Artificial Intelligence and Pay Information Disclosure: Changing How Pay is Communicated

Abstract

Pay information disclosure (PID), or communicating pay information between and among actors, affects employees, organizations, and societies. Disruptions resulting from artificial intelligence (AI) will also change how pay is communicated. Based on AI, AI and human resources (HR), and PID, as well as anecdotal data involving organizations that are integrating AI in their pay practices, we introduce areas of AI relevant to PID and describe opportunities and challenges. HR should play a critical role in developing employee trust in AI systems by protecting employee privacy, training AI on high-quality data, and ensuring AI algorithms are ethical. AI can transform PID by supporting advanced pay methodologies, reducing barriers to sharing information, and educating employees. However, research needs to be conducted on all of these areas and practitioners should strive to keep an open, but cautious mind about both the benefits and challenges of integrating AI into HR practices.

Keywords

pay communication pay information disclosure pay secrecy pay transparency pay outcomes

Introduction

Forces are converging to disrupt the amount and manner that companies engage in pay information disclosure (PID), which is “the communication of relevant pay information between and among actors” (Brown et al., 2022, p. 1662). Changes in technology, which make it easier to post and find pay information, changes in legislation, which are requiring greater pay disclosure around the world, and changes in employee demands are increasing pressure on organizations to share pay information with employees and stakeholders. Pay information helps employees decide what organization to join, whether to stay, and how much effort to exert (Gerhart & Rynes, 2003); hence, PID is garnering increasing attention among employees, employers, governments, popular press, and academics (Arnold & Fulmer, 2019; Bamberger, 2021, 2023).

Research in the PID field uses terms such as pay secrecy (Belogolovsky & Bamberger, 2014; Blumkin & Lagziel, 2019) or pay transparency (Brandes & Darai, 2017; Cullen & Pakzad-Hurson, 2019). Others refer to a pay communication scale with pay secrecy and pay openness (Marasi et al., 2018) or pay transparency representing the two anchors. However, using pay secrecy and pay transparency can create artificial biases because the terms are value-laden with pay secrecy conjuring back-room deals and pay transparency invoking images of open, engaged leaders (Bernstein, 2017; Costas & Grey, 2014). In contrast, using the term, pay information disclosure (PID), removes the value judgment, and focuses on the provision of information and the reduction of information asymmetry (Brown et al., 2022).

Specifically, the PID perspective fosters an understanding that pay information is multidimensional, including instances in which information about one’s own pay (e.g., merit increase) versus the pay of (all) other employees (e.g., in a department or group) is made available, and in which partial (e.g., the median pay of a group) versus full information (e.g., the pay of all individual group members) is disclosed. Communications can be initiated by the company, informal exchanges among employees, as well as through intermediaries. Typically, employers have more and better pay information than employees, giving employers a decision advantage, but PID can help narrow the gap (Brown et al., 2022).

While PID is increasing in importance, artificial intelligence (AI) is capturing the world’s attention as a transformative technology that changes the way companies work and communicate. AI is transforming the workplace, from automating repetitive tasks to improving decision-making and such automation is likely to influence PID. In this manuscript, we explore the intersection of AI and PID. Our goals are to draw attention to the ways that AI can play a role in PID and identify key challenges and opportunities of implementing AI for PID.

Because work at the AI-PID intersection is in its infancy, we use a generative approach in issue identification (Aggarwal et al., 2021). First, we briefly explain PID before we introduce AI and the AI domains most relevant to PID. Second, we examine challenges for AI to evolve to the next level in PID. Third, we explore opportunities to capitalize on the benefits that AI brings to PID. By examining the AI-PID intersection, we suggest ways that organizations can leverage technology to create more transparent, equitable, and effective pay systems. We also provide insight into how AI can help reduce pay information asymmetry, while noting that new forms of asymmetry may emerge. Our insights are drawn from research on AI, AI and Human Resources (HR), and PID.

Introducing PID

PID researchers generally focus on four areas: what pay information is communicated, what the process is of making pay decisions, how pay information is disseminated, and to what extent pay communications among employees are encouraged or discouraged (Fulmer & Arnold, 2020). Even though increasing PID may increase pay system effectiveness, substantive barriers remain, such as organizational fears of revealing secrets to competitors or exposing existing inequities, societal values of individual privacy, and the difficulty of communicating complex topics and disseminating large amounts of data (Brown et al., 2022). To minimize reader fatigue, we point interested readers to Brown et al. (2022) for a complete review of the PID literature.

AI Domains Most Relevant to PID

AI Functionality

AI is the field of computer science that focuses on creating machines that can perform tasks that previously required human intelligence, such as reasoning, learning, and problem-solving. AI is a broad term encompassing subfields, including natural language processing (NLP), computer vision, robotics, and machine learning (ML; Howard, 2019; Wang & Siau, 2019). In HR, AI can automate repetitive tasks, such as resume screening and candidate sourcing, freeing time to focus on more strategic initiatives (Strohmeier & Piazza, 2015). AI can also analyze employee data and identify patterns that can help organizations make decisions to improve employee engagement and retention (Charlwood, 2021; Guenole & Feinzig, 2018).

Although AI is talked about as one tool, it is many different tools that perform different functions. These span a spectrum from knowledge-related tools that discover, represent, and process knowledge, to solution searching tools that mimic problem-solving cognition, to language-related tools including text and speech processing. Each has a specific purpose with informational inputs and outputs. Tools can then be linked to make networks of tools or platforms in support of an organization or across organizations (Strohmeier & Piazza, 2015).

Understanding each tool’s purpose informs whether the tool will have the correct task-technology fit, meet the desired objectives, and provide a sufficient return on investment (ROI). Task-technology fit is where there is high correspondence between the task requirements and the technology functionality, and the major success criterion includes improved efficiency, effectiveness, and quality (Goodhue & Thompson, 1995). Although AI technology is advancing rapidly it can be expensive, especially for niche applications, and therefore, the ROI of AI designed and trained for specific tasks is an important consideration (Furneaux, 2012).

Data

AI tools require extensive data inputs to learn patterns and to put context around data to create information that enables artificially intelligent decisions (Kim & Bodie, 2021). For instance, it takes enormous samples of human communications for NLP. NLP is a subset of AI that focuses on enabling machines to understand, interpret, and generate human language. Some common NLP applications include text analysis, sentiment analysis, speech recognition, translation, and chatbots (Howard, 2019; Wang & Siau, 2019). As AI expands into new areas, such as performance management, it will need to capture new data, and this increase in data collection will be less expensive and more comprehensive (Kim & Bodie, 2021), but can come at a cost to employee privacy. For instance, an amazon warehouse performance management system, through the smart device used to scan picked items, tracks the location, activity status, and productivity of workers, meaning that it is collecting substantial data about an employee’s every movement (Delfanti & Frey, 2021).

Algorithms and decision-making

AI processes data through algorithms producing insights that are expected to help in decision-making that normally requires human cognition, including adaptive decision-making (Tambe et al., 2019). All companies will need to develop strategies to determine when decision-making is fully delegated to AI or aggregated together, and where AI and humans make decisions in parallel (Shrestha et al., 2019). In one example, IBM created a machine learning-based system to be used during pay discussions to give pay recommendations; however, the manager had the final authority to accept the recommendation. It had a 95% acceptance rate and it also resulted in 50% attrition reduction (Sammer, 2019). In another multi-employer example, hiring managers used an algorithmic selection test to assist in selection decisions. Managers who overruled the test recommendations ended up with worse average hires, demonstrating the algorithm’s decision-making support potential (Hoffman et al., 2018).

AI and PID Interactions

Three types of AI tools (i.e., AI-enabled automation, decision support systems, emotional intelligence systems; Bilquise et al., 2022; Jarrahi, 2018; Strohmeier & Piazza, 2015), can affect PID-related outcomes, including pay customization and pay education, by transferring tasks from humans to machines (Zuboff, 1985). For example, AI can automate payroll processing, provide employee pay and benefit information, and answer questions through chatbots.

AI-enabled automation systems currently most relevant to PID are systems that gather data and interface with humans. Data gathering systems range from database systems to web crawling bots, to data gathering electronic surveillance systems (O’Leary, 2013; West, 2019). These systems will increasingly provide more pay information in more sophisticated and customized formats, such as interfacing with humans through chatbots that use NLP in text or verbal form (Luo et al., 2022; Rapp et al., 2021).

Decision support systems have some combination of AI and human interaction to improve decision-making (Jarrahi, 2018). AI can feed information to humans, humans can feed information to AI, or they can work together to make decisions. Part of the decision-making process is the creation of new knowledge from information (Strohmeier & Piazza, 2015). Even with advanced modeling there remains unpredictability (Jarrahi, 2018). For example, economic modeling, which forecasts supply and demand, can assist PID by providing “what if scenarios,” such as considering labor supply and demand. While data scientists are updating AI algorithms (Marwala, 2013), AI will be dependent on data quality, data availability, and historically relevant data to be informative (Florens et al., 2007).

Emotional intelligence systems are AI tools designed to recognize, interpret, and respond to human emotions (Bilquise et al., 2022). Through several means including NLP, sentiment analysis, interpretation of facial expression, tone of voice, or body language, AI systems are increasingly able to interpret and adapt to delivering communications that best fits emotional states (Fu et al., 2017; Gil et al., 2015; Yadegaridehkordi et al., 2019). As AI is increasingly able to effectively deal with human emotions it can help with both excitement and sorrow (Hughes et al., 2019). Thus, the use of AI can extend beyond the delivery of pay information itself to assisting in the response to the emotional reaction of an employee receiving the pay information.

Pay Customization

One of the more profound ways AI can transform PID involves the ability for organizations to use AI to customize employee rewards more precisely, moving, albeit slowly, away from a one-size-fits-all approach to employee rewards (Conroy et al., 2015). Pay customization will have increasing implications for PID because AI-driven pay customization both enables increased PID and requires increased PID for success. As AI supports increasing complexity and customization, the foundations for how pay decisions are determined could be made clearer to employees. Such pay customization, which will be facilitated by increasing automation, can lead to greater PID clarity (Abdulsalam, et al., 2021). At the same time, a highly customized pay system without increased PID will be too opaque for employees to understand.

There are four primary issues in this domain.

AI will drive pay administration efficiency while also allowing increased complexity

AI and automation are making pay administration more efficient, meaning that it can also make it easier to engage in PID. For instance, budgeting, benchmarking, salary administration, and pay structure determination can be automated while becoming substantially more sophisticated. Historically, pay administration and design has been constrained by the resources required to administer a pay plan and the difficulty it takes to explain the plan to employees. The more difficult the plan, the more likely employees will misinterpret the plan leading to lower pay satisfaction (Gerhart, 2023). Additionally, pay methods are path dependent on prior pay philosophies, which emerged before the advent of advanced computer technologies (Nyberg & Reilly, 2019). AI allows for a rebirth in pay philosophy. With the help of AI, complex pay designs can be created with numerous variables while simultaneously facilitating PID.

AI complexity can mean increased employee confidence and trust

Trust in a pay system comes from transparency and predictability of its outcomes and an understanding of how the system works (Day, 2007, 2011). In an AI-based pay system, although the system may be more complex, the system can still be more transparent, predictable, and understandable because the AI system can explain its data inputs, algorithms, outputs to help educate employees. Employees can also ask as many questions as needed to understand. Hence, with chatbot use or other automated systems, employees could have personalized training and answers that cannot occur at scale when there is a need for human explanations, and a better understanding of how their pay is related to their own efforts will likely improve their likelihood of performing those tasks, if the pay system is designed well (Alterman, et al., 2021; Nyberg et al., 2016).

AI increases the possibility of pay customization

AI enables opportunities to shift from the traditional single pay plan for most employees, at least most employees doing similar jobs, toward pay customization such as idiosyncratic deals (i-deals), individualized pay and performance arrangements (Maltarich et al., 2017), or mass customization. As with the marketing field, where mass customization has long been recognized as a leading benefit for customers (Da Silveira et al., 2001), so too is it possible that employees would prefer to have their rewards customized. Customizing rewards can lead to greater employee performance and potentially greater employee rewards (Maltarich et al., 2017) and increased organizational performance (Abdulsalam et al., 2021). Pay customization enables companies to consider and combine multiple factors such as job performance, experience, education, skills, and employee preferences when determining pay. This approach avoids relying on a few limited factors (e.g., job level, tenure) and combining them in limited ways to apply to all employees the same way. Concurrently, it allows for pay to be calculated in more complex, holistic ways—allowing the full rewards to be considered. With such mass customization through AI, it may be possible for many more employees to have customized pay, reducing concerns about special treatment. However, as individualization complicates pay comparisons, customization attempts will increase the importance of PID in terms of trust, and fairness perceptions toward the overall pay system. It is also likely that not all types of occupations and jobs will benefit equally from AI. For example, there is little variation possible at the lower pay margin, and it is unlikely that much experimentation and adaptation will be applied to low-wage jobs (Conroy & Morton, 2023).

AI can reduce pay gaps

Gender and race pay gaps can be more accurately, and regularly analyzed using AI. This increased scrutiny and clarity of gaps across subsets of employees (e.g., sex, race, location, manager) can help identify where unexplained differences occur. Reducing pay gaps can improve PID outcomes as the processes underlying pay differences would be better explained. AI’s creativity, analytical power, and communication ability opens unforeseen possibilities compared to current methods that are limited by human and financial resources.

Pay Education

At its broadest level, PID involves educating employees about pay systems and helping employees understand how their rewards are tied to company goals with the intent of improving the ROI of the pay system (Brown et al., 2022). As AI becomes more powerful pay communications can become more personalized and will move from sending pay information to becoming educational, where the employee is interacting with the AI to learn about how pay works within the larger pay system. HR can then focus more on designing the system rather than conducting the transaction. Given how complex pay systems can be (Burkert et al., 2023), AI can play a substantial role in making pay education more effective and personalized.

There are several ways that AI can facilitate pay education. AI can help create personalized learning paths for employees based on their role, experience, learning preferences, and learning capability (Charlwood & Guenole, 2022; Del Prado, 2015). This could involve interactive modules that teach employees about pay structures, pay plans, and market trends. Gamification techniques can be used to make pay education more engaging and interactive. Emotionally intelligent AI can help employees process and understand their pay. This could involve creating chatbots that can answer employee questions about their pay or using NLP to analyze employee feedback and identify areas where pay education can be improved. AI-powered feedback systems could provide employees with feedback on how their performance affects their pay, which will further educate employees on how the pay system works and how their work contributes to the success of the organization.

Artificial Intelligence and Pay Information Disclosure Challenges

There are notable challenges involving the interaction between AI and PID. For instance, many companies are still struggling with basic HR data requirements as they move to HR analytics capabilities before they move to machine learning and continue to AI (Strohmeier & Piazza, 2015; Tambe et al., 2019).

Data Collection Challenges

It is expensive and difficult to gather data, train the system, test the decisions, and learn how to manage employee reactions. In addition to investing financial resources, there are several means for addressing challenges. Industry collaboration can share data gathering and management costs. Building and using synthetic data to train AI systems can save costs and prevent bias from entering the algorithm (Candemir et al., 2021; James et al., 2021). A related issue is that data science techniques perform poorly when predicting relatively rare outcomes because datasets used to train models are biased toward common occurrences, resulting in limited information to predict rare outcomes, such as are rare diseases or extreme weather (Dwivedi et al., 2021; Tambe et al., 2019). In HR, there may not be enough data to train AI to predict outcomes for highly specialized skills, small groups of employees, stars, or rare workplace situations. This means that many organizations may not be able to provide AI with enough internal data to be able to use tools effectively. It may be most effective if organizations work together to pool data, at least in some instances.

Organizations must also convince employees that the pay information generated by AI is useful and trustworthy. Failing to get employee buy-in can mean that the AI recommendations go unheeded. For instance, IBM used software to drive career advancement by recommending career moves based on employee interests, prior jobs, training, and the characteristics of those previously successful in the job. In 2018, the system’s recommendations were accepted only 27% of the time (Tambe et al., 2019), which is both substantial and limited. On the one hand, individuals are well advised to carefully weigh the potential benefits and challenges of a career move, which explains the high rejection rate. On the other hand, opportunities for career advancement may be rare, making the acceptance rate appear low. Evidence comparing the acceptance rates of traditional career paths with AI-suggested moves at IBM, as well as AI-based suggestions across multiple companies, would be helpful but is not yet available.

Implementing data privacy and security measures to gain employee trust and protect sensitive data will reduce employee resistance to participating in AI data gathering, training, and evolution efforts (Privacy Forum, 2018). Encouraging and rewarding employee participation in AI will also reduce employee resistance (Hughes et al., 2019). Finally, choosing AI initiatives that benefit employees will lower resistance to AI and create momentum and trust (Charlwood & Guenole, 2022). However, research shows that involving employees in trial periods can lead to technostress that risks employee engagement with AI in the long run (Maier et al., 2022). These risks musts be managed to lift AI’s role in PID to its full potential.

Measurement Challenges

Employee performance, which ought to be a key pay driver and PID input, is often difficult to measure. Performance evaluations are complex, involving factors such as job interdependence, group dynamics, human biases, and data quality. Individual performance is hard to disentangle from group performance because most complex jobs are interdependent with other jobs (Tambe et al., 2019). Given the uncertain quality of current human-based performance evaluation systems, using current systems as data, training, or architecture for AI systems may scale unreliable human decisions, making PID much more difficult.

Asymmetry Challenges

There are potential new asymmetry challenges for the future of AI and PID. Management could use AI exclusively and not release its functionality to employees. For instance, there could be strong PID on basic pay (e.g., pay level), but limited on other pay dimensions (e.g., overtime, one-time bonuses, incentives, targets, payouts). These challenges may be further complicated at different levels (e.g., collective pay) when the line of sight between pay and outcomes is already a little more distorted (Nyberg et al., 2018). Alternatively, PID could be broad and deep, but analytical tools kept only accessible to management. For example, employees may be able to see current pay reports, but only management may be able to see the AI future pay predictions. Thus, new PID questions arise, such as: should employees be made aware of all AI capabilities around pay? Should employees be made aware of their future pay projections at the company?

AI requires resources and there will be haves and have nots in the AI space (Brougham & Haar, 2018; Xiang, 2022). Gaps between companies, within companies, between employee tiers, and between employee types will arise from differences in AI investment. Investing in expensive AI for all employees may not be feasible. Further, there may not be enough data to implement effective AI for all employee groups. When resources are lacking, it may require collaborating with other companies with existing AI technology to implement pay-related systems.

As AI becomes more powerful, outside influences could increasingly influence within company pay dynamics. Competitors could actively solicit pay information from employees to train their own AI models or develop recruiting plans. It may be possible to gather pay information to be more competitive or to try to spread false pay information which would affect pay satisfaction. At some point, AI programs and AI networks may also start competing for pay information dominance (Aggarwal et al., 2021; Qiu et al., 2019).

Black Box

How pay decisions are made is often described as a black box, due to the inability to see or understand how pay is determined, particularly from the employee perspective, due to low PID (Brooks, 2021). Likewise, AI decisions are also described as occurring inside a black box due to computer algorithm opacity (Kim & Bodie, 2021). While as noted earlier, AI could create more clarity and consistency in pay determination, it may still create PID difficulties. With AI, even if the algorithms themselves are transparent, including the code used, few may understand and comprehend how codes translate into decisions that affect pay, potentially increasing PID challenges and increasing the black box effect.

Another challenge is that AI processing and algorithms are often poorly understood by those making decisions (Pasquale, 2015). This raises concerns about the ability to access and inspect the data, logic, and decisions of an AI system. It also raises concerns about the ability to anticipate and control the AI outcomes and impact. The lack of understanding about the core of AI recommendations can also lead to unintended (and even unrecognized) biases and errors or distortions, making PID more challenging, but also more important.

Privacy Tensions

One PID challenge is that both employees and companies often view pay to be private (Brown et al., 2022). The necessity of data to fuel AI systems and the inclination to disseminate pay information derived from AI analyses may intensify privacy concerns (Kim & Bodie, 2021). For example, pay data to be distributed through AI could be based on an individual, every employee in the company, and potentially even non-employees. The data used in the model, and subsequently in pay communications, could become increasingly intricate. Pay communications might initially involve basic pay level and PID but could eventually encompass all aspects of pay and performance, including the antecedents of both (such as performance metrics influencing pay) and the projected consequences of pay changes. Internally, companies may gather data that is increasingly invasive but related to performance (e.g., health metrics). Externally, companies may gather information about competitors. Consequently, future PID data could cover a much wider spectrum of pay-related information. Therefore, the data required to power AI systems and the data managed and reported by AI are likely to form a complex web.

Informed Consent

The privacy-related issues of AI-PID data requirements raise informed consent concerns (Kim & Bodie, 2021). Informed consent is the principle that data owners should be notified about what data will be collected and how it will be used before collection (Giannopoulou, 2020). One challenge of today’s data ecosystem is that the assumption is once consent is given it is always given (Froomkin, 2019; Leetaru, 2019). The ethic that consent is an ongoing process (Gupta, 2013; Klykken, 2022) as applied to personal data implies that individuals have the right to assert affirmative consent when their data is going to be shared continuously or that they have the right to revoke consent. In today’s Big Data environment, many people may not know what consent was given or when. The ongoing use of private data is often justified by the argument that the data is anonymized, aggregated, or manipulated in a way that it is not traceable back to its original source (Froomkin, 2019; Giannopoulou, 2020).

Electronic Surveillance

There may also be increasing opportunities to collect data through observing an individual through electronic surveillance (West, 2019). For example, anonymous data such as log on attempts may be gathered (or purchased) and then later tied to the individual and used as data (Krapp, 2022; O’Leary, 2013). Electronic surveillance companies gather data through internet browsers, websites, smart phone, and phones apps with no or limited users’ knowledge (O’Leary, 2013; West, 2019). Another tool is triangulation. In triangulation, a company predicts who an individual is from a data pattern. This is like profiling. A profile can be built around one piece of data (e.g., a phone number) without knowing who someone is specifically and through location tracking identify their work and home locations (Krapp, 2022).

In the PID context, AI systems could, through surveillance methods and through third parties, figure out a vast array of job performance and pay-related information including work schedule, job title, position, work location, and pay level all without asking permission. Corporations could use internal surveilled data to supplement performance measures (e.g., Amazon warehouse workers; Delfanti & Frey, 2021). External third parties could gather and provide surveilled data to supplement data that could feed algorithms to estimate individual pay level or predict any number of pay-related factors to sell as competitive information.

Regulatory Challenges

Depending on the legal and institutional context, employers have broad legal rights to surveil their employees and use their data (Kim & Bodie, 2021), yet much of the legal framework around how AI uses data is unclear. HR-related legal protections regarding employee data are emerging; for example, the United States Health Insurance Portability and Accountability Act protects medical data (Kim & Bodie, 2021), but legal mechanisms to litigate AI’s data use are not robust. For example, in the United States and Europe, AI’s use of employee data have not been clearly protected (Scassa, 2018; Yavorsky, 2022). For example, in the United States, part of the challenge is that most AI collection and data use is seen as a copyright issue, whereas algorithms are seen as patent issues. Additionally, copyright law does not protect the output of AI because it does not protect the output of non-humans (Ernst et al., 2019).

There may be many paradoxes in the future about AI’s challenges and benefits, but HR can affect the future (Charlwood & Guenole, 2022). HR should influence legislation regarding AI and PID (e.g., SIOP’s 2023 recommendations for AI-based assessments; Stark, 2023). As AI tools are deployed across domains ranging from banking to employment to criminal justice systems, regulators are paying particular attention to AI’s potential ill-effects on fundamental rights and taking actions (Rodrigues, 2020). At the same time, pay transparency laws are becoming more popular. For instance, some states have enacted pay transparency laws that require employers to list salaries and benefits on job ads or provide pay scales upon request (California Senate Bill 1162, 2023; Rosenquist & Visconti, 2023). HR professionals should be thinking about how AI and transparency laws overlap, and more broadly about PID and AI.

The ethics and oversight of AI is rapidly evolving. More than just legislation, guidelines, and decision-making protocols are needed; rather, there needs to be inspection of the code and of AI’s output (Bostrom & Yudkowsky, 2018; Dwivedi et al., 2021). Well thought-through, easily identified, safeguards should be placed in the code because it cannot be assumed that regulations or protocols will trickle down into AI protocols unless they are specifically written into the code. When safeguards are written into the code, they are testable and auditable and therefore create a closed loop quality process to ensure compliance (Privacy Forum, 2018).

Part of the challenge of dealing with AI policy making is that the regulatory and oversight domains creating a confusing patchwork of potential regulatory possibilities (Kim & Bodie, 2021). Laws, government branches, agencies, companies, industry coalitions, will all influence AI standards and oversight, and enforcement. With AI, there must be visibility and management at the data input, algorithm, code syntax, and output levels (Kim & Bodie, 2021), but there is substantive skepticism about government’s role in regulating AI (Rampersad, 2020).

AI and PID Opportunities

Having identified the primary domains in which AI may be relevant to PID as well as the challenges of implementing AI in PID, we now focus on how HR and PID can add value to organizations, and how AI can add value to HR especially in PID.

Investing in AI

HR must carefully decide where it is going to invest in AI. Determining ROI requires identifying the specific use for the AI tool and evaluating the task-technology fit (Jarrahi, 2018). The foundational categories of where to invest in AI resources include pay administration and planning (forecasting, scenario analysis, allocation, analytics, transparency reporting, equity analysis), pay learning and education (personalized pay explanations), performance management (goal setting, tracking, evaluation), data gathering and sharing (open source platforms and cross company collaboration), and emotional intelligence systems (individual customization, sentiment analysis, emotionally intelligent information delivery), among others. For instance, some PID-related AI may deliver routine pay information that has little consequence. However, some AI may be involved in managerial pay-related decisions or real time performance feedback that can have long term financial and emotional impact on employees. On the one hand, pay for performance can cause stress and mental health problems and has been shown to increase the usage of antidepressant and antianxiety medication (Dahl & Pierce, 2020). Through increased PID and the use of emotional intelligence, AI might help reduce such concerns. On the other hand, substantial investment in training will be needed to make sure the system works optimally before implementation because if it fails, it can cause considerable pay dissatisfaction that can lead to turnover or personal distress (Nyberg, 2010; Williams et al., 2006).

Building AI Competency

To be effective, HR professionals need to understand how to apply AI to HR broadly and PID specifically. HR must have a plan for upgrading their AI competencies and how to communicate effectively with employees, managers, and external stakeholders about how AI will affect PID issues. HR can partner with educational institutions, experts, or consultants who can provide guidance on integrating AI tools into HR processes and workflows. HR can also create opportunities for employees to collaborate on AI and PID-related projects and initiatives to gain hands-on experience working with AI tools. HR can also create an evidence-based and science-infused AI knowledge base that includes articles, case studies, and best practices that affect how AI can influence PID to build AI competency.

Reducing PID Barriers

Barriers to sharing pay information will continue to shrink. AI will enable individuals to seek and share information more easily and then make it available to others. It is not just AI’s ability to gather and process vast amounts of data, but an individual’s ability to ask for pay information through common language interfaces that makes AI powerful. It will become possible to ask a chatbot, “how much do I make compared to my coworkers?” and AI will be able provide the pay information and present it in understandable formats. Then the individual can ask any number of follow up questions. Currently, to share pay information outside of the company, it takes effort to find an internet site and then provide enough data to establish oneself and then input the data. In the future, bots may come directly to the individual, or may gather the pay information simply by asking, “do you want to share your pay information? Y/N” and the bot will go through the system and gather all the detail needed to upload it into the database.

A particularly difficult part of collaborating about pay across companies is job matching due to job variety and specificity (Weller et al., 2019). For example, jobs titled Programmer and Developer could be the same or different jobs depending on the company. Likewise, Manager could be the same or reflect different levels of responsibility at different companies. AI will make this easier once the system learns about different terminology in context and combines it with other data. Third-party AI may invasively or cooperatively gather many levels of individual data related to pay beyond simple pay level, including attitudes and intentions.

PID as Competitive Advantage

AI’s inherent properties align with and reinforce PID which can become an integral part of a pay system’s sustained competitive advantage as PID is linked to satisfaction, fairness, motivation, and performance (Brown et al., 2022; Fulmer et al., 2023). AI’s ability to improve productivity will continue to push their use, which requires increased data, and improved algorithms. Increased AI usage, to be done ethically, also requires increased oversight into algorithms and data. Therefore, the processes, which will be handling pay data and pay decisions, can be more transparent and auditable. As these systems are more documented and transparent pay information can be more accessible to employees (Zenger, 2017). Further, one reason management resists increasing PID levels is when current systems have inequities that will be exposed (Brown et al., 2022). Thus, the implementation of AI-related tools may not only hasten the implementation of more fair and auditable pay systems but could also address resistance to PID policies by addressing hidden inequity concerns.

Pay Negotiations

AI has the potential to affect how pay negotiations are conducted between employees and companies. With AI providing sophisticated support to create and explain complex algorithms to calculate pay, through increasing PID, the method for deriving pay can become more objective. Firms will be better able to explain and justify their positions (i.e., increasing PID) to employees with well-developed AI. At the same time, employees will be able to see errors in the employer’s methods that are unique to the employee or find additional information that is relevant to the calculation that previously may have been inaccessible. External labor marketplaces will also become more transparent. Therefore, employees will be better able to see opportunities at competitors, thus equalizing bargaining power from an information asymmetry standpoint.

Because of the dominance of the AI algorithm on pay calculations and processes, what may become pay negotiation’s larger focus is the data being fed into the AI systems. These future negotiations may be more focused on what rules are applied to circumstances rather than focusing on the dollar amount. While the dollar amount is the result, the focus becomes the algorithm’s rules and decision points because much decision-making may be within an agreed upon AI framework and decision protocol. Firms may lose bargaining power if the cooperation needed from employees to participate in AI systems requires substantive concessions. For instance, employers may need to give Unions and employees access and input into AI design to get their buy-in and participation in providing their data and help training the AI system.

Trust in AI Through PID

Using AI and PID can incrementally increase employee trust in the pay system if it removes human bias and employees believe that it is removing such bias (Kim & Bodie, 2021; Miller, 2015). This is because AI algorithms are designed to analyze data objectively and make decisions based on the information collected. As a result, fairness perceptions in the decision-making process could increase (Hughes et al., 2019), but this will depend on ensuring that the systems themselves are not perpetuating biases based on the data being collected by the system. Therefore, it is possible that one concern expressed about PID (i.e., trust) will be reduced through AI. This could occur if AI calculates employee pay and provides enough information about the pay system that, if the system is good, employees may have more trust in the system.

Imagine two systems. The first is a high PID system where the rules are open, auditable, governed, and explained by an AI system in such a way that it gains the confidence of those participating in the system. The second is a system that provides simulated or masked data that protects individual privacy but still represents the company in a way that illustrates where someone fits within the pay structure and why. Either way there is enough information and explanation of the system to gain employee confidence about how the system is working.

In the future, effective PID may depend more on emotional and contextual factors than technical ones. AI has the potential to solve the challenge of explaining how pay is calculated and affected by market conditions. In doing so, the interpretation of pay and responding to emotional reactions to pay may become easier. Customized and emotionally intelligent AI could play a role in helping to motivate employees through effective goal setting, clarifying line of sight, personalization, and through emotionally intelligent interactions.

Our primary takeaways are summarized in Table 1.

Table 1.

AI-PID Intersection.

Category	Label	Description
Domains	AI functionality	Data, algorithms, and decision-making.
	AI-PID interactions	AI-enabled automation, decision support systems, and emotional intelligence systems.
	Pay customization	Increased pay administration efficiency enables complexity while increasing employee trust, through means of mass customization.
	Pay education	Personalized, emotionally supportive, pay communications, and interactive learning paths increase employee trust in pay systems.
Challenges	Data collection	Large amounts of data are needed for AI effectiveness and to create employee trust in AI.
	Measurement	Some data is difficult to gather and evaluate.
	Asymmetry	AI resource and ROI requirements inhibit equal investment, distribution, and access, which can perpetuate inequalities.
	Black box	AI perpetuates pay’s black box effect if not addressed.
	Privacy tensions	Increased demand for sensitive data. Privacy tensions may increase.
	Regulatory	Opportunities exist, but knowing those that should be taken advantage of creates substantial questions.
Opportunities	Investing in AI	Investment in AI can have sufficient ROI in multiple areas. For example, AI can both increase pay administration efficiency and increase pay system sophistication.
	Building AI competency	An early competency in AI helps companies gain early mover advantage and a sustained competitive edge.
	Reducing PID barriers	Barriers to exchange pay information will shrink, making it easier for employees and companies to seek and share pay information. Job matching across companies will improve.
	PID as competitive advantage	Companies that recognize the effects of AI and PID can turn it into a sustained competitive advantage.
	Pay negotiations	Increased information can change pay negotiations to focus on procedures and algorithms rather than on pay levels, which is often the current focus.
	Trust in AI through PID	Employees can gain greater trust in AI-enabled pay systems than human managed ones when they understand them. AI-enabled systems have potential for overall better PID and PID outcomes.

Future Research Suggestions

Substantial research on AI and PID is needed across all areas of HR because to date the very limited research about the intersection of HR and AI has been very narrowly focused. For example, some areas that desperately need research attention, including pay customization research, should explore the range of factors that can be considered and under what circumstances multiple factors can be considered while maintaining a sense of benefit to employers and employees. How employees trust AI should also be examined because it is unclear if PID will be easier or more difficult for employees to accept when it is enacted through AI. Research should also be extended and applied directly to compensation-related decisions to learn how employee trust is affected in the compensation decision context, which has its own unique set of variables. AI will offer an increased ability to inform and educate, but are there excessive levels of PID (i.e., too much information) or diminishing returns on pay education? Because of AI’s costs, knowing the ROI threshold is valuable, and it is easy to forget that AI must be trained before it can be functional. Statistical modeling should also be conducted on different performance management measures and methods to identify sample size thresholds required for accuracy in AI decisions.

Summary

HR is at the heart of how AI will impact organizations. A key concern about AI involves the elimination of jobs (inherently an HR function). Simultaneously, organizations grapple with identifying and securing top-quality talent to propel their progress forward (again, an HR responsibility). Moreover, ethical questions surrounding AI and its implications for privacy are fundamental concerns that directly affect employees (again, an HR responsibility). As a result, HR should be the most relevant function for driving AI implementation in organizations.

PID will play a pivotal role in facilitating this transition, and AI is positioned to profoundly transform the PID landscape. Despite inevitable resource challenges, it is imperative for HR to lead the way in the strategic introduction of AI within the competitive arena. HR’s involvement should encompass fostering AI competency, influencing AI legislation and standards, and designing the AI-PID development roadmap. As AI continues its rapid expansion, one of HR’s critical responsibilities lies in cultivating employee trust in AI systems. This can be achieved by safeguarding employee privacy, ensuring AI is trained on high-quality data, and enforcing ethical standards for AI algorithms. By taking these actions, HR can harness the potential of AI to revolutionize PID, supporting advanced pay methodologies, breaking down barriers to seeking and sharing pay information, and providing valuable education to employees. Moreover, there is even a possibility for AI to deliver empathetic and supportive pay communication, all while upholding employee privacy. Simultaneously, companies stand to benefit from greater pay system ROI as employees gain a better understanding of AI and respond positively to its implementation. Therefore, HR’s proactive approach to integrating AI into the PID landscape can usher in transformative and positive changes while bolstering employee satisfaction and trust.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Anthony J. Nyberg

Author Biographies

Anthony J. Nyberg is a Professor at the University of South Carolina who teaches Negotiations, Rewards, and Research Methods. His research examines how organizations compete through people, specifically the strategic role of pay in the attraction, emergence, retention, and motivation of human capital resources. He publishes in top academic journals, and his editorial service has included serving as an Associate Editor for the Academy of Management Journal. He has received awards for teaching, service, and research including awards for best dissertation, best published manuscript (twice), and the Early Career Achievement award from the Academy of Management’s HR Division. His research has been highlighted in major international media outlets. Anthony received his doctorate from the University of Wisconsin-Madison. Prior to that he spent nine years as the managing partner for an international financial services firm, where he held Series 7, 24, 55, and 63 licenses, and earned his Certified Financial Analyst designation.

Ormonde Cragun is an Assistant Professor of Human Resource Management at the University of Minnesota Duluth. His work focuses on how executives and incentives impact the workplace. Specifically, his research includes work on CEO succession, CEO personality, and compensation. His work has been published in the Journal of Management, the Annual Review of Organizational Psychology and Organizational Behavior, and the Journal of Organizational Effectiveness: People and Performance.

Samantha Conroy is an Associate Professor of Management in the Department of Management at Colorado State University. Her research and teaching interests include compensation, frontline workers, and employee well-being. Her work has appeared in journals such as Academy of Management Journal, Academy of Management Review, and Journal of Management. She currently serves as an associate editor at Human Resource Management Review.

Ingo Weller is Professor of Human Resource Management at the LMU Munich School of Management in Munich, Germany. His research and teaching interests include strategic human resource management, matching, and compensation. His work has appeared in journals such as Academy of Management Journal, Academy of Management Annals, Academy of Management Discoveries, Journal of Applied Psychology, Strategic Management Journal, and Journal of Management. He also served as an Associate Editor of the Academy of Management Journal.

References

Abdulsalam

Maltarich

M. A.

Nyberg

A. J.

Reilly

Martin

(2021). Individualized pay-for-performance arrangements: Peer reactions and consequences. Journal of Applied Psychology, 106(8), 1202–1223. https://doi.org/10.1037/apl0000820

Aggarwal

Mittal

Battineni

(2021). Generative adversarial network: An overview of theory and applications. International Journal of Information Management Data Insights, 1(1), 100004. https://doi.org/10.1016/j.jjimei.2020.100004

Alterman

Bamberger

P. A.

Wang

Koopmann

Belogolovsky

Shi

(2021). Best not to know: Pay secrecy, employee voluntary turnover, and the conditioning effect of distributive justice. Academy of Management Journal, 64(2), 482–508. https://doi.org/10.5465/amj.2019.0231

Arnold

Fulmer

I. S.

(2019). Pay transparency. In The Routledge companion to reward management (pp. 87–96). Routledge.

Bamberger

P. A.

(2021). Pay transparency: Conceptualization and implications for employees, employers, and society as a whole. In Oxford research encyclopedia of business and management. Oxford University Press. https://doi.org/10.1093/acrefore/9780190224851.013.347

Bamberger

P. A.

(2023). Exposing pay: Pay transparency and what it means for employees, employers, and public policy. Oxford University Press.

Belogolovsky

Bamberger

P. A.

(2014). Signaling in secret: Pay for performance and the incentive and sorting effects of pay secrecy. Academy of Management Journal, 57(6), 1706–1733. https://doi.org/10.5465/amj.2012.0937

Bernstein

E. S.

(2017). Making transparency transparent: The evolution of observation in management theory. The Academy of Management Annals, 11(1), 217–266. https://doi.org/10.5465/annals.2014.0076

Bilquise

Ibrahim

Shaalan

(2022). Emotionally intelligent chatbots: A systematic literature review. Human Behavior and Emerging Technologies, 2022(3), 1–23. https://doi.org/10.1155/2022/9601630

10.

Blumkin

Lagziel

(2019). Relative ambition and the role of wage secrecy in labor contracts. Working papers 1902, Ben-Gurion University of the Negev, Department of Economics. https://ideas.repec.org/p/bgu/wpaper/1902.html

11.

Bostrom

Yudkowsky

(2018). The ethics of artificial intelligence. In Artificial intelligence safety and security (pp. 57–69). Chapman and Hall/CRC.

12.

Brandes

Darai

(2017). The value and motivating mechanism of transparency in organizations. European Economic Review, 98, 189–198. https://doi.org/10.1016/j.euroecorev.2017.06.014

13.

Brooks

B. A.

(2021). Determining and communicating employee pay: Transparency versus the black box. Nurse Leader, 19(3), 225–226. https://doi.org/10.1016/j.mnl.2021.04.004

14.

Brougham

Haar

(2018). Smart technology, artificial intelligence, robotics, and algorithms (STARA): Employees’ perceptions of our future workplace. Journal of Management and Organization, 24(2), 239–257. https://doi.org/10.1017/jmo.2016.55

15.

Brown

Nyberg

A. J.

Weller

Striver

S. D.

(2022). Pay information disclosure: Review and recommendations for research spanning the pay secrecy–pay transparency continuum. Journal of Management, 48(6), 1661–1694. https://doi.org/10.1177/01492063221079249

16.

Burkert

Oberpaul

Tichy

Weller

(2023). Executive compensation complexity and firm performance. Academy of Management Discoveries, online first publication. https://doi.org/10.5465/amd.2022.0056

17.

California Senate Bill 1162, 432.3 (2023).

18.

Candemir

Nguyen

X. V.

Folio

L. R.

Prevedello

L. M.

(2021). Training strategies for radiology deep learning models in data-limited scenarios. Radiology: Artificial Intelligence, 3(6), Article e210014. https://doi.org/10.1148/ryai.2021210014

19.

Charlwood

(2021). Artificial intelligence and talent management. Navigating the human-technology interface. In Wiblen

(Ed.), Digitalized talent management (1st ed., pp. 122–126). Routledge.

20.

Charlwood

Guenole

(2022). Can HR adapt to the paradoxes of artificial intelligence? Human Resource Management Journal, 32(4), 729–742. https://doi.org/10.1111/1748-8583.12433

21.

Conroy

S. A.

Yeong

J. Y.

Bamberger

P. A.

Gerhart

Gupta

Nyberg

A. J.

Park

T.-Y.

Shaw

J. D.

Sturman

M. C

(2015). Past, present and future compensation research perspectives. Compensation & Benefits Review, 47(5–6), 207–215. https://doi.org/10.1177/0886368716681677.

22.

Conroy

S. A.

Morton

J. W.

(2023). Back to basics in human resource theorizing: A call for greater attention to jobs. Human Resource Management Review. https://doi.org/10.1016/j.hrmr.2023.100973.

23.

Costas

Grey

(2014). Bringing secrecy into the open: Towards a theorization of the social processes of organizational secrecy. Organization Studies, 35(10), 1423–1447. https://doi.org/10.1177/0170840613515470

24.

Cullen

Z. B.

Pakzad-Hurson

(2019). Equilibrium effects of pay transparency in a simple labor market. Proceedings of the 2019 ACM Conference on Economics and Computation (pp. 193–193). https://doi.org/10.1145/3328526.3329645

25.

Dahl

M. S.

Pierce

(2020). Pay-for-performance and employee mental health: Large sample evidence using employee prescription drug usage. Academy of Management Discoveries, 6(1): 12–38. https://doi.org/10.5465/amd.2018.0007

26.

Da Silveira

Borenstein

Fogliatto

F. S.

(2001). Mass customization: Literature review and research directions. International Journal of Production Economics, 72(1), 1–13. https://doi.org/10.1016/S0925-5273(00)00079-7

27.

Day

N. E.

(2007). An investigation into pay communication: Is ignorance bliss? Personnel Review, 36(5), 739–762. https://doi.org/10.1108/00483480710774025

28.

Day

N. E.

(2011). Perceived pay communication, justice and pay satisfaction. Employee Relations, 33(5), 476–497. https://doi.org/10.1108/01425451111153862

29.

Delfanti

Frey

(2021). Humanly extended automation or the future of work seen through Amazon patents. Science, Technology & Human Values, 46(3), 655–682. https://doi.org/10.1177/0162243920943665

30.

Del Prado

(2015, October 26). 18 Artificial intelligence researchers reveal the profound changes coming to our lives. Business Insider. https://www.businessinsider.com/researchers-predictions-future-artificial-intelligence-2015-10

31.

Dwivedi

Y. K.

Hughes

Ismagilova

Aarts

Coombs

Crick

Duan

Dwivedi

Edwards

Eirug

Galanos

Ilavarasan

P. V.

Janssen

Jones

Kar

A. K.

Kizgin

Kronemann

Lal

Lucini

Williams

M. D.

(2021). Artificial intelligence (AI): Multidisciplinary perspectives on emerging challenges, opportunities, and agenda for research, practice and policy. International Journal of Information Management, 57, 101994. https://doi.org/10.1016/j.ijinfomgt.2019.08.002

32.

Ernst

Merola

Samaan

(2019). Economics of artificial intelligence: Implications for the future of work. IZA Journal of Labor Policy, 9(1), 20190004. https://doi.org/10.2478/izajolp-2019-0004

33.

Florens

J.-P.

Marimoutou

Péguin-Feissolle

(2007). Econometric modeling and inference. Cambridge University Press.

34.

Froomkin

A. M.

(2019). Big data: Destroyer of informed consent. Yale JL & Tech, 21(27).

35.

Leong

H. V.

Ngai

Huang

M. X.

Chan

S. C. F.

(2017). Physiological mouse: Toward an emotion-aware mouse. Universal Access in the Information Society, 16(2), 365–379. https://doi.org/10.1007/s10209-016-0469-9

36.

Fulmer

I. S.

Arnold

(2020). Pay communication: A global look at practices and preferences. J. Total Rewards, 29(4), 25–35.

37.

Fulmer

I. S.

Gerhart

Kim

J. H.

(2023). Compensation and performance: A review and recommendations for the future. Personnel Psychology. https://doi.org/10.1111/peps.12583

38.

Furneaux

(2012). Task-technology fit theory: A survey and synopsis of the literature. Information Systems Theory: Explaining and Predicting Our Digital Society, 28, 87–106. https://doi.org/10.1007/978-1-4419-6108-2_5.

39.

Gerhart

(2023). Compensation. McGraw-Hill Higher Education.

40.

Gerhart

Rynes

(2003). Compensation: Theory, evidence, and strategic implications. Sage publications.

41.

Giannopoulou

(2020). Algorithmic systems: The consent is in the detail? Internet Policy Review, 9(1). https://doi.org/10.14763/2020.1.1452

42.

Gil

Virgili-Gomá

García

Mason

(2015). Emotions ontology for collaborative modelling and learning of emotional responses. Computers in Human Behavior, 51, 610–617. https://doi.org/10.1016/j.chb.2014.11.100

43.

Goodhue

D. L.

Thompson

R. L.

(1995). Task-technology fit and individual performance. MIS Quarterly, 19(2), 213. https://doi.org/10.2307/249689

44.

Guenole

Feinzig

(2018). The business case for AI in HR. With insights and tips on getting started. [White paper] Armonk: IBM Smarter Workforce Institute, IBM Corporation. https://www.ibm.com/downloads/cas/A5YLEPBR

45.

Gupta

(2013). Informed consent in clinical research: Revisiting few concepts and areas. Perspectives in Clinical Research, 4(1), 26. https://doi.org/10.4103/2229-3485.106373

46.

Hoffman

Kahn

L. B.

(2018). Discretion in hiring. Quarterly Journal of Economics, 133(2), 765–800. https://doi.org/10.1093/qje/qjx042

47.

Howard

(2019). Artificial intelligence: Implications for the future of work. American Journal of Industrial Medicine, 62(11), 917–926. https://doi.org/10.1002/ajim.23037

48.

Hughes

Robert

Frady

Arroyos

(2019). Artificial intelligence, employee engagement, fairness, and job outcomes. In Hughes

Robert

Frady

Arroyos

, Managing technology and middle- and low-skilled employees (pp. 61–68). Emerald Publishing Limited. https://doi.org/10.1108/978-1-78973-077-720191005

49.

James

Harbron

Branson

Sundler

(2021). Synthetic data use: Exploring use cases to optimize data utility. Discover Artificial Intelligence, 1(1), 15. https://doi.org/10.1007/s44163-021-00016-y

50.

Jarrahi

M. H.

(2018). Artificial intelligence and the future of work: Human-AI symbiosis in organizational decision making. Business Horizons, 61(4), 577–586. https://doi.org/10.1016/j.bushor.2018.03.007

51.

Kim

P. T.

Bodie

M. T.

(2021). Artificial intelligence and the challenges of workplace discrimination and privacy. Journal of Labor and Employment Law, 35(2), 289–315.

52.

Klykken

F. H.

(2022). Implementing continuous consent in qualitative research. Qualitative Research, 22(5), 795–810. https://doi.org/10.1177/14687941211014366

53.

Krapp

(2022, July 27). The future of security: Surveillance is pervasive even if no one is looking for you. Governing. https://www.governing.com/security/surveillance-is-pervasive-even-if-no-one-is-looking-for-you

54.

Leetaru

(2019, February 2). What does it mean to “consent” to the use of our data in the Facebook era. Forbes. https://www.forbes.com/sites/kalevleetaru/2019/02/02/what-does-it-mean-to-consent-to-the-use-of-our-data-in-the-facebook-era/?sh=3d6fcb92b069

55.

Luo

Lau

R. Y. K.

(2022). A critical review of state‐of‐the‐art chatbot designs and applications. WIREs Data Mining and Knowledge Discovery, 12(1). https://doi.org/10.1002/widm.1434

56.

Maier

Laumer

Thatcher

J. B.

Wirth

Weitzel

(2022). Trial-period technostress: A conceptual definition and mixed-methods investigation. Information Systems Research, 33(2): 489–514. https://doi.org/10.1287/isre.2021.1047

57.

Maltarich

Nyberg

A. J.

Reilly

Abdulsalam

D. “D.

Martin

(2017). Pay-for-performance, sometimes: An interdisciplinary approach to integrating economic rationality with psychological emotion to predict individual performance. Academy of Management Journal, 60(6), 2155–2174. https://doi.org/10.5465/amj.2015.0737

58.

Marasi

Wall

Bennett

R. J.

(2018). Pay openness movement: Is it merited? Does it influence more desirable employee outcomes than pay secrecy? Organization Management Journal, 15(2), 58–77. https://doi.org/10.1080/15416518.2018.1471978

59.

Marwala

(2013). Economic modeling using artificial intelligence methods. Springer.

60.

Miller

C. C.

(2015, June 25). Can an algorithm hire better than a human? The New York Times. https://www.nytimes.com/2015/06/26/upshot/can-an-algorithm-hire-better-than-a-human.html

61.

Nyberg

(2010). Retaining your high performers: Moderators of the performance–job satisfaction–voluntary turnover relationship. Journal of Applied Psychology, 95(3), 440. https://doi.org/10.1037/a0018869

62.

Nyberg

A. J.

Maltarich

M. A.

Abdulsalam

D. D.

Essman

S. M.

Cragun

(2018). Collective pay for performance: A cross-disciplinary review and meta-analysis. Journal of Management, 44(6), 2433–2472. https://doi.org/10.1177/0149206318770732.

63.

Nyberg

A. J.

Pieper

J. R.

Trevor

C. O.

(2016). Pay-for-performance’s effect on future employee performance: Integrating psychological and economic principles toward a contingency perspective. Journal of Management, 42(7), 1753–1783. https://doi.org/10.1177/0149206313515520

64.

Nyberg

A. J.

Reilly

(2019). A pay system model for turning human capital resources into action. In Handbook of research on strategic human capital resources (pp. 384–402). Edward Elgar Publishing. https://doi.org/10.4337/9781788116695.00038.

65.

O’Leary

D. E.

(2013). Artificial intelligence and Big data. IEEE Intelligent Systems, 28(2), 96–99. https://doi.org/10.1109/MIS.2013.39

66.

Pasquale

(2015). The black box society: The secret algorithms that control money and information. Harvard University Press.

67.

Privacy Forum . (2018). The privacy expert’s guide to artificial intelligence and machine learning. [White paper]. Future of Privacy Forum. https://fpf.org/wp-content/uploads/2018/10/FPF_Artificial-Intelligence_Digital.pdf

68.

Qiu

Liu

Zhou

(2019). Review of artificial intelligence adversarial attack and defense technologies. Applied Sciences, 9(5), 909. https://doi.org/10.3390/app9050909

69.

Rampersad

(2020). Robot will take your job: Innovation for an era of artificial intelligence. Journal of Business Research, 116(4), 68–74. https://doi.org/10.1016/j.jbusres.2020.05.019.

70.

Rapp

Curti

Boldi

(2021). The human side of human-chatbot interaction: A systematic literature review of ten years of research on text-based chatbots. International Journal of Human-Computer Studies, 151, 102630. https://doi.org/10.1016/j.ijhcs.2021.102630

71.

Rodrigues

(2020). Legal and human rights issues of AI: Gaps, challenges and vulnerabilities. Journal of Responsible Technology, 4, 100005. https://doi.org/10.1016/j.jrt.2020.100005

72.

Rosenquist

J. C.

Visconti

D. M.

(2023, January 3). California releases guidance on pay transparency law. SHRM. SHRM.Org/Resources/State & Local Updates. https://www.shrm.org/resourcesandtools/legal-and-compliance/state-and-local-updates/pages/california-pay-transparency-guidance.aspx

73.

Sammer

(2019, December 10). Bringing artificial intelligence into pay decisions. SHRM.Org/Resources/Compensation. https://www.shrm.org/resourcesandtools/hr-topics/compensation/pages/bringing-artificial-intelligence-into-pay-decisions.aspx

74.

Scassa

(2018). Data ownership. CIGI papers No. 187. Ottawa faculty of law working paper No. 2018-26. https://www.cigionline.org/static/documents/documents/Paperno.187_2.pdf

75.

Shrestha

Y. R.

Ben-Menahem

S. M.

Von Krogh

(2019). Organizational decision-making structures in the age of artificial intelligence. California Management Review, 61(4), 66–83. https://doi.org/10.1177/0008125619862257

76.

Stark

(2023, January 31). SIOP releases recommendations for AI-based assessments. SIOP. https://www.siop.org/Research-Publications/Items-of-Interest/ArtMID/19366/ArticleID/7327/SIOP-Releases-Recommendations-for-AI-Based-Assessments

77.

Strohmeier

Piazza

(2015). Artificial intelligence techniques in human resource management—A conceptual exploration. In Kahraman

Çevik Onar

(Eds.), Intelligent techniques in engineering management (Vol. 87, pp. 149–172). Springer International Publishing. https://doi.org/10.1007/978-3-319-17906-3_7

78.

Tambe

Cappelli

Yakubovich

(2019). Artificial intelligence in human resources management: Challenges and a path forward. California Management Review, 61(4), 15–42. https://doi.org/10.1177/0008125619867910

79.

Wang

Siau

(2019). Artificial intelligence, machine learning, automation, robotics, future of work and future of humanity: A review and research agenda. Journal of Database Management, 30(1), 61–79. https://doi.org/10.4018/JDM.2019010104

80.

Weller

Hymer

C. B.

Nyberg

A. J.

Ebert

(2019). How matching creates value: Cogs and wheels for human capital resources research. The Academy of Management Annals, 13(1), 188–214. https://doi.org/10.5465/annals.2016.0117

81.

West

S. M.

(2019). Data capitalism: Redefining the logics of surveillance and privacy. Business & Society, 58(1), 20–41.

82.

Williams

M. L.

McDaniel

M. A.

Nguyen

N. T.

(2006). A meta-analysis of the antecedents and consequences of pay level satisfaction. Journal of Applied Psychology, 91(2), 392. https://doi.org/10.1037/0021-9010.91.2.392.

83.

Xiang

(2022, December 6). AI isn’t artificial or intelligent: How AI innovation is powered by underpaid workers in foreign countries. Vice. https://www.vice.com/en/article/wxnaqz/ai-isnt-artificial-or-intelligent

84.

Yadegaridehkordi

Noor

N. F. B. M.

Ayub

M. N. B.

Affal

H. B.

Hussin

N. B.

(2019). Affective computing in education: A systematic review and future research. Computers & Education, 142, 103649. https://doi.org/10.1016/j.compedu.2019.103649

85.

Yavorsky

(2022, August 11). New U.S. state privacy laws zero in on artificial intelligence. Orrick: Technology & Innovation, Energy & Infrastructure, Finance. https://www.orrick.com/en/Insights/2022/08/New-State-Privacy-Laws-Zero-in-on-AI

86.

Zenger

(2017, August 13). The downside of full pay transparency. Wall Street Journal. https://www.wsj.com/articles/the-downside-of-full-pay-transparency-1502676360#

87.

Zuboff

(1985). Automatefin-fonnate: The two faces of intelligent technology. Organizational Dynamics, 14(2), 5–18. https://doi.org/10.1016/0090-2616(85)90033-6