Explainable AI in Labor Market Applications

Gabriel Bicharra Santini Pinto, Carlos Eduardo Mello and Ana Cristina Bicharra Garcia

Federal University of Rio de Janeiro (UNIRIO), PPGI, Av. Pasteur, 458, Urca, RJ, Brazil

gabrielbspinto@edu.unirio.br, {mello, cristina.bicharra}@uniriotec.br

Keywords:

Explainable Artiﬁcial Intelligence, Labor Market, Human Resource Systems, Fairness.

Abstract:

The adoption of artiﬁcial intelligence (AI) applications has been accelerating in the labor market, driving

productivity gains, scalability, and efﬁciency in human resource management. This progress has also raised

concerns about AI’s negative impacts, such as ﬂawed decisions, biases, and inaccurate recommendations. In

this context, explainable AI (XAI) plays a crucial role in enhancing users’ understanding, satisfaction, and

trust. This systematic review provides a segmented overview of explainability methods applied in the labor

market. A total of 266 eligible studies were identiﬁed during the search and evaluation process, with 29 studies

selected for in-depth analysis. The review highlights the different explainability requirements expressed by

users of human resource systems. Additionally, it identiﬁes the processes, tasks, and corresponding explain-

ability methods implemented.

1 INTRODUCTION

Artiﬁcial Intelligence (AI)-enabled applications are

increasingly automating and enhancing decision-

making processes in human resource management

within organizations, becoming a core component of

corporate investment strategies (Chowdhury et al.,

2023). The adoption of AI in the labor market is

driven by its potential to boost efﬁciency, minimize

human errors, and forecast future behaviors through

data pattern analysis (Lukacik et al., 2022).

As AI adoption progresses in the labor market, im-

portant issues must be addressed to ensure its ethical

and fair application (Chowdhury et al., 2023). Con-

cerns about AI’s negative impacts, such as ﬂawed

decisions, biases, and inaccurate recommendations,

gain prominence as these technologies become em-

bedded in organizational routines. Despite its poten-

tial to reduce biases, documented discriminatory inci-

dents continue to raise concerns (Tambe et al., 2019).

Simultaneously, human workers face fears and skep-

ticism about potential job losses due to automation

(Webb, 2019).

In this context, Explainable AI (XAI) is essen-

tial to ensure that decisions made by ”black-box”

models are understandable to users (Bertrand et al.,

2022). XAI helps identify and correct biases in train-

ing data, preventing unfair or ﬂawed decisions — a

critical factor for maintaining user trust and the ef-

fectiveness of AI applications across various domains

(Fidel et al., 2020). Clear and comprehensible expla-

nations strengthen the acceptance and credibility of

AI systems from both user and regulatory compliance

perspectives, ensuring that decisions are grounded in

logical and justiﬁable reasoning (Ali et al., 2023a).

The article is organized as follows: Section 2

presents the theoretical framework, providing con-

text for the research problem. Section 3 describes

the methods and procedures employed in conducting

the systematic literature review (SLR), including a de-

tailed explanation of the search terms and criteria used

for selecting relevant studies. Section 4 outlines the

results obtained from applying the search and selec-

tion protocol. In Section 5, the evaluation of qual-

ity criteria for the selected studies is discussed, along

with answers to the research questions. Finally, Sec-

tion 6 summarizes the study’s conclusions, highlights

the work conducted, and offers suggestions for future

research directions.

2 THEORETICAL BACKGROUND

This section presents the literature review addressing

topics inﬂuencing the development of explainability

methods in the labor market. The section is organized

as follows. First, topics related to the application

of artiﬁcial intelligence in the labor market are dis-

cussed. Next, explainability methods for artiﬁcial in-

telligence are presented. Finally, related works found

1450

Pinto, G. B. S., Mello, C. E. and Garcia, A. C. B.

Explainable AI in Labor Market Applications.

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 17th International Conference on Agents and Artiﬁcial Intelligence (ICAART 2025) - Volume 3, pages 1450-1457

ISBN: 978-989-758-737-5; ISSN: 2184-433X

in the literature are reviewed.

2.1 Artiﬁcial Intelligence in the Job

Market

The application of Artiﬁcial Intelligence (AI) in la-

bor market processes has been transforming the job

market while simultaneously provoking an important

ethical debate about its applications. In recruitment,

AI technologies have improved processes across four

different stages: job advertisement, screening, eval-

uation, and facilitation (Black and van Esch, 2020).

The beneﬁts include a larger audience for job postings

with better candidate-job ﬁt (Bogen, 2019), analysis

of large volumes of resumes (Lukacik et al., 2022)

with automatic pattern identiﬁcation based on pre-

deﬁned criteria (Vasconcelos et al., 2018), and in-

creased accuracy (Roemmich et al., 2023). In em-

ployee management, machine learning models are

used to prevent employee attrition, improving reten-

tion strategies and reducing turnover costs (Al-Alawi

and Ghanem, 2024). At the same time, automatic

and personalized skill recommendations make profes-

sional development more informed and suited to ca-

reer paths (B

ohm et al., 2024).

Human resource management is recognized for

high levels of bias and prejudice as they are devel-

oped and manipulated according to human perspec-

tives (Olckers and Walsh, 2024). Consequently, ma-

chine learning models trained on biased employee

databases can perpetuate and magnify discriminatory

practices, privacy issues, and transparency concerns

as they achieve scalability. Ethical application con-

cerns, decision-making biases, discriminatory inci-

dents, and potential job losses due to automation

(Chowdhury et al., 2023) are driving the demand for

explainability in these systems.

Enhancing the interpretability of automated de-

cisions promotes greater trust and fairness, ensuring

decisions are based on accurate and justiﬁable data

while supporting more efﬁcient and equitable hir-

ing, evaluation, and professional development prac-

tices. In ﬁelds where explainability is critical, such

as healthcare and ﬁnance, systematic reviews have

examined XAI methods and the sociotechnical re-

quirements for their applications (

Cernevi

cien

e and

Kaba

sinskas, 2024; Ali et al., 2023b).

2.2 Explainable Artiﬁcial Intelligence

(XAI)

Explainable Artiﬁcial Intelligence (XAI) encom-

passes various methods aimed at making machine

learning models more interpretable. A central

approach is feature relevance explanation, exem-

pliﬁed by Shapley Additive Explanations (SHAP),

which quantiﬁes each feature’s contribution to a

model’s prediction (Lundberg, 2017). While effec-

tive, SHAP’s reliance on individual feature contri-

butions can overlook feature dependencies. Simpli-

ﬁcation methods like decision trees and rule-based

learners approximate complex models through sim-

pler representations, aiding interpretability without

sacriﬁcing performance. Local explanations focus on

speciﬁc instances, with Local Interpretable Model-

Agnostic Explanations (LIME) offering insights by

locally approximating predictions through perturbed

inputs (Ribeiro et al., 2016). Counterfactual expla-

nations further support understanding by highlighting

how different inputs could change outcomes, offering

actionable insights (Johansson et al., 2016).

Transparent models like linear regression and de-

cision trees inherently allow interpretability by ex-

posing how input features inﬂuence predictions (Ar-

rieta et al., 2020). Finally, XAI frameworks integrate

multiple methods to enhance overall explainability.

These frameworks combine feature relevance, local

and global explanations, and visualization techniques,

creating comprehensive interpretation systems that

improve decision-making transparency and user trust

(Linardatos et al., 2020).

2.3 Related Works

(Bujold et al., 2024) highlight the multidisciplinary

nature of AI in Human Resource Management

(HRM), noting that many studies rely on experimental

frameworks without real-world validation. Similarly,

(Trinh and Elbanna, 2023) observe that AI research

in HRM remains fragmented across management dis-

ciplines, limiting comprehensive knowledge build-

ing. In recruitment and selection, (Rigotti and Fosch-

Villaronga, 2024) stress the need for fairness-driven

system design balancing ethical, legal, and technical

aspects. They recommend cross-disciplinary collab-

oration and participatory research to address power

asymmetries between applicants and HR practition-

ers.

3 METHODS AND PROCEDURE

Conducting a Systematic Literature Review (SLR) is

a well-established approach in Software Engineering

for comprehensively surveying existing research on

a speciﬁc topic following the protocol proposed by

(Kitchenham et al., 2015). The following sections

detail this process: Section 3.1 deﬁnes the research

Explainable AI in Labor Market Applications

1451

questions; Section 3.2 outlines the search strategy;

Section 3.3 presents the screening process.

3.1 Research Questions

The primary objective of this review is to provide

valuable insights into models, frameworks, and ap-

proaches of explainable AI applied in the labor mar-

ket, emphasizing key topics, design challenges, and

socio-technical requirements for their effective imple-

mentation. To achieve this objective, the following

research questions were formulated:

• RQ1: What are the socio-technical requirements

and the design, implementation, and evaluation

challenges of explainable artiﬁcial intelligence

models?

• RQ2: Which labor market AI applications have

beneﬁted from the implementation of XAI?

• RQ3: What are the novel XAI approaches that

have been developed and implemented?

• RQ4: How can machine learning and AI-based

frameworks be designed to promote fairness and

reduce bias in automated recruitment and job eval-

uation systems?

3.2 Search Strategy

The search strategy began with a review of the most

relevant systematic literature on explainable artiﬁ-

cial intelligence (XAI) and its pioneering systematic

reviews in speciﬁc domains such as healthcare and

credit scoring. Following this, an initial search was

conducted on Google Scholar using cross-referencing

terms related to the labor market and XAI. This pre-

liminary exploration provided a clearer understanding

of the topic, enabling the deﬁnition of search terms as

outlined in Table 1.

3.3 Screening Process

Searches were conducted across four major

databases: ACM Digital Library, IEEE Xplore,

ScienceDirect, and Springer Link, focusing on arti-

cles classiﬁed under the Computer Science domain.

The review timeframe covers publications from 2019

to 2024, emphasizing recent advancements in the

ﬁeld. The initial search retrieved 261 papers. Ad-

ditionally, ﬁve manually selected articles identiﬁed

during the search strategy development phase were

included. During the title screening process, 199

articles were excluded based on predeﬁned selection

criteria. Subsequently, 67 papers were reviewed

by analyzing their abstracts and assessing their

Table 1: Search Query.

Population

”Employee Veriﬁcation” OR ”Attrition” OR ”Career

Development” OR ”Employability” OR ”Employee

Retention” OR ”Future of Work” OR ”Human Re-

source Management” OR ”Human Resources” OR

”Job Market” OR ”Labor Market”

Intervention

”Explainable AI” OR ”AI Transparency” OR ”Al-

gorithmic Transparency” OR ”Bias Mitigation” OR

”Explainability” OR ”Explainable Artiﬁcial Intelli-

gence” OR ”Fairness in AI” OR ”Interpretability” OR

”LIME” OR ”Model Interpretation” OR ”Post-Hoc

Explanation” OR ”SHAP” OR ”Shapley Additive Ex-

planations” OR ”XAI”

Outcome

”Hiring Process” OR ”Automatic Recruitment” OR

”Career Recommendation Systems” OR ”Job Match-

ing Algorithms” OR ”Performance Evaluation” OR

”Recruitment Automation” OR ”Talent Acquisition”

OR ”Workforce Management” OR ”Workplace Au-

tomation”

Table 2: Inclusion and Exclusion Criteria.

Inclusion Criteria

• Studies with identiﬁed black-box models

• Studies addressing research on artiﬁcial intelligence

responsibility in the job market

• Studies presenting methods to build explainability

Exclusion Criteria

• Papers that do not answer the research questions

• Studies outside the job market application domain

• Studies not written in English

• Studies that do not present a clear methodology

• Studies that do not investigate explainable artiﬁcial

intelligence

availability through institutional access. Ultimately,

29 articles were included in this systematic review, as

illustrated in Figure 1.

Figure 1: Study Selection Flow Diagram.

The quality of the articles was assessed based on

the following criteria, as established by (Kitchenham

et al., 2015): Does the study present clear and unam-

biguous ﬁndings? Is the context of the study clearly

ICAART 2025 - 17th International Conference on Agents and Artiﬁcial Intelligence

1452

described? Are the research objectives explicitly de-

ﬁned?

The extracted data were stored using the Parsifal

software. Data extraction followed 24 unique codes

and was analyzed based on the theoretical framework

and the research questions outlined in this study.

4 RESULTS

As a result of applying the proposed methods and pro-

cedures, this systematic review included 29 articles

published between 2019 and 2024 in the ﬁeld of com-

puter science. Given that this is the ﬁrst SLR focused

on explainability in the labor market, the number of

identiﬁed articles provides a solid foundation for con-

ducting meaningful analyses on the topic. It is im-

portant to note that although the search terms were

speciﬁc to the labor market, many of these terms are

commonly used across various application domains,

leading to relevant articles from other ﬁelds appear-

ing in the search results.

5 DISCUSSION

The analysis was structured as follows: articles ad-

dressing the ﬁrst research question include surveys

with XAI users and contributions based on other

qualitative research methodologies. The responses

to research questions 2, 3, and 4 encompass articles

presenting explainability methods validated through

well-deﬁned methodologies.

5.1 What Are the Socio-Technical

Requirements and the Design,

Implementation, and Evaluation

Challenges of Explainable Artiﬁcial

Intelligence Models?

Based on the analysis of the studies in this research,

three key requirements were identiﬁed to promote

transparency and explainability in artiﬁcial intelli-

gence systems applied to the labor market:

Bargaining: AI systems in the labor market can

beneﬁt if they address employees’ desire to express,

negotiate, and contextualize outcomes, thereby bal-

ancing power asymmetries caused by informational

imbalances with employers. In studies on systems de-

signed to digitally monitor employees during task ex-

ecution (Das Swain et al., 2023), employees agreed to

share generated inferences as long as proper negotia-

tion and contextualization occurred. While employ-

ees do not perceive signiﬁcant differences between

human managers and AI systems regarding objec-

tive aspects such as salaries and bonuses, they expect

greater empathy and emotional support from human

managers in areas like career development (Tomprou

and Lee, 2022). Unmet expectations in these subjec-

tive areas result in greater disappointment when hu-

man interaction is involved, highlighting the impor-

tance of transparent communication in AI-mediated

workplaces.

Job candidates, on the other hand, showed a strong

preference for traditional face-to-face interviews over

those conducted by AI systems (Lashkari and Cheng,

2023; Girona and Yarger, 2024), reﬂecting the im-

portance of contextual interaction in the candidate

selection process. This need for contextualization

also manifests in candidates’ ability to express their

identities and share their lived experiences and as-

pirations autonomously during interviews (Aizenberg

et al., 2023).

Human Resources Biases: AI systems in the la-

bor market can beneﬁt if they identify and address

historical, conscious, or unconscious biases present

in human resources processes. Corporate data re-

lated to recruitment or performance evaluation often

reﬂect historical decisions inﬂuenced by human bi-

ases. AI tools can help reduce such biases by ap-

plying objective and adjustable criteria, improving

job requirement deﬁnitions, structuring interviews,

and conducting skill-based assessments (Lashkari and

Cheng, 2023). Explainability, in this context, helps

both specialists and employees identify errors and bi-

ases in AI outcomes (Park et al., 2022).

However, addressing biases requires particular at-

tention to underrepresented groups. For instance, can-

didates with disabilities face challenges that expose

the need for more equitable and context-sensitive ap-

proaches (Tilmes, 2022). Traditional bias mitigation

strategies often overlook critical differences, under-

scoring the importance of centering candidates’ lived

experiences in achieving algorithmic fairness.

Labor Market Fairness: AI systems in the

labor market can beneﬁt if they incorporate fair-

ness concepts tailored to the context of human re-

sources processes. Issues such as biased predic-

tions, reliance on rigid quantiﬁcation, and limitations

in identifying groups in computational systems have

faced widespread criticism (S

anchez-Monedero et al.,

2020). Using attributes like gender or race to im-

prove predictive performance raises ethical and tech-

nical dilemmas, as well as underlying power dynam-

ics that risk perpetuating structural inequalities unless

participatory approaches are implemented (Park et al.,

Explainable AI in Labor Market Applications

1453

2022).

Deﬁning and measuring algorithmic fairness

poses signiﬁcant challenges. Professionals working

with responsible AI require organizational support to

drive changes that enable ethical practices (Rakova

et al., 2021). The ethical implementation of AI sys-

tems in recruitment is not inherently unethical; the

associated risks stem from inﬂated expectations and

the indiscriminate use of automated recruitment tools.

An ethical approach demands a clear deﬁnition of

what constitutes a fair AI system (Hunkenschroer and

Kriebitz, 2023). In this regard, counterfactual fair-

ness allows for greater ﬂexibility in decision-making

by considering speciﬁc conditions and varied contexts

(Hauer et al., 2021).

These aspects highlight the importance of AI sys-

tems that prioritize transparency, bias mitigation, and

equity in their applications within the labor market.

5.2 Which Labor Market AI

Applications Have Beneﬁted from

the Implementation of XAI?

The reviewed articles propose explainability mod-

els applied to three key processes in the labor mar-

ket: candidate selection and recruitment, professional

skill recommendation, and employee performance

evaluation. Each of these automated processes ad-

dresses speciﬁc tasks inherent to its domain within

the labor market. In the context of automated re-

cruitment, identiﬁed tasks include candidate proﬁle

evaluation (Bhattacharya et al., 2022; Sogancioglu

et al., 2023), interview assessment (Marra and Ku-

biak, 2024; Koutsoumpis et al., 2024), candidate-job

ﬁt prediction (Pessach et al., 2020), resume parsing

(Pe

na et al., 2023), fraud detection in job advertise-

ments (Naud

e et al., 2023), and recruitment demand

forecasting (Zhang et al., 2021).

In the area of professional skill recommendation,

two studies presented methods (Akkasi, 2024; Tran

et al., 2021) designed to extract relevant skills from

job descriptions while integrating user feedback to

enhance recommendation systems through reinforce-

ment learning algorithms (Sun et al., 2021; Sun et al.,

2024). Similarly, the automation of employee eval-

uation processes has been applied to measure job

performance (Sampath et al., 2024; Maurya et al.,

2018) and predict potential employee attrition (Das

et al., 2022; Makanga et al., 2024; Sekaran and Shan-

mugam, 2022), contributing to data-driven human re-

source management strategies.

5.3 What Are the Novel XAI

Approaches That Have Been

Developed and Implemented?

Among the 17 different methods for explainability

presented, three stood out for their transparent ap-

proaches aimed at enhancing model interpretability.

In the ﬁrst approach, explainability was achieved

by applying the Variable-Order Bayesian Network

(VOBN) model to recruitment data (Pessach et al.,

2020). The second method (Tran et al., 2021) uti-

lized counterfactual reasoning to infer the causal ef-

fects of various factors on employment status, rec-

ommending the most effective interventions accord-

ingly. Additionally, the third approach (Maurya et al.,

2018) involved the development of a stylized log-

linear model designed to uncover hidden aspects and

sentiment within an employee peer review corpus.

Another method employed was the use of feature

importance to evaluate the inﬂuence of individual fea-

tures on the model’s outcome, representing a signif-

icant contribution to the ﬁeld. Aiming to identify

fraudulent ads in a comprehensible manner, the most

important features in (Naud

e et al., 2023) were ex-

tracted from the best-performing rule-set-based clas-

siﬁer, the Random Forest with POS tags. Meanwhile,

(Marra and Kubiak, 2024) introduced a novel frame-

work designed to ﬁlter user data, detect gender bias,

and implement a feature importance block to enhance

explainability.

Six studies employed SHAP to provide explain-

ability. Two of these focused on candidate proﬁle

evaluation, using SHAP exclusively (Bhattacharya

et al., 2022; Sogancioglu et al., 2023). Addition-

ally, three studies on employee attrition prediction

(Das et al., 2022; Makanga et al., 2024; Sekaran and

Shanmugam, 2022) and one study on employee per-

formance assessment (Sampath et al., 2024) applied

SHAP in combination with the LIME method. An-

other study, centered on recommending professional

skills, relied solely on the LIME method (Akkasi,

2024).

(Sun et al., 2024) proposed a Self-Explaining

Skill Recommendation framework that identiﬁes and

prototypes prevalent market skill sets into represen-

tative exemplars to support decision-making. This

approach quantitatively decomposes the long-term

learning utility of talents into contributions from each

exemplar. It represents an extended and revised ver-

sion of the authors’ previous study, published three

years earlier (Sun et al., 2021). (Zhang et al.,

2021) implemented a data-driven neural sequential

approach called the Talent Demand Attention Net-

work (TDAN), designed to forecast ﬁne-grained tal-

ICAART 2025 - 17th International Conference on Agents and Artiﬁcial Intelligence

1454

ent demand in the recruitment market.

5.4 How Can Machine Learning and

AI-Based Frameworks Be Designed

to Promote Fairness and Reduce

Bias in Automated Recruitment and

Job Evaluation Systems?

(Pe

na et al., 2023) proposed a methodology demon-

strating how to create fairer AI-based tools, partic-

ularly in automated recruitment systems. Similarly,

(Marra and Kubiak, 2024) introduced a framework

leveraging neural networks as an effective strategy for

enhancing fairness in hiring practices with minimal

predictive accuracy loss. Additionally, (Tran et al.,

2021) integrated a machine learning-based framework

to effectively recommend skill upgrades for disabled

job seekers, identifying both the speciﬁc skills to im-

prove and the optimal upgrade level to maximize their

employment potential.

In a study on the automatic evaluation of job in-

terviews, (Koutsoumpis et al., 2024) examined bias

issues and emphasized the need for developing ma-

chine learning methodologies speciﬁcally designed to

address these challenges.

6 CONCLUSIONS

This study provided a segmented view of explain-

ability methods in the labor market and analyzed the

sociotechnical requirements highlighted by users. A

systematic review was conducted using search strate-

gies applied to major computer science databases.

The search process identiﬁed 266 eligible studies,

from which 29 papers were selected after evaluation,

offering a detailed perspective on explainability and

fairness in the labor market.

AI systems in the labor market can beneﬁt from

integrating three fundamental requirements: bargain-

ing, which enables employees and candidates to ex-

press their perspectives, negotiate outcomes, and con-

textualize decisions to balance informational asym-

metries; bias mitigation, which is crucial for identi-

fying and addressing historical biases in human re-

sources processes, fostering greater inclusion and eq-

uity; and fairness, which demands context-speciﬁc

approaches to address ethical and structural dilem-

mas, avoiding the perpetuation of inequalities. These

elements are essential for developing transparent, eth-

ical, and trustworthy AI systems that not only opti-

mize processes but also promote fairer and more col-

laborative workplace relationships. Three main labor

market processes emerged as more advanced in terms

of explainability development: recruitment, profes-

sional skill recommendation, and employee perfor-

mance evaluation. Within these processes, speciﬁc

tasks involving explainability methods were identi-

ﬁed.

From the perspective of applied explainability

methods, the selected papers included studies where

explainability was integrated by design. Common

explainability techniques and more advanced frame-

works delivering post-hoc explainability were also

frequently used. Notably, none of the papers evalu-

ated user satisfaction.

Overall, the ﬁndings clariﬁed key issues, high-

lighted explainability needs, and mapped methods ad-

dressing different labor market processes and tasks.

However, there is still considerable room for progress,

especially when viewed through the lens of explain-

ability literature. Since the labor market is character-

ized by information asymmetry, balancing explana-

tions could enhance the effective adoption of AI sys-

tems by placing users at the center of explainability

method development.

XAI has been advancing across various sectors,

yet signiﬁcant opportunities remain for its develop-

ment in the labor market. Future research could ex-

plore user perspectives by assessing satisfaction, trust,

and understanding. These measures are essential for

validating proposed methods and fostering greater

adoption of AI systems. Simultaneously, a critical gap

in the literature persists regarding how reducing in-

formational imbalances among different users could

enhance explainability in artiﬁcial intelligence.

ACKNOWLEDGEMENTS

This study was ﬁnanced in part by the Coordenac¸

de Aperfeic¸oamento de Pessoal de N

ıvel Superior –

Brasil (CAPES) – Finance Code 001. Part of the text

revision and reﬁnement was conducted with the assis-

tance of ChatGPT, an AI-based language model de-

veloped by OpenAI, to enhance clarity and coherence.

REFERENCES

Aizenberg, E., Dennis, M. J., and van den Hoven, J. (2023).

Examining the assumptions of ai hiring assessments

and their impact on job seekers’ autonomy over self-

representation. AI & society, pages 1–9.

Akkasi, A. (2024). Job description parsing with explain-

able transformer based ensemble models to extract the

technical and non-technical skills. Natural Language

Processing Journal, 9:100102.

Explainable AI in Labor Market Applications

1455

Al-Alawi, A. I. and Ghanem, Y. A. (2024). Predicting em-

ployee attrition using machine learning: A systematic

literature review. In 2024 ASU International Confer-

ence in Emerging Technologies for Sustainability and

Intelligent Systems (ICETSIS), pages 526–530. IEEE.

Ali, S., Abuhmed, T., El-Sappagh, S., Muhammad, K.,

Alonso-Moral, J. M., Confalonieri, R., Guidotti, R.,

Del Ser, J., D

ıaz-Rodr

ıguez, N., and Herrera, F.

(2023a). Explainable artiﬁcial intelligence (xai):

What we know and what is left to attain trustworthy

artiﬁcial intelligence. Information fusion, 99:101805.

Ali, S., Akhlaq, F., Imran, A. S., Kastrati, Z., Daudpota,

S. M., and Moosa, M. (2023b). The enlightening

role of explainable artiﬁcial intelligence in medical

& healthcare domains: A systematic literature review.

Computers in Biology and Medicine, page 107555.

Arrieta, A. B., D

ıaz-Rodr

ıguez, N., Del Ser, J., Bennetot,

A., Tabik, S., Barbado, A., Garc

ıa, S., Gil-L

opez, S.,

Molina, D., Benjamins, R., et al. (2020). Explainable

artiﬁcial intelligence (xai): Concepts, taxonomies, op-

portunities and challenges toward responsible ai. In-

formation fusion, 58:82–115.

Bertrand, A., Belloum, R., Eagan, J. R., and Maxwell,

W. (2022). How cognitive biases affect xai-assisted

decision-making: A systematic review. In Proceed-

ings of the 2022 AAAI/ACM Conference on AI, Ethics,

and Society, pages 78–91.

Bhattacharya, P., Zuhair, M., Roy, D., Prasad, V. K., and

Savaliya, D. (2022). Aajeevika: Trusted explainable

ai based recruitment scheme in smart organizations. In

2022 5th International Conference on Contemporary

Computing and Informatics (IC3I), pages 1002–1008.

IEEE.

Black, J. S. and van Esch, P. (2020). Ai-enabled recruiting:

What is it and how should a manager use it? Business

Horizons, 63(2):215–226.

Bogen, M. (2019). All the ways hiring algorithms can in-

troduce bias. Harvard Business Review, 6:2019.

ohm, M. J., von Gaudecker, H.-M., and Schran, F. (2024).

Occupation growth, skill prices, and wage inequality.

Journal of Labor Economics, 42(1):201–243.

Bujold, A., Roberge-Maltais, I., Parent-Rocheleau, X.,

Boasen, J., S

ecal, S., and L

eger, P.-M. (2024). Re-

sponsible artiﬁcial intelligence in human resources

management: a review of the empirical literature. AI

and Ethics, 4(4):1185–1200.

Cernevi

cien

e, J. and Kaba

sinskas, A. (2024). Explain-

able artiﬁcial intelligence (xai) in ﬁnance: A system-

atic literature review. Artiﬁcial Intelligence Review,

57(8):216.

Chowdhury, S., Dey, P., Joel-Edgar, S., Bhattacharya, S.,

Rodriguez-Espindola, O., Abadie, A., and Truong, L.

(2023). Unlocking the value of artiﬁcial intelligence

in human resource management through ai capabil-

ity framework. Human resource management review,

33(1):100899.

Das, S., Chakraborty, S., Sajjan, G., Majumder, S., Dey,

N., and Tavares, J. M. R. (2022). Explainable ai for

predictive analytics on employee attrition. In Inter-

national Conference on Soft Computing and its Engi-

neering Applications, pages 147–157. Springer.

Das Swain, V., Gao, L., Wood, W. A., Matli, S. C., Abowd,

G. D., and De Choudhury, M. (2023). Algorithmic

power or punishment: Information worker perspec-

tives on passive sensing enabled ai phenotyping of

performance and wellbeing. In Proceedings of the

2023 CHI Conference on Human Factors in Comput-

ing Systems, pages 1–17.

Fidel, G., Bitton, R., and Shabtai, A. (2020). When explain-

ability meets adversarial learning: Detecting adversar-

ial examples using shap signatures. In 2020 interna-

tional joint conference on neural networks (IJCNN),

pages 1–8. IEEE.

Girona, A. E. and Yarger, L. (2024). To impress an algo-

rithm: Minoritized applicants’ perceptions of fairness

in ai hiring systems. In International Conference on

Information, pages 43–61. Springer.

Hauer, M. P., Kevekordes, J., and Haeri, M. A. (2021).

Legal perspective on possible fairness measures–a le-

gal discussion using the example of hiring decisions.

Computer Law & Security Review, 42:105583.

Hunkenschroer, A. L. and Kriebitz, A. (2023). Is ai recruit-

ing (un) ethical? a human rights perspective on the use

of ai for hiring. AI and Ethics, 3(1):199–213.

Johansson, F., Shalit, U., and Sontag, D. (2016). Learn-

ing representations for counterfactual inference. In

International conference on machine learning, pages

3020–3029. PMLR.

Kitchenham, B. A., Budgen, D., and Brereton, P. (2015).

Evidence-based software engineering and systematic

reviews, volume 4. CRC press.

Koutsoumpis, A., Ghassemi, S., Oostrom, J. K., Holtrop,

D., Van Breda, W., Zhang, T., and de Vries, R. E.

(2024). Beyond traditional interviews: Psychomet-

ric analysis of asynchronous video interviews for per-

sonality and interview performance evaluation using

machine learning. Computers in Human Behavior,

154:108128.

Lashkari, M. and Cheng, J. (2023). “ﬁnding the magic

sauce”: Exploring perspectives of recruiters and job

seekers on recruitment bias and automated tools. In

Proceedings of the 2023 CHI Conference on Human

Factors in Computing Systems, pages 1–16.

Linardatos, P., Papastefanopoulos, V., and Kotsiantis, S.

(2020). Explainable ai: A review of machine learn-

ing interpretability methods. Entropy, 23(1):18.

Lukacik, E.-R., Bourdage, J. S., and Roulin, N. (2022). Into

the void: A conceptual model and research agenda for

the design and use of asynchronous video interviews.

Human Resource Management Review, 32(1):100789.

Lundberg, S. (2017). A uniﬁed approach to interpreting

model predictions. arXiv preprint arXiv:1705.07874.

Makanga, C., Mukwaba, D., Agaba, C. L., Murindanyi, S.,

Joseph, T., Hellen, N., and Marvin, G. (2024). Ex-

plainable machine learning and graph neural network

approaches for predicting employee attrition. In Pro-

ceedings of the 2024 Sixteenth International Confer-

ence on Contemporary Computing, pages 243–255.

Marra, T. and Kubiak, E. (2024). Addressing diversity in

hiring procedures: a generative adversarial network

approach. AI and Ethics, pages 1–25.

ICAART 2025 - 17th International Conference on Agents and Artiﬁcial Intelligence

1456

Maurya, A., Akoglu, L., Krishnan, R., and Bay, D. (2018).

A lens into employee peer reviews via sentiment-

aspect modeling. In 2018 IEEE/ACM International

Conference on Advances in Social Networks Analysis

and Mining (ASONAM), pages 670–677. IEEE.

Naud

e, M., Adebayo, K. J., and Nanda, R. (2023). A ma-

chine learning approach to detecting fraudulent job

types. AI & SOCIETY, 38(2):1013–1024.

Olckers, M. and Walsh, T. (2024). Manipulation and peer

mechanisms: A survey. Artiﬁcial Intelligence, page

104196.

Park, H., Ahn, D., Hosanagar, K., and Lee, J. (2022). De-

signing fair ai in human resource management: Un-

derstanding tensions surrounding algorithmic evalu-

ation and envisioning stakeholder-centered solutions.

In Proceedings of the 2022 CHI conference on human

factors in computing systems, pages 1–22.

na, A., Serna, I., Morales, A., Fierrez, J., Ortega, A., Her-

rarte, A., Alcantara, M., and Ortega-Garcia, J. (2023).

Human-centric multimodal machine learning: Recent

advances and testbed on ai-based recruitment. SN

Computer Science, 4(5):434.

Pessach, D., Singer, G., Avrahami, D., Ben-Gal, H. C.,

Shmueli, E., and Ben-Gal, I. (2020). Employees re-

cruitment: A prescriptive analytics approach via ma-

chine learning and mathematical programming. Deci-

sion Support Systems, 134:113290.

Rakova, B., Yang, J., Cramer, H., and Chowdhury, R.

(2021). Where responsible ai meets reality: Practi-

tioner perspectives on enablers for shifting organiza-

tional practices. Proceedings of the ACM on Human-

Computer Interaction, 5(CSCW1):1–23.

Ribeiro, M. T., Singh, S., and Guestrin, C. (2016). ” why

should i trust you?” explaining the predictions of any

classiﬁer. In Proceedings of the 22nd ACM SIGKDD

international conference on knowledge discovery and

data mining, pages 1135–1144.

Rigotti, C. and Fosch-Villaronga, E. (2024). Fairness, ai

& recruitment. Computer Law & Security Review,

53:105966.

Roemmich, K., Rosenberg, T., Fan, S., and Andalibi, N.

(2023). Values in emotion artiﬁcial intelligence hiring

services: Technosolutions to organizational problems.

Proceedings of the ACM on Human-Computer Inter-

action, 7(CSCW1):1–28.

Sampath, K., Devi, K., Ambuli, T., and Venkatesan, S.

(2024). Ai-powered employee performance evalua-

tion systems in hr management. In 2024 7th Interna-

tional Conference on Circuit Power and Computing

Technologies (ICCPCT), volume 1, pages 703–708.

IEEE.

anchez-Monedero, J., Dencik, L., and Edwards, L. (2020).

What does it mean to’solve’the problem of discrimina-

tion in hiring? social, technical and legal perspectives

from the uk on automated hiring systems. In Proceed-

ings of the 2020 conference on fairness, accountabil-

ity, and transparency, pages 458–468.

Sekaran, K. and Shanmugam, S. (2022). Interpreting the

factors of employee attrition using explainable ai.

In 2022 International Conference on Decision Aid

Sciences and Applications (DASA), pages 932–936.

IEEE.

Sogancioglu, G., Kaya, H., and Salah, A. A. (2023). Using

explainability for bias mitigation: A case study for fair

recruitment assessment. In Proceedings of the 25th

International Conference on Multimodal Interaction,

pages 631–639.

Sun, Y., Ji, Y., Zhu, H., Zhuang, F., He, Q., and Xiong, H.

(2024). Market-aware long-term job skill recommen-

dation with explainable deep reinforcement learning.

ACM Transactions on Information Systems.

Sun, Y., Zhuang, F., Zhu, H., He, Q., and Xiong, H. (2021).

Cost-effective and interpretable job skill recommen-

dation with deep reinforcement learning. In Proceed-

ings of the Web Conference 2021, pages 3827–3838.

Tambe, P., Cappelli, P., and Yakubovich, V. (2019). Ar-

tiﬁcial intelligence in human resources management:

Challenges and a path forward. California Manage-

ment Review, 61(4):15–42.

Tilmes, N. (2022). Disability, fairness, and algorithmic bias

in ai recruitment. Ethics and Information Technology,

24(2):21.

Tomprou, M. and Lee, M. K. (2022). Employment rela-

tionships in algorithmic management: A psychologi-

cal contract perspective. Computers in Human Behav-

ior, 126:106997.

Tran, H. X., Le, T. D., Li, J., Liu, L., Liu, J., Zhao, Y.,

and Waters, T. (2021). Recommending the most ef-

fective intervention to improve employment for job

seekers with disability. In Proceedings of the 27th

ACM SIGKDD Conference on Knowledge Discovery

& Data Mining, pages 3616–3626.

Trinh, N. and Elbanna, A. (2023). Ai and human relation-

ship in the workplace: A literature review and future

research agenda. In International Working Confer-

ence on Transfer and Diffusion of IT, pages 144–156.

Springer.

Vasconcelos, M., Cardonha, C., and Gonc¸alves, B. (2018).

Modeling epistemological principles for bias mitiga-

tion in ai systems: an illustration in hiring decisions.

In Proceedings of the 2018 AAAI/ACM Conference on

AI, Ethics, and Society, pages 323–329.

Webb, M. (2019). The impact of artiﬁcial intelligence on

the labor market. Available at SSRN 3482150.

Zhang, Q., Zhu, H., Sun, Y., Liu, H., Zhuang, F., and Xiong,

H. (2021). Talent demand forecasting with attentive

neural sequential model. In Proceedings of the 27th

ACM SIGKDD Conference on Knowledge Discovery

& Data Mining, pages 3906–3916.

Explainable AI in Labor Market Applications

1457