Open research questions in AI and HR Technologies

What the literature leaves open

• 2 Future Research Directions Future research should focus on empirical validation of the proposed model using quantitative techniques such as Structural Equation Modelling (SEM) or SmartPLS to test relationships between AI capabilities and HR outcomes. Further research is necessary to examine ethical issues, including algorithmic discrimination, system openness, 73 Dash et al.

  AI-ENABLED TRANSFORMATION OF HR: A CONCEPTUAL REVIEW OF TRAINING AND TALENT MANAGEMENT · 2026 · DOI

• Yet, the risks they introduce, understood here as threats to organizational knowledge assets, such as the degradation of knowledge integrity, erosion of knowledge accessibility and compromises in knowledge security within HR processes, remain insufficiently understood.

  Risk mitigation in cloud-based computing in human resource systems: potential apprehensions and critical reflections in knowledge management · 2026 · DOI

• Future research could explore this dynamic more systematically, investi- gating whether a tradeoff exists between model accuracy and skill retention. Future studies could explore the effectiveness of incen- tivizing private, judgments to mitigate potential gaming behaviors and assess the practical implications of integrating this approach into real-world settings. One limitation naturally stems from the controlled exper- imental design. Even though our design allows us to isolate effects, it may not be representative of the development of decision-making across the entire range of the professional environment.

  The Dependency Dilemma: How Machine Learning Decision Aids can Undermine Skill Growth · 2026 · DOI

• This paper introduced a hybrid system for forming diverse and balanced corporate teams using social network analysis and genetic optimization. By optimizing skill diversity, workplace cohesion, and team size balance, the system provides a scalable, data-driven alternative to traditional team formation methods. Simulated organizational networks demonstrated the system’s robustness across various network sizes, including small, medium, and large. Future work will validate the system with real corporate data, integrate behavioral and personality attributes, and explore adaptive team reconfiguration as projects evolve. Expanding to enterprise HR platforms and addressing global, cross-cultural teams will further increase its relevance. These extensions position the system as a practical tool for building inclusive, high-performing teams in modern organizations. 8 ACKNOWLEDGMENTS This work was supported in part by the Commonwealth Cyber Initiative, an investment in the advancement of cyber R&D, innovation, and workforce development. For more information about CCI, visit https://cyberinitiative.org/. 9 REFERENCES B. Vedres and O. Vásárhelyi, “Inclusion unlocks the creative potential of gender diversity in teams,” 2022. https://doi.org/10.31235/osf.io/a3wf9 M. Barak and M. Usher, “The innovation level of engineering students’ team projects in hybrid and MOOC environments,” European Journal of Engineering Education, vol. 47, no. 2, pp. 299–313, 2021. https://doi.org/10.1080/03043797.2021.1920889 S. Hussein, M. Hasan, and M. Murtuza, “A team formation framework for managing diversity in multidisciplinary engineering project,” International Journal of Engineering Pedagogy (IJEP), vol. 7, no. 1, pp. 84–94, 2017. https://doi.org/10.3991/ijep.v7i1.6461 P. García-Sánchez, N. Díaz, and P. Pérez, “Social capital and knowledge sharing in academic research teams,” International Review of Administrative Sciences, vol. 85, no. 1, pp. 191–207, 2017. https://doi.org/10.1177/0020852316689140 M. Sahoo, N. Janardhanan, and S. Ekkirala, “Team ties, embeddedness, and turnover intentions: Integrating social networks and field theory,” Small Group Research, vol. 55, no. 3, pp. 417–456, 2023. https://doi.org/10.1177/10464964231195101 E. Bernstein and S. Turban, “The impact of the ‘open’ workspace on human collaboration,” Philosophical Transactions of the Royal Society B Biological Sciences, vol. 373, no. 1753, p. 20170239, 2018. https://doi.org/10.1098/rstb.2017.0239 S. Reddy et al., “How different forms of social capital created through project team assignments influence employee adoption of sustainability practices,” Organization & Environment, vol. 34, no. 1, pp. 43–73, 2019. https://doi.org/10.1177/1086026619880343 B. Rienties and N. Johan, “Getting the balance right in intercultural groups: A dynamic social network perspective,” Social Networking, vol. 3, no. 3, pp. 173–185, 2014. https://doi.org/10.4236/sn.2014.33022 32 International Journal of Advanced Corporate Learning (iJAC) iJAC | Vol. 19 No. 2 (2026) Using Artificial Intelligence and Social Network Analysis for Building Diverse and Balanced Corporate Teams T. Valente, L. Palinkas, S. Czaja, K. Chu, and C. Brown, “Social network analysis for program implementation,” PLoS ONE, vol. 10, no. 6, p. e0131712, 2015. https://doi.org/ 10.1371/journal.pone.0131712 P. Vahtera, P. Buckley, M. Aliyev, J. Clegg, and A. Cross, “Influence of social identity on negative perceptions in global virtual teams,” Journal of International Management, vol. 23, no. 4, pp. 367–381, 2017.

  Using Artificial Intelligence and Social Network Analysis for Building Diverse and Balanced Corporate Teams · 2026 · DOI

• What was the outcome? Hired, rejected, advanced, placed in different role What actually happened? 30-day performance, 90-day performance, 180-day retention, 1-year advancement This creates a feedback loop. When your AI says "this person will succeed," you track whether they actually do. When they don't, you go back and ask: What did the assessment miss? What did the hiring manager miss? What changed after hire? Organizations that are serious about compliance—and most enterprise organizations will be by 2027—are building this audit trail now. It's not complicated technically. But it requires intentional design. Component 3: Quality-of-Hire Measurement This is where most organizations fall completely flat. Quality of hire comprises four metrics: Performance (0-90 days): Does the hire demonstrate the capability we assessed? This is your accuracy check. If the assessment said "strong Python developer," is this person actually writing Python at the level we expected? Retention (1-year, 2-year, 3-year): Did they stay? This is a crude measure, but important. High-quality hires don't leave in the first year unless something's very wrong. If your AIselected candidates are leaving at twice the rate of manually-selected candidates, something's broken. Advancement (1-year internal mobility): Can they move into higher-responsibility roles? LinkedIn's research found that the companies making the best hires aren't just hiring competent people—they're hiring people who can grow into the next level. That advancement metric is the signal. If you hire someone strong and they're still doing the same work in 18 months, you've missed an opportunity. Cultural Integration: Do they strengthen or weaken team dynamics? This is hard to measure quantitatively, but it's crucial. Team feedback, engagement scores, and peer reviews matter. LinkedIn's framework specifically measures quality of hire as three components: (1) demand—was this candidate highly sought after by other companies? (2) retention—did they stay at least one year? and (3) mobility—did they advance internally within a year? These three metrics combined give you a signal of whether you actually hired well.¹² Here's what's wild: Most companies don't track any of this. They hire someone, they leave the hiring team's view, and nobody ever asks: Did this person work out? Should we hire more people like them or fewer? Component 4: Transition Intelligence Once you measure what works, you predict what comes next. Transition intelligence answers: Based on this person's capability profile and performance trajectory, what's their next best role? What training bridges the gap? If we upskill them in X, what's the probability of successful transition to Y? This is where hiring becomes workforce development. Instead of "fill this open role," you start thinking "hire people with potential, then develop them into the roles we'll need." This transforms the entire economics of hiring.

  "From Automation to Accountability: Why Measurement Infrastructure is the Next Evolutionary Step for AI-Driven Hiring" · 2026 · DOI

• Based on the findings, the following recommendations are proposed: 1. Invest in Upskilling with Focus: HR functions must launch aggressive upskilling programs focused on data literacy, statistical analysis, and business finance. This should be mandated for all senior HR professionals aspiring to advisory roles. 2. Rebrand HR's Value Proposition: HR leadership must proactively communicate its strategic contributions using data-driven business cases. It should move away from reporting HR metrics (e.g., time-to-fill) to presenting business metrics influenced by people strategy (e.g., revenue per employee, impact of engagement on productivity). 3. Develop an AI Ethics Charter: Organizations should form cross-functional committees (including HR, IT, Legal, and Ethics officers) to develop a formal charter for the ethical use of AI in people management. This charter must mandate transparency, algorithmic fairness audits, and human oversight. 4. Foster C-Suite Alignment: CHROs must work closely with CEOs and CFOs to align the HR strategy with the overall business strategy. This involves jointly defining what "strategic partnership" means and setting shared goals that demonstrate HR's impact on business outcomes. 5. Modernize HR Technology Infrastructure: Invest in integrated, AI-powered HCM (Human Capital Management) platforms that provide a single source of truth for people data and have strong predictive analytics capabilities.

  HR Business Advisers 2030: Shaping Strategy in the Age of AI · 2026 · DOI

• The article does not emphasize the specific characteristics of the businesses being studied, and the program development was not tied to local peculiarities. The study represents an initial analysis of the effectiveness and functionality of the proposed model based on the Microsoft Teams chat platform and contains a wide range of opportunities for further improvement and testing. The methodologies used are not yet suitable for scaling in their current form and require preliminary validation and relevance checks in relation to the needs and objectives of other organizations. The study also tested individual potential external factors impacting the results of the experiment, particularly chronic fatigue, etc., which primarily concerned the physical or mental state of the participants. Physiological indicators were collected with the expertise of an independent physician, and the author had no influence or authority to critique the physician's diagnosis. Economic, environmental, cultural, political, or other factors that could have affected interactions with clients, such as specific national or local events that could activate or deactivate interest in insurance among the population, were not taken into account. S.

  AI in Performance Management: AI as a Tool for Instant Assessments and Feedback in the Flow of Employee Transactions · 2026 · DOI

• The moderating role of technological factors highlights that when HRIS/ATS systems are fragmented or data is inconsistent, analytics insights become unreliable and the organisation cannot effectively diagnose or resolve delays.

  <b>The Effect of Human Resource Analytics on Reducing Time-to-Hire in Talent Acquisition: Evidence from FMCG Organisations in Zambia</b> · 2026 · DOI

• While Machine Learning (ML) techniques have demonstrated strong potential in forecasting turnover, empirical benchmarks tailored to specific sectors remain scarce, especially within developing regions.

  Enhancing workforce retention in the engineering sector: machine learning-driven turnover prediction models · 2026 · DOI

• IJDDT, Volume 16 Issue 38s, 2026 Page 606 Digital Transformation of Human Resource Management Practices to Enhance Workforce Performance in Organizations: A Quantitative Assessment • The results may not be applicable to all industries due to differences in organizational practices and work environments.

  Digital Transformation of Human Resource Management Practices to Enhance Workforce Performance in Organizations: A Quantitative Assessment · 2026 · DOI

• personalized AI-powered engagement programs based on the attitudes of an employee, creating a more inclusive and supportive environment within the workplace (Davenport et al., 2020). enable initiatives. For 3.4 Government Initiatives AI technology's adoption in HRM is also not just for private organizations, as this coincidence takes momentum with the government the government declared tremendous investments into AI R&D all through March 2024 as part of a strategic move towards technological pre-eminence and financial boost. Such initiatives are expected to spur AI adoption across sectors, including human resource management through innovation, capacity building and digital infrastructure. in India, instance, These policy-level interventions are imperative to ease the process of AI adoption, especially in emerging economies. Governments can strive to ensure that AI feeds inclusive growth and sustainable development by supporting research, facilitating public-private partnerships, and scaling up workforce reskilling. It fits a wider set of economic the significance of new technologies for improved productivity and labour market outcomes (Bresnahan etal, 2002). theories about AI applications in HRM, in short are numerous and maturing rapidly from recruitment to training, employee engagement to policy frameworks. These technologies not only streamline things operationally, but allow for a more strategic, data driven and human approaches to talent management. With a growing presence of AI, its integration into HRM is likely to become even more entrenched, revolutionising the way organisations attract, develop and retain their talent.

  Artificial Intelligence in Human Resource Management: A Study of Its Impact on Recruitment, Operational Efficiency, Ethical Practices, and Employee Experience · 2026 · DOI

• Future research should address limitations related to cross-sectional, single-source data by using longitudinal or multi-source designs to reduce common method bias and improve causal inference. Larger, probability-based samples and alternative statistical approaches (e.g., marker variables or latent method factors) are recommended to enhance robustness and generalizability. Additionally, testing the model across different sectors and contexts would help validate the stability of the findings. Additionally, integrating topic modeling or explainable AI (XAI) techniques can provide deeper insights into customer expectations and decision-making patterns. DECLARATION Ethical Consideration: This study strictly adhered to the Declaration of Helsinki and relevant national and institutional ethical guidelines. Informed consent was obtained. All procedures performed in this study were consistent with the ethical standards of the Helsinki Declaration.

  The Impact of Human Resource Practices on Employee Job Satisfaction in Pakistan’s Pharmaceutical Industry · 2026 · DOI

• fostering ethical management of also through its the literature by interpretation. The There are some limitations of this review article. First, the study is conceptual and review-based; it does not involve primary data collection, statistical analysis and empirical testing. The arguments are and synthesis elaborated conceptual interpretations should thus be viewed as theoretical and practical, and not as statistical findings. the study emphasises the Asian firms, Second, particularly startups, SMEs and growth-oriented firms. Cultural and historical differences between Asian economies, however, mean that the conditions of the labour markets, the maturity of HR, data protection laws and readiness for AI vary greatly across the region. Due to the differences, the usage of AI-powered HR analytics could differ between countries, industries and organizational size. Third, the review is not specific to particular sectors, but general in scope. The benefits, obstacles, and Doi: 10.66635/g08nn543 1176-8592 Vol. 22 No.3S (2026) May 258/260 By Dr. Annjaan Daash The Journal of Asia Entrepreneurship and Sustainability RESEARCH ARTICLE ethical implications that industries can encounter with the implementation of AI-driven HR systems can vary. Different industries might face varying challenges, benefits, risks when and ethical implementing AI-driven HR systems. Fourth, the study brings to the fore key ethical including algorithmic bias, employees' concerns, privacy, transparency, accountability and human oversight. It does not, however, empirically analyse the perceptions of employees, the preparedness of managers and the reactions of the organisation to these concerns. In conclusion, the implementation of AI in HR analytics might need more exploration for its practical use. Finally, AI technologies are continuously evolving, especially in the era of Generative AI and Automated Decision-Making Solutions. This review does not include new technologies developed since the last review, which may create additional opportunities and risks. Thus, these findings of this research can be applied with a note of caution that the world of artificial Intelligence, HR analytics and digital workforce management is ever- changing.

  <b>Navigating the New Era of Human Capital: The Strategic Convergence of AI, HR Analytics, and Modern Management Education in a Post-Pandemic Economy</b> · 2026 · DOI

• • • • The study was conducted on a limited sample size (154) within a specific geographic area, which may affect the generalizability of the results. Data was collected through self-reported questionnaires, which may be influenced by personal bias or misunderstanding. Employee experience differences among respondents (like tenure variation or dealership exposure) were not considered but could influence perceptions.

  Role of HR Practices in Enhancing Sales Performance · 2026 · DOI

• Ø Expanded AI adoption in recruitment, training, and performance management. Ø Established dedicated HR analytics units within banks. 544 © CINEFORUM CINEFORUM ISSN: 0009-7039 Vol. 66. No. 2, 2026 Ø Conduct regular employees training on AI system. Ø Ensure transparency in AI-based appraisal and promotion decision. Ø Develop hybrid models combining human judgement with AI insights. Ø Monitor Employees satisfactions during digital transition. Ø Allocate strategic budget for HR techonology modernization.

  Impact of Artificial intelligence assisted HR Practices on Employee performance and profitability of public sector bank in India · 2026 · DOI

• Although the present study provides preliminary evi- dence on the feasibility of using AI to assess engagement through natural language, it must be explicitly acknowl- edged that the small sample size (N = 35) represents a significant limitation in terms of external validity. Given that the sample was small and non-random, the results cannot be confidently generalized to other populations, sectors, or organizational contexts. While this type of limitation is common in exploratory and initial validation studies, it should be interpreted with caution, particularly when considering the applicability of the tool in broader or more diverse settings. For future research, it may be worth considering the development of ML models explic- itly based on latent structures that reflect the theoretical dimensions of engagement. This approach would allow models (such as RoBERTa) to be trained to directly esti- mate latent dimensions of engagement previously defined through factor analysis, enabling the rigorous applica- tion of tests such as AVE (Average Variance Extracted), HTMT (Heterotrait-Monotrait Ratio), and latent-level analyses. This integration between ML and psychomet- rics would support validation not only in terms of model accuracy, but also in terms of theoretical consistency and the ability to discriminate between related constructs. In addition, the inclusion of employees from the com- pany that developed the instrument (Erudit AI) may have introduced several potential sources of bias. Participants may have had greater familiarity with the tool, higher motivation to support the project, or increased aware- ness of the study objectives, which could have influenced both their self-report responses and their communica- tion behavior. Although anonymity procedures were implemented, this contextual factor should be considered when interpreting the strength of the observed associa- tions. Future research should prioritize fully independent samples to minimize potential expectancy or allegiance effects. On the other hand, the datasets used in the study con- tain texts that may be influenced by biases related to ide- ology, race, age, gender, and other factors. These biases can introduce distortions into the results of the analysis, affecting the objectivity and validity of the conclusions drawn from the textual data. Although quality and cod- ing criteria were applied, linguistic data may reflect cul- tural, ideological, or communicative patterns linked to variables such as age, gender, educational level, or role within the organization. This presents a significant risk, as the model could learn to associate certain expres- sion styles or linguistic structures with specific levels of engagement, systematically penalizing employees who do not conform to those patterns. As a result, the system could reproduce existing inequalities rather than provide a neutral assessment, affecting both the model’s valid- ity and the fairness of its practical applications. If unde- tected and uncorrected, this type of bias compromises the objectivity of the analysis and limits the generalizabil- ity of the results to more diverse work environments. Given that AI technology is evolving rapidly, the mod- els used in the study could become obsolete in a short García-Navarro and Pulido-Martos BMC Psychology (2026) 14:703 time. In particular, advances in NLP, such as the devel- opment of larger and more context-sensitive models (e.g., GPT, LLaMA, or future versions of BERT-based architec- tures) could significantly outperform current tools like RoBERTa in interpreting nuanced or context-dependent language. While these improvements may lead to more accurate engagement classifications, they also pose a challenge for the generalization and reproducibility of the findings in this study. If future models are trained on different datasets, use different tokenization strategies, or produce updated semantic representations, the same texts could yield different engagement scores. This would hinder the comparability of results over time and across systems, compromising the stability and standardization of AI-based psychological assessment methods. The study has moderate reliability, so it would be important to continue working on improving and modi- fying the instrument in order to achieve higher reliability in future studies. Finally, if the study uses specific AI models, it’s impor- tant to consider that the performance and applicability of these models may vary depending on the context and the specific task for which they were trained. The results may not be extrapolatable to other domains or tasks.

  Erudit AI SaaS: an artificial intelligence tool based on RoBERTa for classifying employee engagement · 2026 · DOI

• related to the frequency of administration, as these instruments are usually administered annually and this does not provide information about the current reality of employees. Although abbreviated versions have been created to solve the problem; their contribution has not been as expected, since, it has reduced the interest in the instruments by participants and also increased the time and economic cost of their administration.

  Erudit AI SaaS: an artificial intelligence tool based on RoBERTa for classifying employee engagement · 2026 · DOI

• This study has results, which means that should be kept in mind when reading its findings. First, the secondary data design cannot access the granular, contextspecific information that primary research would yield. The 51 included studies operationalize "organizational culture", "AI adoption", and "employee engagement" in precise different the comparisons difficult and specificity conclusions. Second, of publication bias is likely present: studies reporting positive effects of AI on to be engagement are more published than those reporting null or negative the evidence synthesis probably overstates the positive case. Third, despite India's growing presence in the dataset, the literature remains predominantly Englishlanguage and relatively concentrated in certain cultural contexts, which limits the generalizability of the thematic findings. Fourth, formal metaanalysis means that effect sizes cannot be of estimated, hypothesized be supported directionally but not quantified. Fifth, nearly all included studies are crosssectional, which means the dynamic, longitudinal character of the culture-AIengagement relationship remains poorly understood.

  Influence of Organizational Culture on Employee Engagement: Exploring the Mediating Role of Artificial Intelligence · 2026 · DOI

• AI strategy, human-AI collaboration, governance Upskilling, adaptability, AI literacy Regulation, reskilling SCIENTIFIC CULTURE, Vol. 12, No 2.1, (2026), pp. 7140-7160 7159 ARTIFICIAL INTELLIGENCE AND WORKFORCE PRODUCTIVITY programs, inclusive adoption 7.3 Future Research Directions The study has pointed out, in a very transparent manner, the different prospective areas for research these being: • Originally designed longitudinal researches to closely monitor AI adoption and meant productivity as they come to exist one after the other. • Causal analysis at the firm level that separates the impact of AI from other changes in the organization. • Comparisons across different sectors and institutional and countries to see how much regulatory differences influence. REFERENCES • Research focused on well-being and equity that will look at how the productivity driven by AI is distributed among the stakeholders. 7.4 Final Reflections in force time not a deterministic AI presents itself as a transformative but at the same the workplace. Its power to double up the output comes from the deliberate plan, human–AI partnership, and flexible organizational ecosystems. On the one hand, technology, along with the policies that support equal opportunities the researchers’ supervision will be the tools to carve out a more productive, satisfying, and fair work practice that will establish the future of labour in knowledge-driven economies. for all and Acemoglu, D., & Restrepo, P. (2020). Artificial intelligence and jobs. Journal of Economic Perspectives, 34(3), 30– 50. Aghion, P., Jones, B. F., & Jones, C. I. (2019). Artificial intelligence and economic growth (NBER Working Paper No. 23928). National Bureau of Economic Research. Autor, D. (2019). Work of the past, work of the future. AEA Papers and Proceedings, 109, 1–32. https://doi.org/10.1257/pandp.20191110 Autor, D. H. (2015). Why are there still so many jobs? Journal of Economic Perspectives, 29(3), 3–30. https://doi.org/10.1257/jep.29.3.3 Autor, D., Mindell, D. A., & Reynolds, E. B. (2022). The work of the future: Building better jobs in an age of intelligent machines. MIT Press. Bessen, J. E. (2019). AI and jobs: The role of demand (NBER Working Paper No. 24235). National Bureau of Economic Research. Brynjolfsson, E., & McAfee, A. (2014). The second machine age. W. W. Norton & Company. Brynjolfsson, E., & McAfee, A. (2017). Machine, platform, crowd: Harnessing our digital future. W. W. Norton & Company. Brynjolfsson, E., Li, D., & Raymond, L. R. (2023). Generative AI at work. Quarterly Journal of Economics, 139(1), 1–48. https://doi.org/10.48550/arXiv.2304.11771 Brynjolfsson, E., Rock, D., & Syverson, C. (2021). The productivity J-curve: How intangibles complement general-purpose technologies. American Economic Journal: Macroeconomics, 13(1), 333–372. Bughin, J., Seong, J., Manyika, J., Chui, M., & Joshi, R. (2018). Notes from the AI frontier: Modeling the impact of AI on the world economy. McKinsey Global Institute. Cowgill, B., & Tucker, C. E. (2020). Algorithmic fairness and economics. Columbia Business School Research Paper. Davenport, T. H., & Ronanki, R. (2018). Artificial intelligence for the real world. Harvard Business Review, 96(1), 108–116. Davis, F. D. (1989). Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Quarterly, 13(3), 319–340. https://doi.org/10.2307/249008 Dell’Acqua, F., et al. (2023). Navigating the jagged technological frontier: Field experimental evidence of the effects of AI on knowledge worker productivity. Harvard Business School Working Paper. Eloundou, T., Manning, S., Mishkin, P., & Rock, D. (2023). Gpts are gpts: An early look at the labor market impact potential of large language models. arXiv preprint arXiv:2303.10130, 10. Evelson, B. (2023). Augmented analytics. Forrester Research. Frank, M. R., Autor, D., Bessen, J. E., Brynjolfsson, E., Cebrian, M., Deming, D. J.,... & Rahwan, I. (2019). Toward understanding the impact of artificial intelligence on labor. Proceedings of the National Academy of Sciences, 116(14), 6531–6539. SCIENTIFIC CULTURE, Vol. 12, No 2.1, (2026), pp. 7140-7160 7160 KULJEET KAUR Frey, C. B., & Osborne, M. A. (2017). The future of employment: How susceptible are jobs to computerisation? 254–280.

  ARTIFICIAL INTELLIGENCE AND WORKFORCE PRODUCTIVITY: A COMPREHENSIVE ANALYSIS OF TRANSFORMATION, OPPORTUNITIES, AND CHALLENGES IN THE MODERN WORKPLACE · 2026 · DOI

• The paper identifies that managerial relationships (Years with Current Manager) contribute significantly to attrition predictions, but does not investigate whether manager-specific characteristics (leadership style, retention rate of their team, tenure in role) would improve predictive performance beyond tenure-based features in the employee attrition model.

  Predicting Employee Attrition: A Machine Learning Approach in Human Resource Analytics · 2026 · DOI

• While the paper notes that Overtime emerges as a major predictor in Gradient Boosting (rank 3) for attrition, reflecting workload effects, it does not empirically measure the threshold at which overtime hours transition from being a retention factor to a significant attrition driver, or examine sector-specific variation in this relationship.

  Predicting Employee Attrition: A Machine Learning Approach in Human Resource Analytics · 2026 · DOI

• The feature importance analysis identifies Monthly Income, Total Working Years, and Age as dominant predictors across both Gradient Boosting and Random Forest models, but does not investigate interaction effects or temporal dynamics between these variables (e.g., whether income trajectories relative to peer compensation better predict attrition than absolute salary levels).

  Predicting Employee Attrition: A Machine Learning Approach in Human Resource Analytics · 2026 · DOI

• The paper demonstrates that Logistic Regression outperforms ensemble methods on the structured IBM HR Analytics Attrition Dataset, but does not validate whether this finding generalizes to organizations with unstructured HR data, different industry sectors, or datasets with varying class imbalance ratios beyond the current imbalanced attrition scenario.

  Predicting Employee Attrition: A Machine Learning Approach in Human Resource Analytics · 2026 · DOI

• Decision Tree performance showed poor generalization (AUC of 0.54) compared to Logistic Regression (AUC 0.82), attributed to overfitting, but the paper does not systematically investigate whether ensemble method advantages (Random Forest, Gradient Boosting) would emerge when applied to complex or high-dimensional HR datasets beyond the IBM structured dataset used in this study.

  Predicting Employee Attrition: A Machine Learning Approach in Human Resource Analytics · 2026 · DOI

• The feature importance analysis reveals divergent rankings between Gradient Boosting and Random Forest models for secondary attrition predictors, with Gradient Boosting prioritizing workload variables (Overtime, Stock Option Level) while Random Forest emphasizes demographic factors (Distance from Home, Years at Company). The paper does not investigate whether these differences stem from model architecture bias or represent genuine contextual variations in attrition mechanisms across employee subgroups.

  Predicting Employee Attrition: A Machine Learning Approach in Human Resource Analytics · 2026 · DOI

Most-cited papers in AI and HR Technologies

• Human Resource Management and Labor Productivity: Does Industry Matter? · Academy of Management Journal · 2005 · 963 citations
• MTurk Research: Review and Recommendations · Journal of Management · 2020 · 757 citations
• A multilevel review of artificial intelligence in organizations: Implications for organizational behavior research and practice · Journal of Organizational Behavior · 2023 · 528 citations
• The Janus face of artificial intelligence feedback: Deployment versus disclosure effects on employee performance · Strategic Management Journal · 2021 · 435 citations
• IMPROVING LABOR PRODUCTIVITY: HUMAN RESOURCE MANAGEMENT POLICIES DO MATTER · Strategic Management Journal · 1996 · 361 citations
• Artificial Intelligence Coaches for Sales Agents: Caveats and Solutions · Journal of Marketing · 2020 · 296 citations
• A review of machine learning applications in human resource management · International Journal of Productivity and Performance Management · 2021 · 263 citations
• Substituting Human Decision-Making with Machine Learning: Implications for Organizational Learning · Academy of Management Review · 2020 · 244 citations
• Employees recruitment: A prescriptive analytics approach via machine learning and mathematical programming · Decision Support Systems · 2020 · 242 citations
• Complexities and Controversies in Linking HRM with Organizational Outcomes · Journal of Management Studies · 2001 · 228 citations

Most recent work

• Artificial intelligence usage at work: understanding dual self-regulation mechanisms and goal orientation in career growth · Career Development International · 2026
• Erudit AI SaaS: an artificial intelligence tool based on RoBERTa for classifying employee engagement · BMC Psychology · 2026
• Company Benefits, Perks, and Their Advantages for Employees: A PRISMA Review · Financial Metrics in Business · 2026
• The implementation of artificial intelligence in organizations by functional areas: A review and conceptual model · Journal of Management & Organization · 2026
• AI in Performance Management: AI as a Tool for Instant Assessments and Feedback in the Flow of Employee Transactions · Public Organization Review · 2026
• Imagining AI at work: the impact of polarized imaginaries on AI use in human resources management · Journal of Workplace Learning · 2026
• The Effects of Artificial Intelligence Applications on Effective Management Processes in Organizations: A Literature Review · Uluslararası Yönetim Akademisi Dergisi · 2026
• Hierarchical dilated parrot causal convolutional networks for optimizing human resource recommendations · Quality & Quantity · 2026
• AI-Powered HR: Transforming Performance Management Through Skills-Based Hiring · Scientific Societal & Behavioral Research Journal · 2026
• Cognitive HRM: Harnessing Generative AI for Emotionally Intelligent and Inclusive Workforce Ecosystems · Journal of Artificial Intelligence General science (JAIGS) ISSN:3006-4023 · 2026

Related topics in Business, Management and Accounting