time)

Research direction #3: evaluating effectiveness and ROIEvaluating the effectiveness of HRM-focused AI applications
1. Can the replacement of the manual method of measuring psychological capital with a logistic regression model save many resources and guarantee accuracy?

2. Can the study of pioneering organizations in the field of HRM-focused robotic technology implementation help in gauging the effectiveness and subsequence of using robots?

3. Can elaboration on the substantiate characteristics and relevance of the focused HR problem(s) and demonstration of the domain success after implementing a technique be done?

4. Can a comparison be made between traditional methods for decision making and a model comprising of a combination of AI techniques to facilitate decision making in different domains of HR?

Return on Investment (ROI)
1. Is the substantial cost of development and implementation of AI techniques like ESs justified?

2. Is the ROI of developing models and computer software rational when employees may not have the skills to access what is required?

Research direction #4: perspective of leadersPerspective of leaders and stakeholders for implementing effective AI solutions
1. How has the perceived value of the usage of robots changed over time (a longitudinal study) for the leaders? What is the role of the classification (e.g. large, medium, or small) and type of the organization (e.g. academic, manufacturing, etc.) in leaders’ perceptions of using robots?

2. What is the difference in leaders’ perceptions of preimplementation and postimplementation of robots?

3. What are the top managers take on AI implementation in HRM?

## References

Abdeldayem, M.M. and Aldulaimi, S.H. (2020), “Trends and opportunities of artificial intelligence in human resource management: aspirations for public sector in Bahrain”, International Journal of Scientific and Technology Research, Scopus, Vol. 9 No. 1, pp. 3867-3871.

Acharyya, S. and Datta, A.K. (2020), “Matching formulation of the staff transfer problem: meta-heuristic approaches”, Opsearch, Vol. 57, pp. 629-668, doi: 10.1007/s12597-019-00432-w.

Akhondzadeh-Noughabi, E., Amin-Naseri, M.R., Albadvi, A. and Saeedi, M. (2016), “Human resource performance evaluation from CRM perspective: a two-step association rule analysis”, International Journal of Business Performance Management, Vol. 17 No. 1, pp. 89-102, doi: 10.1504/IJBPM.2016.073340.

Al-Attar, A. (1990), “Knowledge-base programming in high-level languages”, Information and Software Technology, Vol. 32 No. 2, pp. 127-132.

Alola, U.V. and Atsa'am, D.D. (2020), “Measuring employees' psychological capital using data mining approach”, Journal of Public Affairs, Vol. 20 No. 2, doi: 10.1002/pa.2050.

Arrieta, A.B., Díaz-Rodríguez, N., Del Ser, J., Bennetot, A., Tabik, S., Barbado, A., García, S., Gil-López, S., Molina, D. and Benjamins, R. (2020), “Explainable Artificial Intelligence (XAI): concepts, taxonomies, opportunities and challenges toward responsible AI”, Information Fusion, Vol. 58, pp. 82-115.

Aznoli, F. and Navimipour, N.J. (2017), “Cloud services recommendation: reviewing the recent advances and suggesting the future research directions”, Journal of Network and Computer Applications, Vol. 77, pp. 73-86.

Bankins, S. and Formosa, P. (2020), “When AI meets PC: exploring the implications of workplace social robots and a human-robot psychological contract”, European Journal of Work and Organizational Psychology, Vol. 29 No. 2, pp. 215-229.

Barro, S. and Davenport, T.H. (2019), “People and machines: partners in innovation”, MIT Sloan Management Review, Vol. 60 No. 4, pp. 22-28.

Behera, R.K., Bala, P.K. and Dhir, A. (2019), “The emerging role of cognitive computing in healthcare: a systematic literature review”, International Journal of Medical Informatics, Vol. 129, pp. 154-166.

Berhil, S., Benlahmar, H. and Labani, N. (2020), “A review paper on artificial intelligence at the service of human resources management”, Indonesian Journal of Electrical Engineering and Computer Science, Vol. 18 No. 1, pp. 32-40, doi: 10.11591/ijeecs.v18.i1.pp32-40.

Bohlouli, M., Mittas, N., Kakarontzas, G., Theodosiou, T., Angelis, L. and Fathi, M. (2017), “Competence assessment as an expert system for human resource management: a mathematical approach”, Expert Systems with Applications, Scopus, Vol. 70, pp. 83-102, doi: 10.1016/j.eswa.2016.10.046.

Brendel, A.B., Mirbabaie, M., Lembcke, T.-B. and Hofeditz, L. (2021), “Ethical management of artificial intelligence”, Sustainability, Vol. 13 No. 4, p. 1974.

Byun, D.‐H. and Suh, E.‐H. (1994), “Human resource management expert systems technology”, Expert Systems, Vol. 11 No. 2, pp. 109-119, doi: 10.1111/j.1468-0394.1994.tb00004.x.

Byun, D.-H. and Suh, E.-H. (1996), “Assessing key knowledge representation techniques for use in HRM problem domains”, Expert Systems with Applications, Vol. 11 No. 3, pp. 287-299, doi: 10.1016/S0957-4174(96)00045-0.

Canós, L. and Liern, V. (2004), “Some fuzzy models for human resource management”, International Journal of Technology, Policy and Management, Vol. 4 No. 4, pp. 291-308, doi: 10.1504/IJTPM.2004.006613.

Chakraborty, S., Giri, A., Aich, A. and Biswas, S. (2020), “Evaluating influence of artificial intelligence on human resource management using PLS-SEM (Partial least squares-structural equation modeling)”, International Journal of Scientific and Technology Research, Scopus, Vol. 9 No. 3, pp. 5876-5880.

Chang, V. (2015), “Towards a big data system disaster recovery in a private cloud”, Ad Hoc Networks, Vol. 35, pp. 65-82.

Chang, V. (2021), “An ethical framework for big data and smart cities”, Technological Forecasting and Social Change, Vol. 165, 120559.

Chang, V., Walters, R.J. and Wills, G.B. (2016), “Organisational sustainability modelling—an emerging service and analytics model for evaluating Cloud Computing adoption with two case studies”, International Journal of Information Management, Vol. 36 No. 1, pp. 167-179.

Chen, L.-F. and Chien, C.-F. (2011), “Manufacturing intelligence for class prediction and rule generation to support human capital decisions for high-tech industries”, Flexible Services and Manufacturing Journal, Vol. 23 No. 3, pp. 263-289, doi: 10.1007/s10696-010-9068-x.

Chen, N.-S., Wei, C.-W. and Chen, H.-J. (2008), “Mining e-Learning domain concept map from academic articles”, Computers and Education, Vol. 50 No. 3, pp. 1009-1021.

Chien, C.-F. and Chen, L.-F. (2008), “Data mining to improve personnel selection and enhance human capital: a case study in high-technology industry”, Expert Systems with Applications, Vol. 34 No. 1, pp. 280-290, doi: 10.1016/j.eswa.2006.09.003.

Danping, Z. and Jin, D. (2011), “The data mining of the human resources data warehouse in university based on association rule”, Journal of Computers, Vol. 6 No. 1, pp. 139-146, doi: 10.4304/jcp.6.1.139-146.

Davies, M. (2011), “Concept mapping, mind mapping and argument mapping: what are the differences and do they matter?”, Higher Education, Vol. 62 No. 3, pp. 279-301.

Dean, M.D., Payne, D.M. and Landry, B.J. (2016), “Data mining: an ethical baseline for online privacy policies”, Journal of Enterprise Information Management, Vol. 29 No. 4, pp. 482-504, doi: 10.1108/JEIM-04-2014-0040.

Deane, P., Odendahl, N., Quinlan, T., Fowles, M., Welsh, C. and Bivens‐Tatum, J. (2008), “Cognitive models of writing: writing proficiency as a complex integrated skill”, ETS Research Report Series, Vol. 2008 No. 2, pp. i-36.

Devanna, M.A., Fombrun, C., Tichy, N. and Warren, L. (1982), “Strategic planning and human resource management”, Human Resource Management, Vol. 21 No. 1, pp. 11-17.

Diao, Y. and Shwartz, L. (2017), “Building automated data driven systems for IT service management”, Journal of Network and Systems Management, Scopus, Vol. 25 No. 4, pp. 848-883, doi: 10.1007/s10922-017-9430-3.

Dulebohn, J.H. and Johnson, R.D. (2013), “Human resource metrics and decision support: a classification framework”, Human Resource Management Review, Vol. 23 No. 1, pp. 71-83.

Elacio, A.A., Lacatan, L.L., Vinluan, A.A. and Balazon, F.G. (2020), “Machine learning integration of Herzberg’s theory using c4.5 algorithm”, International Journal of Advanced Trends in Computer Science and Engineering, Scopus, Vol. 9 No. 1.1 Special Issue, pp. 57-63, doi: 10.30534/ijatcse/2020/1191.12020.

Escolar-Jimenez, C.C., Matsuzaki, K., Okada, K. and Gustilo, R.C. (2019), “Data-driven decisions in employee compensation utilizing a neuro-fuzzy inference system”, International Journal of Emerging Trends in Engineering Research, Vol. 7 No. 8, pp. 163-169, doi: 10.30534/ijeter/2019/10782019.

Fahimnia, B., Sarkis, J. and Davarzani, H. (2015), “Green supply chain management: a review and bibliometric analysis”, International Journal of Production Economics, Vol. 162 No. C, pp. 101-114.

Ferrell, O.C. and Ferrell, L. (2021), “Applying the Hunt Vitell ethics model to artificial intelligence ethics”, Journal of Global Scholars of Marketing Science, Vol. 31 No. 2, pp. 178-188.

Fountaine, T., McCarthy, B. and Saleh, T. (2019), “Building the AI-powered organization”, Harvard Business Review, Vol. 97 No. 4, pp. 62-73, available at: https://hbr.org/2019/07/building-the-ai-powered-organization.

Fowler, J.W., Wirojanagud, P. and Gel, E.S. (2008), “Heuristics for workforce planning with worker differences”, European Journal of Operational Research, Vol. 190 No. 3, pp. 724-740, doi: 10.1016/j.ejor.2007.06.038.

Gupta, P., Fernandes, S.F. and Jain, M. (2018), “Automation in recruitment: a new Frontier”, Journal of Information Technology Teaching Cases, Scopus, Vol. 8 No. 2, pp. 118-125, doi: 10.1057/s41266-018-0042-x.

Hajiha, A., Alavi, S.H. and Makvandi, P. (2006), “Improving behavioral based job assignment by using radial basis function network”, WSEAS Transactions on Systems, Scopus, Vol. 5 No. 11, pp. 2542-2547.

Han, Y. (2016), “Improved BIRCH clustering algorithm and human resource management efficiency: an organizational learning perspective”, International Journal of Security and Its Applications, Vol. 10 No. 8, pp. 385-394, doi: 10.14257/ijsia.2016.10.8.34.

Han, J., Kamber, M. and Pei, J. (2011), Data Mining: Concepts and Techniques, 3rd ed., Morgan Kaufmann, Waltham, MA.

Hirsch, J.E. (2005), “An index to quantify an individual's scientific research output”, Proceedings of the National Academy of Sciences, Vol. 102 No. 46, pp. 16569-16572.

Hochrein, S. and Glock, C.H. (2012), “Systematic literature reviews in purchasing and supply management research: a tertiary study”, International Journal of Integrated Supply Management, Vol. 7 No. 4, pp. 215-245.

Holland, J.H. (1992), Adaptation in Natural and Artificial Systems: An Introductory Analysis with Applications to Biology, Control, and Artificial Intelligence, MIT Press.

Hooper, R.S., Galvin, T.P., Kilmer, R.A. and Liebowitz, J. (1998), “Use of an expert system in a personnel selection process”, Expert Systems with Applications, Scopus, Vol. 14 No. 4, pp. 425-432.

Hosmer, D.W. Jr, Lemeshow, S. and Sturdivant, R.X. (2013), Applied Logistic Regression, John Wiley & Sons, Vol. 398.

Hsieh, H.-F. and Shannon, S.E. (2005), “Three approaches to qualitative content analysis”, Qualitative Health Research, Vol. 15 No. 9, pp. 1277-1288.

Huang, L.C., Wu, P., Kuo, R.J. and Huang, H.C. (2001), “A neural network modelling on human resource talent selection”, International Journal of Human Resources Development and Management, Scopus, Vol. 1 Nos 2-4, pp. 206-219.

Huang, M.-J., Tsou, Y.-L. and Lee, S.-C. (2006), “Integrating fuzzy data mining and fuzzy artificial neural networks for discovering implicit knowledge”, Knowledge-Based Systems, Scopus, Vol. 19 No. 6, pp. 396-403, doi: 10.1016/j.knosys.2006.04.003.

Jantan, H., Hamdan, A.R. and Othman, Z.A. (2009), “Knowledge discovery techniques for talent forecasting in human resource application”, World Academy of Science, Engineering and Technology, Vol. 38, pp. 803-811.

Jantan, H., Azemi, N.H.S.C. and Othman, Z.A. (2019), “Artificial immune-based algorithm for academic leadership assessment”, International Journal of Advanced Computer Science and Applications, Vol. 10 No. 8, pp. 364-370.

Jerman, A., Pejić Bach, M. and Aleksić, A. (2020), “Transformation towards smart factory system: examining new job profiles and competencies”, Systems Research and Behavioral Science, Vol. 37 No. 2, pp. 388-402, doi: 10.1002/sres.2657.

Jia, Q., Guo, Y., Li, R., Li, Y. and Chen, Y. (2018), “A conceptual artificial intelligence application framework in human resource management”, International Consortium for Electronic Business, Scopus, in Chang, F.-K., Li, E.Y. and Li, E.Y. (Eds), Proceedings of the International Conference on Electronic Business (ICEB), Vol. 2018, pp. 106-114, available at: https://www.scopus.com/inward/record.uri?eid=2-s2.0-85061925662&partnerID=40&md5=5b8c7463a6202415e7d4c010ecee3ef0.

Kaplan, A. and Haenlein, M. (2019), “Siri, Siri, in my hand: who's the fairest in the land? On the interpretations, illustrations, and implications of artificial intelligence”, Business Horizons, Vol. 62 No. 1, pp. 15-25.

Karam, S., Nagahi, M., Dayarathna, V.L., Ma, J., Jaradat, R. and Hamilton, M. (2020), “Integrating systems thinking skills with multi-criteria decision-making technology to recruit employee candidates”, Expert Systems with Applications, Vol. 160, p. 113585, doi: 10.1016/j.eswa.2020.113585.

Karatop, B., Kubat, C. and Uygun, Ö. (2015), “Talent management in manufacturing system using fuzzy logic approach”, Computers and Industrial Engineering, Vol. 86, pp. 127-136, doi: 10.1016/j.cie.2014.09.015.

Khatri, N., Fern, C.T. and Budhwar, P. (2001), “Explaining employee turnover in an Asian context”, Human Resource Management Journal, Vol. 11 No. 1, pp. 54-74.

Kimseng, T., Javed, A., Jeenanunta, C. and Kohda, Y. (2020), “Applications of fuzzy logic to reconfigure human resource management practices for promoting product innovation in formal and non-formal R&D Firms”, Journal of Open Innovation: Technology, Market, and Complexity, Vol. 6 No. 2, doi: 10.3390/JOITMC6020038.

Krippendorff, K. (2018), Content Analysis: An Introduction to its Methodology, Sage Publications.

Lamarca, B.I.J. and Ambat, S.C. (2018), “The development of a performance appraisal system using decision tree analysis and Fuzzy Logic”, International Journal of Intelligent Engineering and Systems, Vol. 11 No. 4, pp. 11-19, doi: 10.22266/ijies2018.0831.02.

Lawler, J.J. and Elliot, R. (1996), “Artificial intelligence in HRM: an experimental study of an expert system”, Journal of Management, Scopus, Vol. 22 No. 1, pp. 85-111, doi: 10.1016/S0149-2063(96)90013-6.

Lee, D. and Ahn, C. (2020), “Industrial human resource management optimization based on skills and characteristics”, Computers and Industrial Engineering, Scopus, Vol. 144, doi: 10.1016/j.cie.2020.106463.

Lin, Y., Wang, X. and Xu, R. (2020), “Semi-supervised human resource scheduling based on deep presentation in the cloud”, EURASIP Journal on Wireless Communications and Networking, Vol. 2020 No. 1, doi: 10.1186/s13638-020-01677-6.

Lopes, S.A., Duarte, M.E. and Almeida Lopes, J. (2018), “Can artificial neural networks predict lawyers' performance rankings?”, International Journal of Productivity and Performance Management, Scopus, Vol. 67 No. 9, pp. 1940-1958, doi: 10.1108/IJPPM-08-2017-0212.

Lucia-Casademunt, A.M., Ariza-Montes, A. and Becerra-Alonso, D. (2013), “Exploring emotional involvement in workplace by applying artificial neural networks: European study”, Actual Problems of Economics, Scopus, Vol. 143 No. 5, pp. 376-385.

Malik, A. (2009), “Training drivers, competitive strategy and clients’ needs: case studies of three business process outsourcing organisations”, Journal of European Industrial Training, Vol. 33 No. 2, pp. 160-177, doi: 10.1108/03090590910939058.

Malik, A., Srikanth, N.R. and Budhwar, P. (2020), “Digitisation, AI and HRM”, in Jonathan, C., Pawan, B. and Ann, D. (Eds), Human Resource Management: Strategic and International Perspectives, Sage, pp. 88-110.

Mallett, R., Hagen-Zanker, J., Slater, R. and Duvendack, M. (2012), “The benefits and challenges of using systematic reviews in international development research”, Journal of Development Effectiveness, Vol. 4 No. 3, pp. 445-455.

Manafi, M. and Subramaniam, I.D. (2015), “Balancing performance by human resource management practices”, Asian Social Science, Scopus, Vol. 11 No. 10, pp. 386-400, doi: 10.5539/ass.v11n10p386.

Maree, M., Kmail, A.B. and Belkhatir, M. (2019), “Analysis and shortcomings of e-recruitment systems: towards a semantics-based approach addressing knowledge incompleteness and limited domain coverage”, Journal of Information Science, Vol. 45 No. 6, pp. 713-735, doi: 10.1177/0165551518811449.

Markham, K.M., Mintzes, J.J. and Jones, M.G. (1994), “The concept map as a research and evaluation tool: further evidence of validity”, Journal of Research in Science Teaching, Vol. 31 No. 1, pp. 91-101.

Marksberry, P., Church, J. and Schmidt, M. (2014), “The employee suggestion system: a new approach using latent semantic analysis”, Human Factors and Ergonomics in Manufacturing, Vol. 24 No. 1, pp. 29-39, doi: 10.1002/hfm.20351.

Masum, A.-K., Beh, L.-S., Azad, A.-K. and Hoque, K. (2018), “Intelligent human resource information system (i-HRIS): a holistic decision support framework for HR excellence”, The International Arab Journal of Information Technology, Scopus, Vol. 15 No. 1, pp. 121-130.

Mavi, N.K. and Mavi, R.K. (2014), “Talent pool membership in sport organisations with fuzzy analytic hierarchy process”, International Journal of Logistics Systems and Management, Vol. 17 No. 1, pp. 1-21, doi: 10.1504/IJLSM.2014.057978.

Min, H. (2010), “Artificial intelligence in supply chain management: theory and applications”, International Journal of Logistics Research and Applications, Vol. 13 No. 1, pp. 13-39, doi: 10.1080/13675560902736537.

Moon, C., Lee, J. and Lim, S. (2010), “A performance appraisal and promotion ranking system based on fuzzy logic: an implementation case in military organizations”, Applied Soft Computing Journal, Vol. 10 No. 2, pp. 512-519, doi: 10.1016/j.asoc.2009.08.035.

Moradinasab, N., Soofifard, R. and Asili, G.R. (2016), “Developing an effective mathematical model for leadership styles selection by using fuzzy logic: a case on RIPI HR characteristics”, International Journal of Productivity and Quality Management, Vol. 19 No. 4, pp. 466-484, doi: 10.1504/IJPQM.2016.080153.

Nawaz, N. (2019), “Artificial intelligence interchange human intervention in the recruitment process in Indian software industry”, International Journal of Advanced Trends in Computer Science and Engineering, Scopus, Vol. 8 No. 4, pp. 1433-1442, doi: 10.30534/ijatcse/2019/62842019.

Nazri, M.Z.A., Ghani, R.A., Abdullah, S., Ayu, M. and Nor Samsiah, R. (2019), “Predicting academician publication performance using decision tree”, International Journal of Recent Technology and Engineering, Vol. 8 No. 2, pp. 180-185.

Niu, L., Lu, J. and Zhang, G. (2009), “Cognition-driven decision support for business intelligence”, Models, Techniques, Systems and Applications. Studies in Computational Intelligence, Springer, Berlin, pp. 4-5.

Oswald, F.L., Behrend, T.S., Putka, D.J. and Sinar, E. (2020), “Big data in industrial-organizational Psychology and human resource management: forward progress for organizational research and practice”, Annual Review of Organizational Psychology and Organizational Behavior, Scopus, Vol. 7, pp. 505-533, doi: 10.1146/annurev-orgpsych-032117-104553.

Paladini, E.P. (2009), “A fuzzy approach to compare human performance in industrial plants and service-providing companies”, WSEAS Transactions on Business and Economics, Vol. 6 No. 11, pp. 557-569.

Partridge, D. and Hussain, K.M. (1992), Artificial Intelligence and Business Management, Intellect Books, Ablex Publishing Corporation, Norwood, New Jersey.

Pereira, V., Sharma, K. and Malik, A. (2013), “Colliding conflicting perspectives: the rhetoric and reality of employee turnover-evidence from the Indian BPO industry”, 3rd Biennial Conference of the Indian Academy of Management (IAM), 2013 held at IIMA during 12-14 December.

Pickering, C. and Byrne, J. (2014), “The benefits of publishing systematic quantitative literature reviews for PhD candidates and other early-career researchers”, Higher Education Research and Development, Vol. 33 No. 3, pp. 534-548.

Rab-Kettler, K. and Lehnervp, B. (2019), “Recruitment in the times of machine learning”, Management Systems in Production Engineering, Scopus, Vol. 27 No. 2, pp. 105-109, doi: 10.1515/mspe-2019-0018.

Rajpurohit, N., Saxena, M., Yadav, R. and Chande, P.K. (2020), “Investigating impact of artificial intelligence in deployment of effective project teams”, International Journal of Advanced Science and Technology, Scopus, Vol. 29 No. 8, pp. 382-391.

Ramos‐Rodríguez, A.-R. and Ruíz‐Navarro, J. (2004), “Changes in the intellectual structure of strategic management research: a bibliometric study of the Strategic Management Journal, 1980–2000”, Strategic Management Journal, Vol. 25 No. 10, pp. 981-1004, doi: 10.1002/smj.397.

Ranjan, J., Goyal, D.P. and Ahson, S.I. (2008), “Data mining techniques for better decisions in human resource management systems”, International Journal of Business Information Systems, Vol. 3 No. 5, pp. 464-481, doi: 10.1504/IJBIS.2008.018597.

Reddy, A.J.M., Rani, R. and Chaudhary, V. (2019), “Technology for sustainable HRM: an empirical research of health care sector”, International Journal of Innovative Technology and Exploring Engineering, Vol. 9 No. 1, pp. 2919-2924, doi: 10.35940/ijitee.A9108.119119.

Richter, Y., Naveh, Y., Gresh, D.L. and Connors, D.P. (2008), “Optimatch: applying constraint programming to workforce management of highly skilled employees”, International Journal of Services Operations and Informatics, Scopus, Vol. 3 Nos 3-4, pp. 258-270, doi: 10.1504/IJSOI.2008.021338.

Rogers, E.M. (2003), Diffusion of Innovations, 5th ed, Free Press, New York, NY.

Roth, W.-M. and Roychoudhury, A. (1993), “The concept map as a tool for the collaborative construction of knowledge: a microanalysis of high school physics students”, Journal of Research in Science Teaching, Vol. 30 No. 5, pp. 503-534.

Ryan, M. and Stahl, B.C. (2021), “Artificial intelligence ethics guidelines for developers and users: clarifying their content and normative implications”, Journal of Information, Communication and Ethics in Society, Vol. 19 No. 1, pp. 61-68, doi: 10.1108/JICES-12-2019-0138.

Saidi Mehrabad, M. and Fathian Brojeny, M. (2007), “The development of an expert system for effective selection and appointment of the jobs applicants in human resource management”, Computers and Industrial Engineering, Scopus, Vol. 53 No. 2, pp. 306-312, doi: 10.1016/j.cie.2007.06.023.

Seal, A. (2019), “The most interesting AI statistics of 2019”, available at: https://www.vxchnge.com/blog/ai-statistics.

Secundo, G., Ndou, V., Del Vecchio, P. and De Pascale, G. (2020), “Sustainable development, intellectual capital and technology policies: a structured literature review and future research agenda”, Technological Forecasting and Social Change, Elsevier, Vol. 153, p. 119917.

Selden, S.C., Jacobson, W., Ammar, S.H. and Wright, R.H. (2000), “A new approach to assessing performance of state human resource management systems: a multi-level fuzzy rule-based system”, Review of Public Personnel Administration, Vol. 20 No. 3, pp. 58-74, doi: 10.1177/0734371X0002000305.

Sexton, R.S., McMurtrey, S., Michalopoulos, J.O. and Smith, A.M. (2005), “Employee turnover: a neural network solution”, Computers and Operations Research, Vol. 32 No. 10, pp. 2635-2651.

Shahhosseini, V. and Sebt, M.H. (2011), “Competency-based selection and assignment of human resources to construction projects”, Scientia Iranica, Scopus, Vol. 18 No. 2 A, pp. 163-180, doi: 10.1016/j.scient.2011.03.026.

Siau, K. and Wang, W. (2020), “Artificial intelligence (AI) ethics: ethics of AI and ethical AI”, Journal of Database Management, Vol. 31 No. 2, pp. 74-87.

Stavrou, E.T., Charalambous, C. and Spiliotis, S. (2007), “Human resource management and performance: a neural network analysis”, European Journal of Operational Research, Scopus, Vol. 181 No. 1, pp. 453-467, doi: 10.1016/j.ejor.2006.06.006.

Strohmeier, S. and Piazza, F. (2013), “Domain driven data mining in human resource management: a review of current research”, Expert Systems with Applications, Vol. 40 No. 7, pp. 2410-2420, doi: 10.1016/j.eswa.2012.10.059.

Strohmeier, S. and Piazza, F. (2015), “Artificial intelligence techniques in human resource management—a conceptual exploration”, Intelligent Systems Reference Library, Springer Science and Business Media Deutschland GmbH, Scopus, Vol. 87, p. 172, doi: 10.1007/978-3-319-17906-3_7.

Sturman, M.C., Hannon, J.M. and Milkovich, G.T. (1996), “Computerized decision aids for flexible benefits decisions: the effects of an expert system and decision support system on employee intentions and satisfaction with benefits”, Personnel Psychology, Vol. 49 No. 4, pp. 883-908, doi: 10.1111/j.1744-6570.1996.tb02453.x.

Tambe, P., Cappelli, P. and Yakubovich, V. (2019), “Artificial intelligence in human resources management: challenges and a path forward”, California Management Review, Vol. 61 No. 4, pp. 15-42.

Tranfield, D., Denyer, D. and Smart, P. (2003), “Towards a methodology for developing evidence-informed management knowledge by means of systematic review”, British Journal of Management, Vol. 14 No. 3, pp. 207-222, doi: 10.1111/1467-8551.00375.

Tursunbayeva, A., Di Lauro, S. and Pagliari, C. (2018), “People analytics—a scoping review of conceptual boundaries and value propositions”, International Journal of Information Management, Vol. 43, pp. 224-247, doi: 10.1016/j.ijinfomgt.2018.08.002.

Tursunbayeva, A., Pagliari, C., Di Lauro, S. and Antonelli, G. (2021), “The ethics of people analytics: risks, opportunities and recommendations”, Personnel Review.

Varalakshmi, R. and Dhivya, R.S. (2019), “A survey on big data applicability in prediction using absence information for workforce management”, International Journal of Recent Technology and Engineering, Scopus, Vol. 7 No. 5, pp. 97-100.

Vinichenko, M.V., Rybakova, M.V., Сhulanova, O.L., Kuznetsova, I.V., Makushkin, S.A. and Lobacheva, A.S. (2019), “Using natural and artificial intelligence in the talent management system”, International Journal of Recent Technology and Engineering, Scopus, Vol. 8 No. 3, pp. 7417-7423, doi: 10.35940/ijrte.C6152.098319.

Von Krogh, G., Rossi-Lamastra, C. and Haefliger, S. (2012), “Phenomenon-based research in management and organisation science: when is it rigorous and does it matter?”, Long Range Planning, Vol. 45 No. 4, pp. 277-298.

Wallace, J.D. and Mintzes, J.J. (1990), “The concept map as a research tool: exploring conceptual change in biology”, Journal of Research in Science Teaching, Vol. 27 No. 10, pp. 1033-1052.

Wan Chik, W.N.A. and Arokiasamy, L. (2019), “Perceived higher education climate of academics in Malaysian private institutions in industry 4.0”, Global Business and Management Research: International Journal, Vol. 11 No. 1, pp. 488-504.

Wi, H., Oh, S., Mun, J. and Jung, M. (2009), “A team formation model based on knowledge and collaboration”, Expert Systems with Applications, Vol. 36 No. 5, pp. 9121-9134, doi: 10.1016/j.eswa.2008.12.031.

Wright, P.M. and Snell, S.A. (1991), “Toward an integrative view of strategic human resource management”, Human Resource Management Review, Vol. 1 No. 3, pp. 203-225.

Xu, S., Stienmetz, J. and Ashton, M. (2020), “How will service robots redefine leadership in hotel management? A Delphi approach”, International Journal of Contemporary Hospitality Management, Vol. 32 No. 6, pp. 2217-2237, doi: 10.1108/IJCHM-05-2019-0505.

Yong-Hak, J. (2013), “The definitive resource for global research web of science access powerful cited reference searching and multidisciplinary content”, available at: https://libraryresearch.weebly.com/uploads/3/1/ 9/5/3195843/web_of_science_factsheet.pdf (accessed on 16 July 2020).

Zadeh, L.A. (1965), “Fuzzy sets”, Information and Control, Vol. 8 No. 3, pp. 338-353.

Zhang, H., Yuan, W. and Jiang, H. (2012), “Performance evaluation on human resource management of China's commercial banks based on improved Bp neural networks”, International Journal of Advancements in Computing Technology, Vol. 4 No. 11, pp. 304-310, doi: 10.4156/ijact.vol4.issue11.32.

## Corresponding author

Charbel Jose Chiappetta Jabbour can be contacted at: cchiappettajabbour@lincoln.ac.uk