The Ethical Role of Generative Artificial Intelligence in Modern HR Decision-Making: A Systematic Literature Review

The rapid development of generative artificial intelligence (AI) has led to the recognition of tools like ChatGPT and its potential to transform human resource (HR) management processes, particularly in decision-making. This review study aims to assess the effectiveness and benefits of ChatGPT in enhancing HR functions, particularly decision-making, and to identify any challenges and ethical considerations involved. Additionally, the study seeks to establish a hybrid framework that combines AI-driven decision-making with human oversight. A systematic literature review was conducted using PRISMA guidelines, selecting 50 articles from Scopus and Google Scholar databases. The literature review includes a synthesis analysis to assess publication trends and a keyword analysis to identify key themes such as ChatGPT’s impact on decision-making in HR management. The study reveals that ChatGPT can streamline HR processes, improve communication, and support personalized learning and decision-making, eventually contributing to enhanced performance and engagement. However, the technology requires human input for moral judgment and empathy, presenting challenges like resistance to adoption, algorithmic bias, and data privacy concerns. This study uniquely contributes to the literature by providing a systematic analysis of ChatGPT’s role in HR decision-making and proposing a hybrid framework that addresses AI’s limitations through ethical guidelines and human oversight. The findings emphasize the need for empirical research in larger, diverse settings and future enhancements to ChatGPT’s contextual understanding of HR.

Introduction

The success of a business or organization is significantly influenced by human resource management (HRM), which directly affects workforce productivity, satisfaction, and retention. With effective and efficient HRM, organizations can better accomplish their goals, ensuring that every team member performs at their best. However, human resource (HR) managers also face complex issues, such as managing large teams, addressing diverse demands and expectations among team members, and adapting to rapid changes in the business and technological context (Vahdat, 2022). In recent years, artificial intelligence (AI) has become transformative for HRM, prompting experts to suggest that HRM must undergo substantial, rather than incremental, changes to remain effective in a technology-driven era (Sutrisno & Rijal, 2024). In addition to the changes in fundamental HRM processes like hiring, selection, and training, researchers have examined AI’s impact on areas like diversity management and health and safety monitoring (Rabenu & Baruch, 2024).

One of the most noteworthy advancements in AI is ChatGPT, a model that has recently made a significant impact in the tech world. As a trained AI model capable of engaging in human-like text-based conversations, ChatGPT has proven valuable in applications such as research, marketing, brainstorming, customer engagement, and learning (Bukaret al., 2024). Its natural language processing capabilities allow it to follow up on questions, provide historical and contextual information, and even respond to complex inquiries with significant accuracy. Transformer models like ChatGPT enable the creation of chatbots or virtual assistants capable of assisting in decision-making processes. ChatGPT, in particular, can understand user inquiries, follow instructions based on the information provided, and deliver relevant responses (Saiet al., 2024). While ChatGPT offers various benefits to businesses, it also presents certain challenges, such as potential issues with non-standard terms or unfamiliar language, which can be problematic for some users. Thus, HRM, like other fields, has been significantly impacted by ChatGPT and other generative AI systems, which have become essential, evolving tools in organizational operations.

Several studies have explored the role of ChatGPT across various fields, such as healthcare (Liuet al., 2023), finance (Ahangar & Fietko, 2023), business and management (Korzynskiet al., 2023), supply chains (Haddud, 2024), tourism (Stergiou & Nella, 2024), and education (Montenegro-Ruedaet al., 2023). However, research examining ChatGPT’s application specifically within HRM is still limited. Some studies have addressed general HRM practices (Rane, 2023a; Iswahyudiet al., 2023), but they have yet to focus on the decision-making capabilities of ChatGPT in HR contexts. The current literature suggests a notable gap in understanding how generative AI can aid HR decision-making specifically. In the field of HR, professionals, decision-makers, and executives are actively working to determine effective policies and practices for using AI in decision-making (Iswahyudiet al., 2023). Thus, this review study aims to bridge the research gap by identifying strategies to address the challenges and maximize the benefits of ChatGPT as a decision-support tool in HRM.

The purpose of this study is to fill the research gap on assessing the role of ChatGPT and its potential applications as a decision-support tool in HRM (Iswahyudiet al., 2023). This review employs a qualitative approach, analyzing and interpreting data from a wide range of sources. The study’s contributions include synthesizing ChatGPT’s role in HR decision-making, identifying the challenges it presents, evaluating the quality of responses generated by ChatGPT, and offering insights for organizations on effectively using ChatGPT for HR decision-making. Additionally, this review provides recommendations to help organizations use generative AI effectively and responsibly. By assessing the current landscape of generative AI in HR, this study aims to provide a foundation for understanding ChatGPT’s potential benefits and limitations, empowering organizations to leverage it without hesitation. To guide this literature review, the following research questions have been formulated to investigate the applications of generative AI, particularly ChatGPT, in HR decision-making:.

Thus, the following are the research questions framed to carry out the review study on the leadership approaches for post-pandemic recovery:

RQ1: What are the primary ways ChatGPT supports HR decision-making?

RQ2: What challenges do HR managers face when employing ChatGPT?

RQ3: What impact does ChatGPT have on the quality and speed of HR decisions?

RQ4: How can organizations ensure that generative AI tools support fair and unbiased HR decisions?

The structure of the paper is as follows: Section 2 presents the theoretical background study, providing a foundation for understanding the impact of ChatGPT on HRM. Section 3 details the methodology used for this review study. Section 4 discusses the analysis and answers to the research questions, evaluating ChatGPT’s role in HR decision-making. Section 5 discusses the results in relation to managerial implications, providing actionable insights for HR managers. Finally, Section 6 concludes with a summary of findings and recommendations for further research.

Theoretical Background

Overview of ChatGPT

The AI-powered chatbot ChatGPT (Generative Pretrained Transformer), developed by OpenAI, was introduced in November 2022. Designed for conversational text-based interactions, ChatGPT quickly gained attention for its human-like dialogue capabilities (Liuet al., 2023). The initial model, based on GPT-3.5, offered impressive conversational abilities, and in March 2023, OpenAI released ChatGPT-4, which introduced improvements in handling complex tasks, creativity, and accuracy. ChatGPT’s dialogue format allows it to respond to follow-up inquiries, recognize errors, challenge incorrect assumptions, and decline inappropriate requests. Beyond its conversational abilities, ChatGPT supports various applications, including research, marketing content generation, idea development, and customer engagement (Jainet al., 2024). Key characteristics of ChatGPT include natural language understanding, contextual awareness, adaptability, responsiveness, scalability, and efficiency. However, it also has limitations, such as privacy concerns, ethical boundaries, non-personalized responses across sessions, and sensitivity to user input.

Applications of ChatGPT Across Sectors

ChatGPT has been studied across various industries. In healthcare, large language models (LLMs) like ChatGPT enhance shared decision-making (SDM) by enabling patients and clinicians to collaborate based on the best available evidence and to weigh options for well-informed decisions (Mahmudin, 2023). ChatGPT in education aids text summarization and reduces workload but also raises risks of plagiarism, misinformation, and copyright infringement (Bukaret al., 2024; Bhatet al., 2024). These risks highlight the need for institutions and policymakers to implement flexible decision-making approaches to adapt to evolving challenges raised by ChatGPT.

In business and industry, research on ChatGPT’s performance remains limited. Wanget al. (2023) reported that while ChatGPT provides useful information on established concepts like Industry 5.0, it struggles with more abstract ideas and requires clear prompts for optimal performance. A key limitation of ChatGPT is its dependency on specific, explicit information from users, which affects its ability to interpret implied cues. However, its seamless integration into business processes offers advantages, enhancing customer experience by delivering personalized recommendations, comprehensive support, and increased engagement, ultimately driving revenue (Kumaret al., 2024).

In the strategic business environment, researchers emphasize the importance of training management teams to understand both the capabilities and limitations of digital technologies for effective use in decision-making (Mahmudin, 2023; Porkodiet al., 2023). Successful implementation depends on data security, transparency, and adherence to privacy laws, especially for startups that must monitor and assess ChatGPT’s performance to ensure sustainable growth. Furthermore, as ChatGPT has the potential to learn biases from data, it requires continuous refinement with clear boundaries and security measures to support reliable decision-making in business applications. This is equally relevant to HR decision-making, where ChatGPT’s role could bring significant value to companies.

ChatGPT’s potential in business management and strategic decision-making is demonstrated through rapid responses, data analysis, and customized interactions, which enhance productivity and efficiency. Common applications include business process automation, improved customer service, and data-driven decision-making (Jusmanet al., 2023). Some employees believe AI could replace humans in routine tasks and decision-making, while managers argue that humans should retain ultimate decision-making authority. Employees and managers recognize the importance of developing generative AI but view human learning as less critical in routine, repetitive tasks (Yamamura & Ohtake, 2024).

ChatGPT in HR Decision-Making

HR decision-making varies by context and factors, often involving areas like recruitment and selection, talent development, performance management, compensation, employee relations, workforce planning, diversity and inclusion, change management, compliance, technology adoption, and employee well-being (Toplaket al., 2010). These contexts influence the types of decisions HR managers must make, which vary based on factors like company size, industry, workforce composition, and strategic objectives.

Studies exploring ChatGPT’s role in HRM are relatively scarce. Srivastavaet al. (2024) proposed a multi-criteria decision-making (MCDM) method to determine the optimal chatbot for specific tasks. Their study used a combination of literature, expert input, and analytical methods (CRITIC, WASPAS, and EDAS) to rank different chatbots, with ChatGPT emerging as the preferred choice over others like YOU, PerplexityAI, ChatSonic, and CharacterAI. Similarly, Ramanet al. (2024) found ChatGPT outperforms other generative AI tools like Bard in accuracy, thus assessing it as the primary AI tool in this study.

Although ChatGPT shows potential to replace Bard in transactional HR roles, Bard may offer greater security against misuse, making the two tools complementary rather than direct competitors (Ramanet al., 2024). While accuracy, relevance, and clarity vary slightly between ChatGPT and Bard, they can complement each other based on organizational needs. Hamoucheet al. (2023) examined AI’s impact on HR development (HRD) using a bibliometric approach to analyze the literature on ChatGPT and machine learning developments. Their study, with a focus on machine learning’s role in HRD, encourages interdisciplinary collaboration and emphasizes the growing role of technology in HR education and workplace development.

Korzynskiet al. (2023) highlighted the transformative potential of AI in HR by identifying possible shifts in traditional management theories related to customer service, HR practices, and decision-making. They suggested that the adoption of AI ensures a re-assessment of these theories, particularly as Industry 4.0 and 5.0 evolve. Generative AI, including ChatGPT, has the potential to alter data analysis and decision-making processes fundamentally, making this field of study increasingly relevant. Consequently, a thorough evaluation of ChatGPT’s strengths, limitations, and challenges is necessary before integrating it into practice. Thus, this study aims to assess the application of generative AI, specifically ChatGPT, in HR decision-making.

Methodology

A systematic literature review was conducted to achieve the objectives of this proposed study and provide a comprehensive overview of the findings. This systematic literature review follows standard PRISMA guidelines, ensuring transparency at each review step (Porkodi & Raman, 2024). This approach involves three phases: (1) locating the studies from data sources, (2) screening and assessing the selected studies, and (3) including selected studies for analysis. This systematic process includes keyword identification, study selection, bibliometric analysis, full-text analysis, and a review of findings to reduce the bias present in non-systematic reviews. Fig. 1 displays the comprehensive PRISMA flowchart for the systematic review carried out for the proposed study.

Fig. 1. PRISMA flowchart for the proposed systematic review.

Initially, relevant research works were extracted from various standard databases, such as Scopus and Google Scholar, between October 29 and 31, 2024. In order to capture ChatGPT’s essence from a wide range of literature sources, this study uses Google Scholar as well as Scopus, a well-known database for study selection. The keywords used for locating the studies include “human resource management,” “human resources,” “generative AI,” “ChatGPT,” “decision-making,” and “HR decisions.” Since ChatGPT was introduced in 2022, no time frame limitations were applied for selecting studies. The total number of articles initially identified for the study was 319, comprising 231 articles retrieved from Scopus and 88 articles from Google Scholar. The initial screening included duplicate removal, as articles identified from various sources often overlapped. This step eliminated 74 duplicates, resulting in a total of 245 unique articles. Subsequently, in the screening and eligibility phase, articles irrelevant to the study were excluded. After screening 245 articles based on their titles and abstracts, 97 articles relevant to the study were retained, while 148 irrelevant articles were excluded. Of the 97 articles selected for retrieval, five could not be retrieved except for their title, resulting in a total of 92 articles. The eligibility of these articles was then evaluated by examining their complete content, leading to the selection of 48 articles for further analysis. Several exclusion criteria were applied to filter out irrelevant articles from the review study, including:

• Articles from non-peer-reviewed sources, such as blogs and non-English articles.

• Articles that do not focus on HRM, generative AI, and decision-making.

• Articles with unclear methodologies or evident biases.

• Articles lacking consideration of managerial implications on HR decision-making.

• Articles lacking empirical evidence or valid research methods.

Additionally, two articles were identified through the references of the selected studies, which ended up in 50 articles.

Synthesis Analysis

Utilizing bibliometric data analysis and visualization tools like Biblioshiny and VOSviewer, this section provides a basic synthesis analysis of the selected studies for easy interpretation. Among the 50 articles selected, approximately 19 were published in 2023, 30 in 2024, and one article is an early publication for 2025. This trend clearly indicates that the academic production of studies related to ChatGPT in HR decision-making is increasing year over year and is likely to continue growing in the future. Furthermore, 70% of the studies are articles, 14% are simple reviews, 12% are conference papers, and 4% are book chapters. This highlights that the topic under study is both timely and necessary for further exploration in this field.

The keywords listed in the selected studies are visualized using a network diagram in Fig. 2. As seen in Fig. 2, ‘ChatGPT’ (35) is the most frequently used term, with related terms like ‘generative AI’ (16), ‘artificial intelligence’ (12), ‘decision making’ (11), and ‘human resource management’ (11) also appearing prominently. Other common terms include ‘HRM’ (5), ‘chatbots’ (4), ‘large language models’ (4), ‘ethics’ (3), ‘business’ (3), ‘AI’ (2), ‘AI chatbots’ (2), ‘HR bots’ (2), ‘human resource technology’ (2), and ‘machine learning’ (2). The analysis of significant words listed in the titles of the selected studies is further illustrated through a word cloud in Fig. 3, which highlights words such as ‘ChatGPT,’ ‘AI,’ ‘human resource management,’ ‘artificial intelligence,’ ‘decision making,’ ‘business,’ ‘challenges,’ and ‘chatbots.’ This highlights the research trends and key topics in the field of study.

Fig. 2. Network visualization of authors’ keywords.

Fig. 3. Word cloud of title of selected studies.

The above analysis, including year-wise distribution, article types, and keyword network analysis, provides insights into research trends, growth patterns, dominant topics, and conceptual relationships within the field. However, it does not yet offer an in-depth analysis or critical evaluation of ChatGPT’s role in HR decision-making to fully address the research questions. Next section reviews articles on ChatGPT’s strengths, challenges, and impact on HR decision-making quality, addressing research questions.

Findings

ChatGPT can consistently offer significant opportunities for companies that strategically leverage this groundbreaking technology. GPT can enhance business processes by improving client relationships, fostering creativity, and optimizing workflows while also transforming financial brokers’ roles by enhancing productivity and decision-making (Daret al., 2024). Given the growing popularity of AI tools like ChatGPT and DALL-E-2, marketers often struggle to integrate them into marketing campaigns and operations. Zhang and Agnihotri (2024) suggested a customer-centric strategy to assist managers in making these decisions. AlQershiet al. (2024) discovered that organizational support and managerial productivity significantly impact business sustainability, while decision aids do not.

AI and generative AI impact HR development practices and the organizational ecosystem by raising concerns about bias, fairness, transparency, safety, job loss, privacy intrusion, and agency. Extensive frameworks and guidelines in HR development fail to adequately address the ethical issues surrounding AI and generative AI. A study by Yorks and Jester (2024) highlighted the necessity of an extensive structure that promotes justice, openness, and privacy while directing the moral application of AI and generative AI in HR development procedures. The Multi-Criteria Decision Making (MCDM) aspect of ChatGPT on supplier evaluation was assessed by Wang and Wu (2024). The ChatGPT model demonstrated its multi-criteria decision-making capabilities, speed, and cost-effectiveness by aligning supplier evaluations with human experts after examining the supplier dataset using conventional MCDM models.

Several studies have proposed frameworks to facilitate the effective use of ChatGPT. A framework for adopting generative AI, specifically tools like ChatGPT, within entrepreneurial and innovation ecosystems was proposed, involving three phases of adoption: pre-perception and perception, assessment, and outcome (Gupta & Yang, 2024). This model aids in directing strategic choices regarding the integration of ChatGPT and other AI into workplace environments. In order to improve organizational decision-making across a range of domains, Sansanee and Kiattisin (2024) proposed the AI PROMPT framework, which makes use of prompt engineering techniques to improve the efficacy and quality of text-to-text prompt engineering. Kromidha and Davison (2024) suggest that generative AI aids decision-making by organizing information, but without a moral and ethical stance, the responsibility remains with the human actor. To transform generative AI-augmented decision-making, a cooperative approach between humans and generative AI must be proposed, emphasizing the necessity of modifications in learning and adaptation patterns on both sides. To improve startup interactions with generative AI, Wanget al. (2022) used prompt engineering. The study found that customized prompts support brainstorming, a crucial startup activity, but the outcomes vary due to human factors. ChatGPT, while useful for HRD planning, has limitations in generic responses and lack of contextual awareness, emphasizing the need for human expertise to enhance its effectiveness (Ardichviliet al., 2024).

ChatGPT’s Support in HR Decision-Making

Numerous research studies have highlighted the significance of ChatGPT in decision-making, particularly in HRM (Rajet al., 2023). Generally, AI technologies revolutionize HR procedures by supporting employee training, performance analysis, and talent acquisition in HRM (Rane, 2023b). ChatGPT aids in employee development and training by creating personalized materials based on individual needs and utilizing generative AI tools to continuously enhance knowledge and skills (Jusmanet al., 2023). ChatGPT, with higher HR literacy levels, is a promising robotic advisor for transactional HR roles, providing individualized feedback for skill development and human capital building (Ramanet al., 2024). A study among senior HR managers underscored the significance of AI in global competitiveness, emphasizing workforce empowerment and efficient technology management (Poisatet al., 2024).

By responding to commonly asked questions, offering company information, and promoting smooth communication throughout the hiring process, ChatGPT can automate candidate evaluation, minimize HR effort, and provide personalized experiences (Sebastian, 2023). ChatGPT answers regular HR questions, enhancing employee support and engagement while freeing up HR professionals to focus on strategic projects. ChatGPT improves communication quality in HR teams, enhancing teamwork, performance management, training, and recruitment while also improving comprehension in HRM (Sutrisno & Rijal, 2024).

Kumaret al. (2025) found that a leader’s use of ChatGPT partially moderates the relationship between communicative leadership and employee engagement, bolstered by employees’ perception of communication. By using ChatGPT, leaders can increase trust, engagement, and performance, ultimately helping staff members more successfully accomplish company goals and objectives. ChatGPT, viewed as user-friendly AI technology, can enhance communication and provide real-time solutions, aiding HRM and decision-making (Porkodi, 2022).

ChatGPT can also improve workplace learning and organizational performance by suggesting learning resources, accelerating learning, and assisting staff in developing skills based on job demands and career goals (Diantoroet al., 2024). Its prompt insights and assistance encourage well-informed decision-making, empowering HR professionals to proactively address difficulties and seize opportunities. Rane (2024) explored the application of ChatGPT and other generative AI in HRM, emphasizing their potential to enhance hiring, employee training, and communication effectiveness. Sakibet al. (2024) conclude that ChatGPT offers HR benefits like automating tasks and optimizing strategies but may face challenges like limited understanding, emotional intelligence, privacy, and security concerns.

Iswahyudiet al. (2023) suggest ChatGPT, an AI model, as a potential tool for enhancing HR processes, specifically in recruitment, employee development, performance management, and support. In addition to providing text-based responses, it serves as a virtual assistant that collects important information about workers and the workplace, assisting HR managers in making better decisions. To improve HR strategies and decision-making, ChatGPT gathers and analyzes data on interactions between employees and job candidates (Sakibet al., 2024). It improves candidate experiences by ensuring consistency, reducing human error, and streamlining HR processes. AI chatbots are expected to revolutionize HR tasks like candidate interviews and hiring decisions due to the rapid advancement of technology.

According to Diantoroet al. (2024), medium-sized businesses can fully utilize ChatGPT to assist with strategic decision-making and achieve long-term success in the digital era by taking an integrated and comprehensive approach. Additionally, HR professionals can make well-informed decisions about resource allocation and talent development initiatives by using ChatGPT’s performance data analysis to spot trends and patterns. Businesses should regularly gather user feedback from employees, customers, and other users to continuously improve the service quality and performance of ChatGPT. Through comprehensive and detailed data support, the ChatGPT model enhances decision-making effectiveness. Its powerful semantic understanding and generation capabilities enable managers to make more intelligent HR decisions (Zhou & Cen, 2023). Managers are better able to understand employee performance due to the platform’s data analysis and predictive capabilities, which facilitate scientific decisions about talent management, promotions, and rewards.

Challenges Faced by HR Managers in Employing ChatGPT

Though many studies acknowledge the benefits and applications of ChatGPT in HR decision-making, they also highlight various challenges faced by HR managers when employing it. Studies show that innovation orientation and ChatGPT are positively correlated, indicating that managers who are more innovative and creative are more likely to use and customize ChatGPT (Ciminoet al., 2024). However, research indicates that human contributions to tasks involving creativity, moral judgment, empathy, and interpersonal skills may be more difficult for AI to replicate effectively (Adiasto, 2024). These findings underscore the need for a balanced approach in HR, where AI complements rather than replaces human contributions. The study explored various challenges in AI-powered HRM, including data privacy, algorithmic bias, and human-AI balance, offering valuable insights for HR professionals and policymakers (Rane, 2023a). Zhou and Cen (2023) revealed that ChatGPT-based digital HRM platforms offer better value and potential than traditional HRM platforms by personalizing employee experiences, improving decision-making efficiency, and enhancing productivity. However, the platform’s potential should be enhanced through improvements in data privacy, security, training, and performance to advance HRM theory.

Primarily, maintaining a language model like ChatGPT necessitates substantial computational resources, expertise, and continuous updates to stay up-to-date of new advancements (Ayindeet al., 2023). To optimize performance, organizations should allocate resources, invest in continuous training, hire experienced personnel, and establish a performance monitoring system to identify and address errors. Especially if HR employees are not familiar with AI technology, it may take time and resources to integrate ChatGPT into HR systems and train the team on its efficient use. Similarly, Abu-Shanabet al. (2023) highlight AI’s significant role in streamlining recruitment, personalizing employee experiences, reducing biases, and improving decision-making, though they also note resource limitations. Even large tech companies face restrictions that limit access to LLM and foundational models necessary for deeper understanding. The challenge of organizational resistance to change is also reported. Poisatet al. (2024) emphasize the importance of employee education and managerial support for successful AI integration despite challenges such as ignorance and resistance to change.

Language accuracy in ChatGPT prompts can also pose challenges. Tang and Kejriwal (2023) revealed that grammatical errors in prompts may lead to unintended behavior, but ChatGPT often remains consistent by ignoring minor errors, as it can adapt to small syntactic variations. Nevertheless, training ChatGPT effectively is essential, as proper training can lead to positive outcomes, while improper training can result in negative consequences. Even with adequate training, Sakibet al. (2024) found that ChatGPT may face challenges in understanding human context, emotional intelligence, privacy, and security concerns, underscoring the importance of targeted training for optimal results. ChatGPT gains knowledge from past data, which often includes biases. In such cases, ChatGPT’s suggestions may potentially reinforce biases in training data, potentially compromising equity in hiring or promotion decisions.

Bias remains a prominent issue with AI applications. Chenet al. (2023) conducted an experiment using GPT-3.5 and GPT-4 and observed that these models mirrored human biases in half of the standard context experiments while diverging in the remaining experiments. GPT models exhibit consistency in operations management tasks, with dual-edged progressions showing increased decision-making accuracy for mathematical solutions and behavioral biases for preference-based issues. Thus, while ChatGPT demonstrates high consistency and significant workflow advantages, preference-based decisions require caution due to susceptibility to biases across different contexts.

Moreover, ChatGPT lacks explainability and transparency, which limits its suitability for critical applications (Krügelet al., 2023). Ayindeet al. (2023) emphasize that an inexplicable decision-making procedure hinders the acceptance of AI in such applications. Basiret al. (2023) further highlight ethical challenges associated with using ChatGPT in leadership and strategic decision-making despite its advantages in risk management, transparency, fairness, privacy, and long-term impact considerations. The study explores ethical challenges, including privacy issues and the need for unbiased AI-driven decisions, alongside technical obstacles such as data fragmentation and the safeguarding of sensitive information.

There is also concern about over-reliance on technology potentially weakening critical and creative thinking among employees (Sumbalet al., 2024). This dependence could lead to reduced innovation and adaptability within the workforce. Yamamura and Ohtake (2024) strengthen this concern by reporting that while managers hold positive views of generative AI, they do not perceive it as suitable for responsibility in human decision-making, indicating a preference for a human-led approach in critical HR functions.

Through a detailed examination of several generative AI applications in Industry 5.0 and an analysis of real-world products integrating generative AI, these studies shed light on real-time benefits and limitations encountered. ChatGPT, while potentially improving HR functions and decision-making, requires careful implementation to overcome limitations and effectively address ethical, technical, and operational challenges (Saiet al., 2024).

Impact of ChatGPT on the Quality and Speed of HR Decisions

ChatGPT’s effectiveness in decision-making is largely unexplored due to limited focus on its quality, limiting its ability to accurately and reliably inform HR decision-making processes. Some of the studies that assessed the response quality are highlighted here. A study by Chuma and De Oliveira (2023) assessed ChatGPT’s effectiveness in various business decision-making scenarios, including hypothetical supermarket chain mergers in Sweden, Brazilian oil company investment recommendations, and online shopping behavior factors. The authors concluded that while ChatGPT enhances output and saves time, it cannot replace professional judgment in business decision-making. Although its development supports future business decision-making, it also raises concerns about search engines like Google, as it makes information easier for humans to organize. In their continued study, the authors asserted that the higher version of ChatGPT-4) effectively aids decision-making by providing a comprehensive overview of topics, but it does not replace the need for a business expert (Chumaet al., 2024).

In a case study, Budhwaret al. (2023) identified drawbacks of ChatGPT as a methodology assistant, including offering either too many, too few, or too general suggestions. Managers seeking precise, context-specific guidance may face decision paralysis, lack of understanding, and unspecific recommendations. A study by Ikeda (2024) revealed that ChatGPT impacts moral decision-making similarly to expert advice, though decisions driven by negative emotions are less affected. Though it has little impact on low-reward decisions, its influence goes beyond moral assessments. Trust in AI has no effect on this influence, but personal uncertainty about making appropriate decisions does.

The potential of utilizing ChatGPT in decision-making scenarios to shorten the amount of time needed for decision-making was investigated by Seita and Kurahashi (2024). According to the findings of a game-based experiment, a single user of ChatGPT could make decisions that were on par with comprehensive discussions by several users. Similarly, the use of ChatGPT to speed up decision-making scenarios was investigated in a study by Seita and Kurahashi (2024), which found that a single interaction can produce outcomes comparable to extended discussions by several people. An important finding of this study is that it proved a substantial decrease in decision-making costs is possible.

According to a study by Khanet al. (2024), the majority of interviewees think ChatGPT will speed up data collection, save time and effort, and enhance the quality of decisions. Despite ChatGPT’s potential advantages, most experts were concerned that it would harm impact assessment. Lawmakers should take these concerns into account when drafting regulations, guidelines, and laws pertaining to ChatGPT use.

Ensuring Fair and Unbiased HR Decisions with Generative AI Tools

ChatGPT, a decision-support tool in HRM, can enhance organizational performance, employee satisfaction, and business goal achievement by combining AI and human policies. However, several authors have discussed various suggestions for making fair and unbiased decisions when implementing AI tools like ChatGPT in HR. AI bias is a common problem that arises when biased data is used to train the system, which can result in unfair decision-making (Sebastian, 2023). Recent research suggests methods to reduce biases in data and algorithms, with the aim of establishing fair responses from AI. Regular updates to the AI models can help resolve bias issues and ensure fairness in decision-making. Primarily, organizations should invest in higher-performing models for objective problem-solving despite the trade-off in cost and performance. Business success can be supported by ChatGPT’s integration with knowledge management, which improves organizational decision-making through effective information retrieval, personalized learning, collaborative learning, real-time decision support, and continuous improvement (Sumbalet al., 2024). However, to ensure fairness, ChatGPT training data must be diverse and reflect various groups and perspectives. This will minimize the risk of prolonging bias and ensure that HR decisions are inclusive.

Before implementing ChatGPT in HRM, it is necessary to ensure the model is trained with relevant organizational data and undergoes comprehensive testing to validate its responses. The model should be continually refined and updated to align with the organization’s evolving needs and diversity goals. Ensuring the accuracy and quality of the data used to train the system is also critical in avoiding unintended biases. Even though ChatGPT is an AI tool, users may take advantage of its innocence as it learns through training. Therefore, providing accurate and proper training is essential for HR professionals who use the tool. This guarantees the responsible use of AI, preventing it from becoming an infallible decision-maker. Businesses should establish ethical guidelines for the use of ChatGPT to avoid its potential for negative impact on society, such as reinforcing biases or violating privacy. As per a study by Krügelet al. (2023), users’ judgment is influenced by ChatGPT’s inconsistent moral advice; therefore, it is imperative to enhance users’ digital literacy and assist them in comprehending the limitations of AI. One must be knowledgeable about when to ask AI for advice and when to disregard it. Moreover, providing employees with a comprehensive explanation of ChatGPT’s use for HR support is crucial. The explanation should encompass its benefits, drawbacks, and the scope of its application in different HR roles. It is essential to have a thorough discussion of the advantages, drawbacks, and concerns related to ChatGPT and other generative AI technologies to address policies regarding their use.

Continuously monitoring ChatGPT’s functionality and managing interactions with employees is crucial for companies to ensure accurate responses and compliance with business regulations. Due to its inability to verify information sources or conduct critical evaluations, an AI system may find data that isn’t accurate or plausible, which could reduce the information value. However, companies should identify these issues and fix them without delay (Jainet al., 2024). Staying updated with the latest advancements in AI technology, including ChatGPT models, will enhance the quality and effectiveness of using this tool for HR decision-making.

HR managers should consider human roles in decision-making processes and interpret ChatGPT results before making final decisions. However, moral dilemmas may arise when ChatGPT makes career or well-being decisions, highlighting the need for human oversight and accountability in AI-driven HRM processes (Sutrisno & Rijal, 2024). In situations with moral labels, ChatGPT-3’s decision-making differs greatly from human decision-making, which emphasizes the necessity of incorporating ethical theories and procedures in AI development (Rehmanet al., 2023). The ethical considerations of AI use will aid regulators and policymakers in making informed decisions regarding its use in moral regulations and decision-making processes.

Strict privacy policies are necessary to safeguard employee data accessed by ChatGPT, and AI algorithms should not be used for discriminatory or adverse purposes. Protecting the privacy of employees’ personal information and ensuring that AI-driven decisions do not overstep on their rights is essential for creating a fair HR environment. In any case, managers must develop strategies to manage the risks associated with decisions made by generative AI. AI decisions should not be blindly accepted because they might be based on inaccurate or biased information that could serve to further preconceived notions in areas like hiring and promotion (Aguiniset al., 2024). Maintaining human oversight throughout the decision-making process will mitigate the potential risks associated with AI-driven HR decisions.

Discussion

This study reviews existing research on ChatGPT and its support in decision-making within HRM. Similar studies have been conducted in various fields, such as medical advice (Liuet al., 2023), financial suggestions (Ahangar & Fietko, 2023), business and management (Korzynskiet al., 2023), supply chains (Haddud, 2024), tourism (Stergiou & Nella, 2024), and education (Montenegro-Ruedaet al., 2023). However, limited research has focused specifically on HR field. Some studies have addressed general HRM practices (Rane, 2023a, 2023b; 2024; Iswahyudiet al., 2023), but these have not explored the decision-making abilities of ChatGPT. This proposed review study, therefore, aims to fill this gap by focusing on the role of ChatGPT in HR decision-making, analyzing the challenges it poses, evaluating the quality of the responses provided, and offering insights for organizations to effectively utilize it.

This study carried out a systematic literature review in accordance with PRISMA guidelines in order to accomplish its objectives and present a thorough summary of findings. The review process comprised selecting keywords, locating studies, and evaluating their results. The studies were located in popular academic databases such as Scopus and Google Scholar. The literature review is based on 50 studies related to ChatGPT and its use in HRM decision-making. The bibliometric analysis indicates that the topic is gaining attention from researchers, with 40% of articles from 2023 and 60% from 2024 showing an increase in article production over the years. Furthermore, the types of studies indicate that the majority are published in journals. However, most of these articles focus on simple quantitative research, leaving a dearth of more detailed qualitative and evidence-based studies. The keyword analysis reveals that the terms ‘generative AI,’ ‘decision making,’ and ‘human resource management’ are frequently used, offering valuable insights into research trends and securing dominant topics.

To answer RQ1, a comprehensive analysis was carried out by analyzing the entire content of the studies to assess the strengths of ChatGPT in making decisions in HRM. Research revealed that ChatGPT enhances organizational performance and employee training by serving as a crucial element in HR decision-making (Rane, 2023b). Through task automation, enhanced communication, and individualized learning, it facilitates employee training and development, all of which support HRM (Sebastian, 2023). ChatGPT speeds up HR procedures, enhances decision-making, and increases employee engagement by supporting the hiring, performance management, and skill development (Rane, 2024). Its real-time insights and data analysis capabilities assist HR managers in making well-informed decisions regarding performance, retention, and talent management (Zhou & Cen, 2023). For future organizational success, ChatGPT’s potential to enhance productivity, communication, and strategic HR decisions remains significant.

As a response to RQ2, numerous studies address the drawbacks of ChatGPT while highlighting its advantages in HR decision-making. The positive relationship between ChatGPT and innovation (Ciminoet al., 2024) indicates that it can be helpful for creative managers; however, human input may be more beneficial for tasks requiring moral judgment and empathy (Adiasto, 2024). Organizational resistance to change, algorithmic bias, data privacy, and the balance between AI and human roles are among the challenges (Rane, 2023a). Problems with bias, transparency, and excessive dependence on AI can impact the quality of decisions and critical thinking (Chenet al., 2023; Sumbalet al., 2024). Thus, despite their promise, ChatGPT’s HR applications need constant management and ethical, well-balanced integration.

As a response to RQ3, the quality and speed of ChatGPT’s responses have been evaluated. ChatGPT has the potential to significantly enhance HR decision-making by providing HR professionals with timely and relevant insights, which enhances their ability to save time and effort and encourages them to make well-informed decisions. However, challenges associated with AI-driven decision-making—such as accuracy, sensitivity to context, and ethical implications—bring to light the importance of carefully integrating human expertise. Additionally, the answer to RQ4 has been assessed. Organizations need to place a strong emphasis on diverse data, continuous monitoring, ethical guidelines, and human oversight in order to guarantee that generative AI tools, such as ChatGPT, support decisions that are fair and unbiased in the field of human resources. Businesses can utilize AI by proactively addressing areas of concern, minimizing risks, and ensuring fairness and equity in HR practices.

While assessing the literature, the decision-making abilities should involve 16 aspects, which can be grouped under the main skills of analytical skills, strategic orientation, interpersonal and ethical skills, and adaptability and innovation, as shown in Fig. 4. However, ChatGPT partially supports decision-making through logical tasks, but lacks deep subjective, ethical, or strategic judgment skills. Table I lists abilities and limitations. As a result, human expertise continues to be indispensable for making higher-level decisions that require subjective judgment, creative thinking, or ethical concerns.

Fig. 4. Various aspects of decision-making abilities.

Table I. The Abilities and Limitations of ChatGPT with Decision Making Aspects

Decision-Making aspects	Capabilities	Limitations
Critical thinking	Conducts data analysis and evaluation using specific patterns.	Lacks independent opinions and complex ethical reasoning.
Problem-Solving	Suggests solutions using predefined logic or knowledge.	Struggles with new or unclear issues.
Data interpretation	Processes and displays data effectively.	Lacks interpreting incomplete data and inferring deeper meanings.
Risk assessment	Identifies risks from collected data.	Lacks understanding of human emotions, organisational context, and long-term effects.
Subjective judgment	Uses facts or patterns to provide insights.	Lacks context-based reasoning or human experience to make subjective judgements.
Strategic thinking	Uses historical data to make strategic recommendations.	Lacks strategic creativity and struggles to set long-term goals and align decisions.
Prioritization	Determines priorities from inputs.	Lacks emotional or organisational context needed to determine high-impact priorities.
Resource management	Offers resource-management solutions.	Ineffective at real-time resource allocation and dynamic constraint management.
Communication	Articulates concise ideas and decisions.	Lacks context-sensitivity and may miss tone or intent.
Ethical judgment	Ethics can be identified theoretically.	Lacks understanding and action on complex emotional, cultural, and circumstance-based ethical issues.
Emotional intelligence	Simulates emotion comprehension from the input.	Unable to empathise and understand emotions.
Responsibility	Make data-driven suggestions.	Lacks moral responsibility and decision-making.
Creativity	Helps generate new ideas from existing patterns.	Relies heavily on existing data, hindering innovation and creativity.
Flexibility	Adjust responses based on the provided information.	Lacks situational awareness and cannot adapt to fast-changing environments.
Learning from feedback	Adjust responses based on feedback.	Unable to learn naturally through personal growth or experiential learning.
Time management	Utilise data to suggest time management strategies.	Inability to manage time or adjust schedules or priorities in real-time.

Managerial Implications

This study has several implications for managers and policymakers who are integrating ChatGPT for decision-making in HRM. To influence managerial work at the strategic, functional, and administrative levels, it is necessary to study certain management theories and concepts in the context of generative AI (Korzynskiet al., 2023). Understanding the factors influencing generative AI adoption and customization can enhance innovation outcomes and decision-making processes by guiding better integration strategies (Ciminoet al., 2024). To ensure transparency and well-informed decision-making, it is essential to demand an explanation for the decisions made by AIs. AI systems employed in HRM should be open and transparent, offering concise justifications for hiring, promotion, and termination decisions to foster understanding and accountability. AI cannot replace human empathy and moral judgment, so it should not be relied upon entirely (Aguiniset al., 2024).

Here are some suggestions for enhancing ChatGPT’s potential in HR decision-making:

• Provide continuous training and updates with recent HR policies and organizational specifics.

• Ensure professionals are properly trained to understand ChatGPT’s limitations.

• Use diverse and representative training data and address biases in data appropriately.

• Use AI but encourage human HR professionals for final decisions.

• Develop clear ethical guidelines around the use of AI.

• Collect feedback from HR professionals to identify areas for improvement.

• Regularly monitor ChatGPT’s recommendations and ensure its suggestions are explainable.

These suggestions imply that companies can successfully incorporate ChatGPT into HR, utilizing its advantages while managing any risks, thus improving decision-making procedures in an ethical manner. To effectively utilize the advantages of generative AI, the proposed work advocates a hybrid HR decision-making framework, as shown in Fig. 5.

Fig. 5. Proposed hybrid HR decision-making framework with generative AI.

This decision-making framework has six steps, including problem identification, determination of possible solutions, data collection, evidence analysis, selection of the best option, and review of the decisions. The HR manager identifies HR concerns and provides detailed specifications, while generative AI streamlines the process while maintaining human oversight. The hybrid framework involves a human decision-maker and AI assistance, with grey numbers representing the detailed steps involved. HR managers must ensure proper AI training with diverse data and gather feedback for future improvements, allowing for a better understanding of AI’s advantages and limitations.

Conclusion

This study evaluates ChatGPT’s role in HR decision-making through a systematic literature review of 50 articles, highlighting increased interest but limited qualitative and evidence-based studies. Despite the lack of in-depth qualitative research, it is evident that generative AI, decision-making, and HRM are receiving increased attention. ChatGPT enhances organizational performance and employee training by automating tasks, improving communication, and enabling personalized learning, streamlining HR processes, aiding decision-making, and increasing employee engagement. However, ChatGPT, requiring human input for moral judgment, faces challenges like organizational resistance, algorithmic bias, and data privacy. Prioritizing diverse data, continuous monitoring, ethical guidelines, and human oversight ensures fair decisions. Therefore, the study proposes a hybrid HR decision-making framework incorporating generative AI to overcome these challenges.

Limitations and Future Directions

While the study provides a comprehensive view of ChatGPT’s role in decision-making, it has limitations. The review primarily explores ChatGPT’s role in HR decision-making, suggesting future studies should compare it with other generative AI tools like Google’s Bard and Microsoft’s Bing. The research’s narrow focus on HR decision-making overlooks other HRM aspects and business processes, suggesting the need for a broader exploration of HRM areas and processes. The limited database selection and English-language focus in the literature may lead to missing relevant studies, with limited empirical research in diverse, large-scale organizational settings. Future research should expand databases and conduct empirical studies to explore generative AI’s role in HR decision-making, enhancing ChatGPT’s understanding of human behavior and ethical HR decision-making. ChatGPT could also be combined with machine learning, sentiment analysis, and voice recognition to offer more personalized and user-centered experiences in HR.