Keywords

1 Introduction

After the Covid-19 pandemic, it became clear how important it can be to identify emerging trends in Human Resources Management (HRM), as it enables organizations to stay ahead of the curve, adapt to the changing needs and expectations of employees, and proactively implement innovative practices that improve employee engagement, talent acquisition, and overall organizational effectiveness [1, 2]. There are several approaches that can be adopted to reveal these trends, and bibliometric analysis is one that helps researchers identify emerging research trends within the field. By analyzing publication patterns, citation networks, and keyword co-occurrence, researchers can gain insight into the most prominent and influential topics and themes in human resources research. This knowledge can guide future research directions and highlight areas that require further exploration. Since ChatGPT emerged, Artificial Intelligence (AI) tools have been compared and used to develop this type of analysis, leading to question if the gains from this new approach overcome the ones from traditional bibliometric analysis. Thus, this paper, employing both a traditional approach and a combined use of AI tools, aims to unveil the gains and throwbacks of both approaches, using the research developed in the HRM field after Covid-19 as the scope of analysis. This work provides a comparative snapshot of the research conducted in this field by (i) reviewing the concepts related to HRM, (ii) analyzing the research reported in a three years window using a traditional versus an AI method, and (iii) unveiling research gains and future directions.

2 Literature Review

2.1 Human Resources Management

Human Resources Management (HRM) is the strategic and operational process of effectively managing the workforce to achieve organizational goals and objectives, encompassing several activities and responsibilities related to the management and development of human capital within an organization. Its main focus is to optimize employees’ performance and productivity, ensuring their well-being and job satisfaction [3,4,5,6]. Different definitions and interpretations of the concept of HRM arise from various perspectives and evolving organizational practices, depending on the context, the organization’s size, industry, and management philosophy [7, 8]. It is essential to recognize that HRM is a dynamic field, where business practices and employee expectations evolve; in addition, different organizations may adopt a combination of these perspectives to suit their specific needs and organizational culture. The COVID-19 pandemic has brought about significant challenges for HRM, many resulting from the sudden and unexpected changes in the work environment and the workforce [1, 2, 9,10,11]. Navigating these challenges requires HRM to be adaptable, empathetic, and proactive. By embracing new strategies and technologies, focusing on employee well-being, and maintaining open communication with the workforce, Human Resources (HR) professionals can play a critical role in helping organizations succeed in the post-pandemic period.

2.2 Bibliometric Analysis as a Source of Knowledge

Bibliometric analysis is a valuable source of knowledge that utilizes quantitative and statistical methods to assess and analyze scholarly publications and their citation patterns [12]. It provides researchers with an overview of the research landscape in a specific field, including the volume of publications, growth rate, and distribution of research across journals and institutions. The analysis uncovers emerging research trends and hotspots by examining keyword co-occurrence and citation networks, guiding researchers toward relevant and influential topics. Overall, bibliometric analysis enriches our understanding of scholarly research, facilitates collaboration, and guides future research directions. To draw accurate and meaningful conclusions, researchers must interpret bibliometric data cautiously, considering its limitations and potential biases. Human and AI bibliometric analysis differ in critical aspects, including data collection, processing, analysis, and interpretative capabilities (see Table 1).

Table 1 Differences between Human and AI bibliometric analysis
Full size table

3 Study Methods

Research in HRM extends beyond journals in the HR field, leading to scattered relevant materials across various academic publications, becoming the perfect field to conduct this study. To address this issue, the present study incorporates three stages: (i) traditional bibliometric analysis; (ii) AI analysis; and (iii) a comparison of the results of the first two phases. A traditional bibliometric analysis systematically gathers, evaluates, and analyzes scholarly publications and their citation patterns [12]. In this work, a general protocol comprehending five stages was adopted: (1) The research objectives were set by clearly defining the scope of the bibliometric analysis, focusing on HRM articles published between 2020 and 2023; (2) to identify relevant publications, a comprehensive literature search was conducted using the Web of Science database, known for its academic credibility, to enable a comparison of outcomes [13]; (3) data collection involved extracting essential information, such as publication details and citation data, from chosen publications; (4) data cleaning and pre-processing were performed to eliminate duplicates, correct errors, and ensure data consistency; (5) the selection of bibliometric indicators included citation counts, h-index, journal impact factors, co-citation analysis, and keyword co-occurrence analysis.

On the other hand, a bibliometric analysis using AI tools can be a powerful and efficient way to gain insights into scholarly publications and their citation patterns faster. When conducting a bibliometric analysis using AI tools, some of the protocol tasks are similar to those adopted in the traditional bibliometric analysis, but processed by AI tools, as seen below: (1) clearly define research objectives; (2) for data collection, researchers can utilize AI tools like SciScape, Iris, Litmap, and Dimensions, offering access to vast scholarly data, including publications and citations; (3) input data depends on the chosen AI tool—researchers can specify parameters like keywords, authors, and publication years to gather relevant publications; (4) AI-driven data analysis autonomously conducts bibliometric analysis, extracting bibliographic information, performing network analysis, and identifying trends; (5) built-in bibliometric indicators are used; (6) AI tools provide visualizations, such as network maps and trend graphs, for a more intuitive understanding of the research landscape; (7) interpretation involves analyzing AI-generated results to gain insights into influential authors, publications, and emerging trends in the field. The third phase of this study is concerned with comparing the results of the two bibliometric analyses. In both approaches, the researcher must draw meaningful conclusions from the analysis and relate them to the initial research objectives. The same happens in addressing any limitations of the study and suggesting potential future research directions, keeping in mind the specific capabilities and limitations of the AI tools used and the database access limits that the traditional approach may have.

4 Results

The traditional process on Web of Science used as search words “human resources” and presented a total result of 3.656 entries. The highly cited papers’ filter was actioned to find the most relevant papers, totalling 269 entries; each entry was analyzed (title and/or abstract) to learn that only 10 papers reconciled both human resources and covid themes. Due to this result, it was detected that most of the papers presented only one of the keywords searched, motivating a refinement of the search. The publication year of the 172 papers obtained varies between 2020 (total 33), 2021 (total 126), and 2022 (total 13). Regarding the field of study of the source (journal), this item presents a broad spectrum of themes as the following results attest: Business, Development Studies, Economics, Education Educational Research, Ergonomics, Ethics, Gerontology, Health Policy Services, Healthcare Sciences Services, Hospitality Leisure Sport Tourism, HR, HR for Health, Industrial Relations Labour, Management, Medicine Research Experimental, Multidisciplinary Sciences, Nursing, Psychiatry, Psychology, Public Administration, Public Environmental Occupational Health, Social Sciences Interdisciplinary, and Urban Studies.

Looking at the paper’s content, it is possible to notice that over 90% of the articles were published in thematic journals. Human resources and human resources for health were the main fields of publication (21.5%), followed by business and management fields (12.7%) and multidisciplinary sciences (9.7%). The publications are mainly produced in the USA, followed by China, England, Canada, and India. After the initial descriptive analysis, a keyword co-occurrence network was developed using VosViewer software (see Fig. 1). To better understand these streams of research, a clustering graph was composed based on abstracts and titles, showing a total of three clusters (see Fig. 1): Cluster 1 (red) is formed of 45 items; Cluster 2 (green), 23 items; and Cluster 3 (blue), 19 items. In Cluster 1, the item resource stands out; this item correlates more expressively within the cluster with the items analysis and data and the other two clusters, namely, with the items change, HRM, employee, organization, and work. Cluster 1 presents the research focus on human resources—data analysis and the inherent research process. Cluster 2 is more focused on the human resource development (HRD) issue given by employee, work, and organization items. HRD is a central issue in the research under study; it is also a pillar of organizations in what concerns their sustainability, the increasing importance of human resources (employees), and their sense of belonging and loyalty to an employer. Cluster 3 is centered on change, context, and HRM. HRM is vital in recognizing all contextual influences on an organization and its employees. The context of change is synonymous with adapting all stakeholders for the general good. HR departments and/or responsible people assume responsibility for managing different interests that should culminate in a healthy and prosperous organization supported by happy and fulfilled employees.

Fig. 1
A color-coded network visualization map of the keywords related to the streams of research. They are connected, with lines representing connections between them. It has 3 clusters. Some of the keywords are resource, management, period, factor, sample, change, employee, article, and organization.

Keyword co-occurrence network

Full size image

The outcome of the AI bibliometric tools to the same research query, using the following path integrating the different AI tools, Dimension, Litmap, and Iris, was the following (Table 2).

Table 2 Number of articles found in Dimensions.ai
Full size table

The number of works found was higher than in the traditional bibliometric process, even delimiting only to commerce, management, tourism, and services. Using the results of the analysis and the Litmap application, a co-authorship map was found (see Fig. 2). Litmap maps two strong clusters anchored in (1) the risks driven by broader technology adoption and (2) the new challenges teleworking poses. The map on Fig. 3 was produced by combining these outcomes and using the Iris tool. The left side (blue squares) shows the external elements conditioning firms’ HRM, ranging from financial constraints, economic and global recession, and public and healthcare issues. According to the different authors, acknowledging the challenges posed by the external environment can be useful to better and quickly adjust to new work conditions and leverage some of the characteristics of the people working in the firm.

Fig. 2
A thematic co-occurrence network map of co-authors related to H R M. Some of the authors are Harney 2022, Arora 2020, Kaushik 2020, Cooke 2021, Jiang 2018, Adam 2021, Jackson 2014, Wang 2020, and Caligiuri 2022.

Authors and thematic co-occurrence network

Full size image
Fig. 3
A rectangular-shaped thematic map is divided into 7 parts with 4 gradient shades. On top are 2 squares containing elements related to adjustment and resources. Below are 5 blocks containing elements related to crisis, pandemic, competency, leadership, and H R D.

Thematic predominance and network

Full size image

Nonetheless, when focusing on the right side of Fig. 3, the HR competencies and leadership capabilities, combined with HRD, are also critical trends.

5 Discussion and Conclusions

Traditional bibliometric analysis used data retrieved from the Web of Science. Most of the articles published focus on the North American and European realities, and the most researched topics included the adjustments needed and a set of differential issues that arise from remote work. The topics could be aggregated into nine clusters using VosViewer. This analysis allowed to identify as future HR trends the employee well-being, the adoption of remote and hybrid work models, the enhancing employee experience, the offer of flexible benefits, and the integration of HR technology seamlessly.

The AI approach was conducted using a combination of three different AI tools (SciScape, Iris, and Litmap). The outcome produced came from a wider range of data sources, and the Europe and America dichotomy disappeared. Nonetheless, there was a convergence in terms of gaps and future trends mapped, with the emphasis on the gains from the use of AI and automation, data-driven decision-making, upskilling of employees, and focusing on diversity and inclusion.

While AI offers numerous advantages, it is important to note that traditional bibliometric analysis still provides valuable insights and may be preferred in certain situations where manual curation or context-specific knowledge is crucial. The choice between AI and traditional methods depends on the specific research objectives, data availability, and the expertise and resources available to researchers.

This study contributes to the advancement of bibliometrics as a field of study and potentially enhancing the quality of research evaluation and decision-making, by pointing out the gains and differences between the two methods. Considering that AI tools can identify connections and patterns across different disciplines, facilitating interdisciplinary collaborations and knowledge transfer, future research should try to assess how this interdisciplinary approach can be valued by firms and policy makers.