Artificial Intelligence in Human Resources Management

Seventy-seven percent of the consumers are knowingly or unknowingly using AI technologies ranging from interactive chatbots and smart wearables to personal digital assistants.

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. 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.

In about 20 years, 50% of jobs will be outdated or not needed anymore. More than 800 million employees worldwide will lose their jobs (one-fifth of the global workforce) and be replaced by a robot until 2030, according to the McKinsey Global Institute study conducted in 46 countries.

Primary school students have to be prepared for change because by 2030, 85% they will work in careers that do not yet exist. However, over the next five years 'growth' will be included in the rate of using artificial intelligence as an indicator within global economic growth indicators as well per capita technology of the Fourth Industrial Revolution, along with rates national income, GDP, inflation, and other indicators which measures the state's economic power.

Since the late 1970s, AI has demonstrated tremendous potential in improving human decision-making processes and in the ensuing efficiency in various business settings. 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. 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.

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). AI has been successfully applied in various HRM functions such as human resource (HR) performance evaluation, employee selection, employee turnover, prediction of the level of employees’ emotional involvement, and employee assignment.

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.

Artificial Intelligence (AI) is increasingly adopted within Human Resource management (HRM) due to its potential to create value for consumers, employees, and organisations. However, recent studies have found that organisations are yet to experience the anticipated benefits from AI adoption, despite investing time, effort, and resources. The existing studies in HRM have examined the applications of AI, anticipated benefits, and its impact on human workforce and organisations.

The most prevalent AI techniques in the reviewed literature are fuzzy logic, data mining and expert systems. The row of “other techniques” includes the articles that do not deal with one AI technique but discuss the global influence AI has on HRM.

AI in staffing

The application of IoT in HRM involves changes and modifications in HR technologies (hardware, software, and data), HR activities (flexibilization of employee working times, improvement of employee performance, personalization of employee working environments) and HR actors (tasks and qualifications). Employee self-service (ESS) technology allows employees to manage their own data rather than rely on HR professionals and register for training with the objective of efficiency-related gains. Electronic performance monitoring (EPM) has the potentiality to change various HR practices including evaluation, selection, and training.

Many EPM forms are already widely used (e.g., call and internet usage monitoring, electronic medication administration records) and it is increasingly argued that technologies such as microchip wrist implants and body heat sensor desk hardware may be the future of work monitoring. Algorithmic technologies can also help employers direct, evaluate and discipline workers.

Of contemporary interest are virtualization technologies virtual representation of individuals who interact with each other in 3D digital environments. Although they have emerged from the computer games industry, they are increasingly being used to enhance interpersonal and organizational interaction and facilitate organizational learning.

Reviewing the functions shows detailed research attention on staffing – 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).

The first and the largest subcategory of research contributions concerns employees’ selection. This subcategory addressed selection either based on competency, or a specific employee segment (such as army personnel), or with the focus on reduction of cost and time involved in the selection process, or based on performance ranking, or predicting work behaviors.?Another subcategory discussed automation in the staffing of employees, followed by problems of the likelihood of absenteeism by the employee, allocation of employees to specific jobs, employee transfer problem, and finally a study on understanding the importance of AI in talent acquisition.

Research reveals that companies establish an e-recruiting system to reduce costs, to access more people, get quicker response and increase applicants’ positive perceptions about the company Automatic matching between job offers and suitable candidate profiles provides several advantages including reduced effort (in terms of cost and time) and elimination of the need for HR professionals to have knowledge pertaining to a specific professional field or skill. However, although the trend towards the use of information technology for e-recruitment has transformed the way firms recruit, select, and retain employees several issues impede its effectiveness.

AI in compensation

Compensation focuses on the employees’ remuneration discretionally complemented by equity sharing, profit, or both. AI can facilitate extensive numeric data-driven decision-making to design an efficient AI-based compensation system.?This subfunction of HRM deals with an enormous amount of numeric data, AI could potentially design a robust compensation system.

AI in performance management

Managing performance refers to systematically specifying the performance goals and subsequently examining these goals’ realization. Being the next largest subcategory after staffing, it mainly discussed employee performance prediction. 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 or positive or negative factors influencing the performance. A significant number of studies related to employee segments, namely, lawyers, bank employees, call center employees and academicians. Discussions on performance from an organizational perspective were meant to forecast how a group of HRM practices affects future organizational performance.

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. Studies included competency assessment based on visualization of the competency gaps or past data, either for evaluation or training. Research contributions also refer to the forecasting of future competencies and job profiles needed for organizational transformations, career matching based on skill preferences and measuring psychological capital for recruitment and appraisal.

AI in other HRM functions

The overall influence of AI application was a frequently studied aspect, which includes AI’s ability to improve decision-making, presenting AI as an essential approach, and the widespread influence AI has on HRM. Some peculiar subjects like exploring emotional involvement, developing an employee suggestion system, perception of service robots were also addressed. Specific discussions on applications were also attempted by some authors in this subcategory.

The AI – HRM concept map

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. Concept maps can be of various shapes, ranging from non-hierarchical to hierarchical structures and even data-driven maps where the input determines the map’s shape. 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. A concept map’s hierarchical levels, the number of relationships, cross-links and branchings are an estimate of cohesion or integration in the knowledge base.

AI application in HRM is made to facilitate or support HR decision-making. Based on HR decision-making framework, the map has been classified into two parts – namely, unstructured/semi-structured and structured.

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. Decisions, such as employee performance ranking by prediction, would be considered a very structured decision.

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.

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.

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. The advancement of ES in HRM can be seen in studied literature as early as 1994, where was 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. ES development in the HR domain helps solve unstructured HRM problems and contributes to developing complete human resource information systems (HRIS). Evidence of the development of a mathematical model for competence assessment, basic rule-based ES for selection called BOARDEX, using ES as a decision aid, and using ES for efficacious assignment and selection of the job seekers can be found in the literature.

Fuzzy logic in HRM

Fuzzy logic is based on the fuzzy set theory. 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. With these sets, fuzzy logic can quantify the data’s uncertainty and forecast future scenarios, which further facilitates decision-making.

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. In AI – HRM applications, it can be used for personnel selection and optimal staff design, differentiating between a suitable and nonsuitable job applicant, eliminating inaccuracy in evaluating the proportionate significance of traits and the performance gradings of the choices. 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.

Data mining in HRM

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. By implementing data mining methods, organizations can transform useful information and patterns to achieve a competitive advantage.

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 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. Data mining is one of the best ways to examine documents in databases as the outcomes can be utilized for prediction and future planning.

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. 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. Clustering analysis has also been broadly utilized in image processing, pattern recognition, etc.

Data mining has been used for employee selection, performance evaluation, competency evaluation, talent management, and various other HR functions.

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. ANN is a simplified model that has been developed to imitate the function of a brain. It is designed using a simple structure comprising of a processing element, layer, and network to reenact the human learning process. ANN is the most popular intelligent technique for prediction, which can help in solving models developed for predicting the HR functions like selection, recruitment and performance.

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

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. 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.

Genetic algorithm (GA) in HRM

GAs are search techniques that include discerning experimentation, seeking to determine a global optimal. They use strategies found in nature, for example, replication, mutation and gene crossover, to arrive at optimal solutions for mathematical problems. GAs have been relatively less used in HRM functions, with a few studies that implemented GA to solve the workforce planning problem and evaluate the performance. Also, GA can be used for problems related to the transfer of staff using real-life constraints, and for the formation of a framework to analyze the knowledge of the candidates.

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. Studies have revealed that the adoption of machine learning in decision-making is beneficial. Data mining is likewise a specific kind of machine learning (learning from examples). 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. It has been used for assessing the psychological capital of employees. Very few studies have used this technique for HR decision-making.

Ethical concerns around AI in HRM

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. 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 and can result in opportunities for a data-miner to take advantage of data subject’s vulnerabilities. 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.

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.

Toward a framework for ethical AI in HRM

Most organizations have values, policies and codes of ethics that aim to create an ethical culture. 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.

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; 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 it is considered that the quality of being self-learning is one central distinguishing characteristic of AI in organizations.

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.

Closing remarks on the state of AI – HRM domain

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. The sophistication of AI-controlled frameworks has lately expanded to such a degree that no human intercession is required for their structure and deployment. The impact of the sophistication can be seen from the employment of AI techniques in the recruitment process (such as shortlisting of CVs from career sites, direct and break down video interviews), predicting performance, and automation of tasks.

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.

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. 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, the need to integrate the DSS, ES and behavioral research; and the need to understand employees’ willingness to work with robots.

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 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.

Practical implications

While information technology and other advanced technological innovations have offered several benefits (cost savings, harmonization and integration of HR activities, efficiency, support of international strategy), they have also created extra barriers (more HR administration, work stress, disappointments with technological properties).

The consequences of technology largely depend on context. For some companies depending on the size, the industry and the country, technology enabled HRM will have a negligible effect. For others, it can be seen as a key factor for success and survival in a highly competitive market.

As for employees, the implications remain unclear. The number of jobs will increase, but the nature of these jobs will change. New skills will be required including problem-solving and communication that are particularly hard for computers to match. A decline in standard full-time employment and a growth in contingent forms of work are inevitable.

HRM transformations undoubtedly eliminate distance constraints, but the risk for an increasing lack of direct contact between the various stakeholders is lurking. Although there are several advantages, there is a danger lurking behind technology focused HRM and suggest that technology should be viewed as a decision support tool that enhances and does not replace the HR professionals in organizations.

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.

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.

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. 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.

Ethical and privacy issues can also be a significant challenge for HR managers. A broader view of the ethical issues and their management will help managers consider organizational and technological arrangements to manage these issues better. Organizations are the ultimate bearers of privacy, and ethical concerns, and 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.

AI, having the ability to perform human tasks and being able to think and feel like humans, will replace human labor entirely and, thus, human interactions will fade from sight. Consider, for example, the potential impact of virtual assis[1]tants like Siri. Dealing with queries and customer support internationally, they may enable organizations to operate 24 h a day, without engaging human employees as representatives at physical locations. Due to the dramatic advances in AI, automation and digitalization, unskilled workers in advanced economies may become unemployed, as human tasks and jobs are either offshored, cease to exist altogether or decline.

The progress of AI may change the fundamental nature of work and pose a serious threat to human employment. However, it can also create significant opportunities for human-machine collaboration and integration. AI can be of great value in facilitating service or sales and creating more favorable, customized, and valued service interactions.

Machine learning can assist in processing interaction-based knowledge, analyze variability across interactions and clarify ambiguous patterns using data from frontline employee (FLE) – customer interactions.?In this way, it gives FLEs the possibility to use this data for the provision of efficient, effective, and customized solutions to customers.

Similarly, artificially smart technologies, being capable of natural language processing and real time learning, play an important role in complementing human interactions and increasing problem-solving effectiveness. AI algorithms in the realm of journalism beyond the initial programming can also assist journalists in basic works, allowing them to focus on more investigative reporting, generating at the same time news faster, at a larger scale and with less errors.

In addition, people can use AI, often in the form of personal digital assistants, to facilitate work activities regardless of temporal and spatial location. Overall, these observations are consistent with the view that the effect of automation technology on staffing decisions greatly depends on a facility’s vertical position in the local marketplace, thereby supporting the argument that automate intelligent technologies do not lead necessarily to reduced job opportunities.

AI concerns, inter alia, information processing, logical reasoning, and mathematical skills. For employees, those challenging skills could be obtained through expertise and training.

AI applications could be of pivotal utility in HRM for training purposes. Simulations, defined as AI environments, can provide high degree of interactivity with other users, and enhance learning opportunities.

Despite the increased cost of using such technologies, simulation-based applications allow employees to interact and comprehend how to adapt their decisions to the interactive effects of the environment and multiple competitors.

The use of intelligent animated characters for training purposes, giving feedback, and providing support like a human trainer. These intelligent agents can learn in real time and amend their training to employees’ preferences and external information, addressing issues related to low engagement and isolation in web-based training.

AI computer agents have been examined as important tools in enhancing employees’ skills when interacting in strategic and negotiation settings, saving considerable effort, and offering better performance.

Early studies in considering AI as a decision-making tool in HRM suggest that expert systems – AI applications that embody the knowledge and decision-making abilities of a human expert – can increase the accuracy of HRM decisions made by non-experts and eliminate the time, required by them to make HRM decisions. Explanations produced by AI expert systems are useful to managers who are firstly assisted by this decision-making process and are able secondly to learn why a particular decision was made.

Other studies pertaining to AI applications in HRM decision-making highlight the ability to process large amounts of data at high speeds, the possibility to help salespeople to acquire new customers more efficiently and the potentiality to effectively evaluate and manage employee turnover risk. However, even when AI improves task performance and poses no immediate threat, its extensive use in HRM decision-making is likely to be perceived as a threat to human employees’ autonomy, status, and job security because it can provide more options to them, and confuse them, increasing perceived complexity.?

The introduction of AI applications in HRM allows HR employees to conduct background checks of job applications and develop compensation packages for certain positions. AI-enabled recruitment platforms can also extrapolate possible behaviors in terms of job fit and performance while being less biased and more objective than humans.

Machine learning can greatly assist HR practitioners and firms by transforming the selection process into a more systematic process by eliminating the occurrence of recruiters’ biases or even applicants’ influence methods to deviate the selection process. The numerous advantages that AI provides to HRM recruiting constitute a positive development for HRM.

However, these positive effects have been questioned in reference to ethicality of acquiring and progressing of data as well as in terms of favorability among applicants. AI machine learning and deep learning applications in HRM raise questions of privacy and offer a fruitful discussion of ethical challenges. Notably, direct applications in the employment and HR context through AI machine learning, including the analysis and collection of digital records to support traditional psychometric tests in evaluating talent and predicting work-related issues, entail several questions concerning human privacy.

Similar privacy issues arise when employing image and video recognition in digital interviews through AI deep learning to capture verbal and other interpersonal behaviors and amend them to create a psychological profile and predict possible fit.?

Overall, with the increasing involvement of AI in the HRM field, we are witnessing a shift from eHRM to a new phase. In this phase, AI intelligent automation constitutes the tool that drives the transformation of HRM by utilizing AI applications in recruiting, training, and decision-making.

Robotics

Research on robotic technologies has predicted that many jobs will soon disappear and be replaced by automation and robotics. The jobs that are more possible to suffer the greatest effects of job displacement are those of welding, painting, and assembling jobs as well as those employees that are less educated, experienced, and skilled. It is also plausible that humanoid robots like robot waiters in restaurants and virtual assistants that provide guidance to customers through a company’s website will fully substitute human frontline employees. Other studies suggest that the impact of robotics might be of great importance for HRM and more specifically for unemployment; however, this might occur in the future.

Given the way that AI, digitalization, and robotic technologies are being shaped by socioeconomic and organizational forces, predictions about mass joblessness and replacement by robotics are not likely to be realized.?Situations defined by strong needs for empathy, in which developing original and creative solutions is required or that necessitate high levels of social intelligence are not at high risk for automation and replacement.

Shifting away from job replacement, several researchers emphasize the need to combine human capabilities with robotic technologies in HRM to bring more insightful HR solutions. In this regard, more skilled and educated employees are needed in the era of human and robots’ symbiosis and collaboration to be benefited from possible opportunities and reverse potential threats.

Robotics can support human employees by offering them opportunities for more technical positions that are either created or enhanced by robotic technologies. Robotic surgery is a notable example. Although robotics can enhance precision and reduce errors if applied correctly, the human knowledge remains a vital component. Importantly, the features of technology as well as the manipulations and knowledge of the doctor are required.

Robotic technologies have also brought several learning opportunities to HRM. Work on robotics emphasizes the ways in which robotic technologies can eliminate repetitive and routine activities handled by human employees, offering to them the possibility to engage in opportunities to use their skills more effectively. At the same time, this creates new learning opportunities combined with extensive training for the employees to meet their altered responsibilities and acquire the skills required to work with a robot.

However, employees may exhibit differential perceptions towards robots based on their occupations. High-skilled employees have more positive attitudes towards robots and their implementation as they offer them opportunities to expand further their skills and knowledge. Inevitably, jobs designed based on AI and robotic technologies are about to bring uncertainty. Yet, these technologies offer the opportunity for the design of problem-solving strategies that will be of great value.

Managerial implications

Firms should establish an organizational environment in which human employees and technology could coexist. Firms should focus on training and ongoing development of employees for them to meet the criteria and skills needed for working with AI agents.

It is the managers’ responsibility to assist employees in being more engaged in such activities that will offer them the technological knowledge required in the competitive international market. The technological knowledge required could be acquired through numerous flexible alliances with various government and public organizations, research centers and universities.

Notably, although there are several advantages, there is a danger lurking behind technology focused HRM, suggesting that technology should be viewed as a supporting tool that enhances and does not replace the HRM professionals in organizations.

Artificial Intelligence can boost the positive effects for HRM, if managers refrain from letting technology dominate and substitute the core meaning and role of HRM.

HR practitioners should focus on the mutual development of HRM strengths and intelligent technologies. Technological developments in HRM, including the introduction of AI, machine learning and deep learning applications for the analysis and collection of digital records in predicting work-related issues, have raised several concerns pertaining to human privacy. Considering the privacy and ethical challenges that these positions hold, there is an emerging necessity for the development of regulations that guarantee the rights of employees or potential employees for the protection of their data.

Although considerable progress has been made with the General Data Protection Regulation (EU, 2016/679), the rapid technological developments imply ongoing updates that will raise the awareness of society and employees.

Finally, in the context of global HRM, thinking of how AI technologies eliminate distance constraints but at the same time how they minimize the direct contact between the various stakeholders involving digital mediations, managers need to consider ways to use these technologies for the benefit of firms and employees. This entails assisting diverse actors to use different technologies to coherently perform shared work arrangements.