Faculty of Engineering & Information Technology School of Information, Systems and Modelling

AI-based Recruitment

Adoption, Implications, and Source of Competitive Advantage.

BY ANTON KRAVETS

Introduction

Business is a constantly changing environment, adapting to external factors. As a result of recent radical changes in lifestyle and the way people work, novel technologies have been introduced to the international business arena, such as Artificial Intelligence (AI) (Minbaeva, 2021). In this report, the phenomenon of AI will be considered by its’ implications in Recruitment and Selection (R&S) to figure out whether the current state of AI-based recruitment can be a competitive advantage for companies. The paper will investigate an academic foundation in Digitalization, the role of modern recruitment, AI and Machine Learning, Implications of AI in recruitment, Recommendations for HR practitioners.

One of the essential goals of this report is to identify knowledge gaps that should be discovered at the intersection between the aforementioned areas. Thus, AI-powered tools, stakeholders, issues will be revealed, based on which a topic for further research will be determined. The report will advance knowledge of the practical implications of AI in HR and its problems and discover effective adoption approaches for companies.

Literature Review

The primary sources of technological changes and comprehensive integration of AI-powered tools in HR practices result from the external forces that affect the working environment. Derous and Fruyt (2016) emphasized that Recruitment and Selection (R&S) has a strategic role for the organization during the current so-called “war” for talent where technological advances will increase the efficiency of recruiting, particularly in the assessment process. Minbaeva (2021, p.1), in a discussion on “Disruption in HR, ” considered integrating new technologies into HR practices as a construct of three mega-trends as Digitalization of business models, AI, and Machine Learning, and a flexible workforce

1. The Digitalization of Business Models. Disruption in HR practices

Minbaeva (2021) emphasized that COVID-19 has been identified as a significant cause and a global trend that changes the way people work today, leading to a digitalized and flexible workforce. Amonkwah-Amoah et al. (2021, p.601) considered this mega-trend as a positive source for a great introduction of new technologies that transform the lifestyle. Amankwah-Amoah et al. (2021) aimed to determine the phenomenon of Digitalization, analyze its’ opportunities, challenges, and contribute to the academic foundation of COVID-19 impact on the business environment. This discussion can be a key to understanding why HR practices are becoming heavily based on digital tools. While considering the role of Digitalization, Ritter and Peresen (2020) found that the main reason for the development and implementation of AI is that business practices are experiencing a data-driven disruption, which allows access to a large amount of data and efficiently analyze it through powerful computing capabilities.

2. Role of Modern Recruitment

In the late 20th century, the most crucial success factor were tangible assets: technological equipment, plants, and other properties. However, in the 2000s, there was a shift in the source of success towards the quality of human resources, which have become the primary source of competitive advantage (Black & Esch, 2020).

Breaugh (2013) pointed out the importance of research in the field of “External Recruitment” (employer recruits workers from outside of the organization) and determined the primary intention of such approach: draw candidates’ attention to the job ad, influence them to apply, maintain interest in position until the offer is accepted. Due to the nature of recruitment, there are always two interacting parties involved in this process: the employer (company) and the employee (candidate) (Breaugh, 2013). An important factor has been identified by Breaugh (2013) related to employees’ psychological consideration of recruitment. In this vein, during the process of recruitment, employees form an impression of the organization and its’ working environment. This aspect could become crucial when evaluating the implications of AI-powered recruitment on candidates’ perception of the organization.

Companies are also looking for a way to reduce biases in their approaches to recruitment (prejudices and beliefs); however, relying solely on AI in decision-making, the threat of discrimination and unfairness becomes more pronounced (Kochling & Wehner, 2020). One of the cases supporting this threat occurred in 2018 when Amazon’s machine-learning specialists discovered that their AI-powered recruiting engine was shut down due to the constant disadvantaging of women candidates (Dastin, 2018). Kochling and Wehner (2020) revealed undeveloped areas of unfairness and discrimination that accompany the Ai-based approach. To understand the source of potential issues, Kochling and Wehner (2020) carried out a detailed analysis of the operational principles of AI-based recruiting systems. They found that AI aims to collect, process, and act (make a decision) on data.

This is where the key to understanding biases lies because algorithms rely on historical data as an example for future decisions. Therefore, AI accepts human behaviors and decisions accurately and translates them into its’ decision-making algorithm, including stereotypes, biases, and judgments. Thus, if the criteria were taken with pre-existing tendencies to discrimination, then the decision would be the same. However, it is worth stating that AI is a neutral mechanism from the beginning. If the AI-learning stage is based on the neutral data, which does not have discrimination/biases inclinations, the AI will then make a decision without such flaws

3. AI-Based Recruitment as a Competitive Advantage

According to IndustryARC (2022), the AI in Recruitment has an estimated value of $580 million in 2019, projecting approximately 7% growth for the period 2020-2025. Market research reported an emerging need for AI-powered recruiting systems such as process automation, chatbots and screening tools.

The current AI-focused research domains are Machine learning (ML), Natural Language Processing (NLP), Image and Speech Recognition (Kochling & Wehner, 2020; Pillai & Sivathanu, 2020). There are three AI types of intelligence applied to the HR sector: Assisted, Augmented, and Autonomous (Charlier & Kloppenburg, 2017).

Assisted AI aims to standardize repetitive tasks, e.g., via chatbots. Augmented intelligence synthesizes man and computer in a single work unit to establish an immersive experience for the candidates. Autonomous intelligence is a top-level transformation that takes care of all the work of selecting a candidate based on the criteria. Algorithms of the last type are mostly considered “black boxes”, in which the HR department sets the requirements for the outcome, for instance, personality traits, skills, turnover tendencies, and receives the result without understanding how the machine made the calculations and on what logic.

Albert (2019) identified 11 AI tools currently applied in Recruitment & Selection. However, only three of them are usually adopted by organizations: chatbots, task automation admins, and CV screening software. Thus, Albert (2019) stated that despite such favorable opportunities and benefits of AI, there is still a low level of adaptation of AI tools and capabilities.

According to recent reports, finding a suitable candidate costs on average $4,000 and takes 42 days for companies (Bika, 2021). Kochling and Wehner (2020)  found that standardizing routine work decisions through an automated decision-making approach helps discover talent and process a high flow of applications. AI is undoubtedly a competitive advantage; it increases recruiting efficiency (reduces tomorrow, time, risks, and increases the accuracy of decisions). With the right approach to implementation, AI potentially increases companies’ profitability by 30%, which is why more than 60% of business continuedly to integrate AI-based tools across their HR departments (Albert 2019). Today, more and more companies seek opportunities to gain a competitive advantage through utilizing AI-based recruitment systems. This trend is supported by academics, who aim to develop algorithms and methodologies; organizations can further use that.

4 Existed AI tools, Algorithms, and Models for Enterprises.

Since most of the implications for AI are related to analyzing and comprehension of text, Natural Language Processing (NLP) is primarily used in Talent Acquisition tasks (Ivashenko & Milutkin 2019). Qin et al. (2018) highlighted Convolutional Neural Network (CNN) and Recurrent Neural Network (RNN) as the two mainly used architectures to solve NLP problems. In experimental research, they developed a “Person-Job-Fit” solution, which would serve as a bridge connecting candidate and employer.

The” Person-Job-Fit” model aims to reduce recruiters’ routine work and improve matching results by providing better candidate-job fit results. The model represents a Neural Network based on information extracted from the text and relates to Natural Language Processing (NLP) method that aims to classify, find a similarity, and comprehend the text.

It is worth structurally analyzing the “Person-Job-Fit” approach developed by Qin et al. (2018) to establish the underlying model of how AI-based HR departments will acquire talent in the future. Researchers were the first to propose developing such a solution as the construction of three representations: Word-Level, Hierarchical Ability-Aware, and Person-Job Fit. Such formation allows utilizing a Recurrent Neural Network (RNN) at the Word-Level representation to extrapolate words collected from job posting onto vectorized format through a Word2Vec algorithm.

Each representation measures the overall performance of a “Person-Job-Fit” model. The key phrases related to the essential skills become highlighted at the Word Level against an array of common words. Ability-Level helps categorize abilities/skills with their importance measured by scarcity (e.g. bachelor degree has the lowest value, because it is considered an essential requirement, although data analysis skill would be more beneficial since only a few candidates can master this skill). At the Matching Level, where the system will understand whether the job requirements and candidate experience can be paired, the new end-to-end Ability-Aware Person-Job Fit Neural Network (APJFNN) showed promising results in interpreting key matching clues. It has been proved that APJFNN indeed reduces the need for manual labour of recruiters while improving the quality of matching candidate-employer.

1Ilustration of the Person-Job-Fit model of the network algorithm made by Qin et al. (2018).

Progressing in developing AI-based recruitment and utilizing Qin et al. (2018) findings, Nigam and Saxena (2019) proposed an advanced methodology for talent acquisition utilizing a synthesis of Machine Learning and sub-recommendations to produce job recommendations for a job-seeker. In their view, Machine Learning has the potential to capture hidden (subtle) motivates of the job-seekers during their interaction with the job post.

Recommender System proposed by Nigam & Saxena (2019)

As it can be seen, both scientist groups used a Word2Vec model to vectorize words. Jang et al. (2020) considered this model (2020) through Word2Vec as a popular Deep Learning Model that converts natural language into the representation of vectors. Word2Vec, as a “sequence embedding” method, can find conceptual connections between words in a multidimensional space. This method is mainly used to establish a first approach in information capturing tasks for further predictions/decisions, as done by Qin et.al (2018) during the Word-Level Representation stage.

Recommendations

Considering the novelty of the technology, academics aim to provide industrial-applicable recommendations for various stakeholders in terms of the approaches to adaptation of AI tools. Black and Esch (2021) indicated the importance of managerial success in AI-based recruitment adoption. The main recommendations for managers from Black and Esch (2021) are:

  • Consider AI as a tool to reduce employees’ switching costs
  • AI-enabled recruiting tools are beneficial and should be employed by managers and expanded and explored’ implications beyond screening, interviewing, and assessing.

Albert (2019) framed his explorative research with recommendations for leading groups of practitioners:

  • For HR managers regarding how to maneuver in a changing technological landscape.
  • For Vendors, how to be competitive.
  • For Recruitment agencies to keep in mind, their existence is not stable since AI-vendors produce more efficient solutions in the industry.
  • For Candidates to take advantage of AI in R&S, especially by disadvantaged groups.
  • For Entrepreneurs to take advantage of unexplored niches.

Discussion

The environment of the HR industry is going through a rapid technological change due to the introduction of AI into the tactical, strategic, and operational tasks of organizations. This paper aims to provide a critical review of the established literature on AI and Talent Acquisition intersection. It has been found that the leading cause of the ongoing disruption in the HR industry is Digitalization, which keeps emerging under the influence of global factors, such as COVID-19. AI can be a tool to combat discrimination if the engine is trained on good quality data. Research agrees that AI-based recruitment tools can be a competitive advantage for organizations if the right adoption approach is employed.

The significant knowledge gap and recommendations for further research have also been identified. It is suggested to expand research to more sectors and study the impact of AI-based recruitment on organizational performance, employer reputation, and satisfaction of HR management. The geographic components should also be taken into account for further research. A critical gap is that the current focus of the study was mainly on the employer/organization, losing sight of the candidate’s perspective in the TA.

References

Albert. (2019). AI in talent acquisition: a review of AI-applications used in recruitment and selection. Strategic HR Review, 18(5), 215–221. https://doi.org/10.1108/SHR-04-2019-0024

Amankwah-Amoah, Khan, Z., Wood, G., & Knight, G. (2021). COVID-19 and Digitalization: The great acceleration. Journal of Business Research, 136, 602–611. https://doi.org/10.1016/j.jbusres.2021.08.011

Artificial Intelligence in HR: A no-brainer – PWC. (n.d.). Retrieved March 29, 2022, from https://www.pwc.nl/nl/assets/documents/artificial-intelligence-in-hr-a-no-brainer.pdf

Artificial Intelligence in recruitment market research report: Market size, industry outlook, Market Forecast, demand analysis,market share, market report 2021-2026. IndustryARC. (n.d.). Retrieved March 29, 2022, from https://www.industryarc.com/Report/19231/artificial-intelligence-in-recruitment-market.htm006C

Bika Contributor, N., & Contributor. (2021, November 19). Recruiting costs FAQ: Budget and cost per hire. Recruiting Resources: How to Recruit and Hire Better. Retrieved March 29, 2022, from https://resources.workable.com/tutorial/faq-recruitment-budget-metrics

Black, & van Esch, P. (2020). AI-enabled recruiting: What is it, and how should a manager use it? Business Horizons, 63(2), 215–226. https://doi.org/10.1016/j.bushor.2019.12.001

 

Black, & van Esch, P. (2021). AI-enabled recruiting in the war for talent. Business Horizons, 64(4), 513–524. https://doi.org/10.1016/j.bushor.2021.02.015

BREAUGH. (2013). Employee Recruitment. Annual Review of Psychology, 64(1), 389–416. https://doi.org/10.1146/annurev-psych-113011-143757

 

Dastin, J. (2018, October 10). Amazon scraps secret AI recruiting tool that showed bias against women. Reuters. Retrieved March 29, 2022, from https://www.reuters.com/article/us-amazon-com-jobs-automation-insight-idUSKCN1MK08G

 

Derous, & De Fruyt, F. (2016). Developments in Recruitment and Selection Research. International Journal of Selection and Assessment, 24(1), 1–3. https://doi.org/10.1111/ijsa.12123

 

Ivaschenko, & Milutkin, M. (2019). HR Decision-Making Support Based on Natural Language Processing. In Creativity in Intelligent Technologies and Data Science (pp. 152–161). Springer International Publishing. https://doi.org/10.1007/978-3-030-29743-5_12

Jang, Kim, M., Harerimana, G., Kang, S., & Kim, J. W. (2020). Bi-LSTM Model to Increase Accuracy in Text Classification: Combining Word2vec CNN and Attention Mechanism. Applied Sciences, 10(17), 1–14. https://doi.org/10.3390/app10175841

Köchling, & Wehner, M. C. (2020). Discriminated by an algorithm: a systematic review of discrimination and fairness by algorithmic decision-making in the context of HR recruitment and HR development. Business Research (Göttingen), 13(3), 795–848. https://doi.org/10.1007/s40685-020-00134-w

Minbaeva. (2021). Disrupted HR? Human Resource Management Review, 31(4), 1–8. https://doi.org/10.1016/j.hrmr.2020.100820 

Nigam, A., Roy, A., Saxena, A., & Singh, H. (2020, June 2). Job recommendation through progression of job selection. arXiv.org. Retrieved March 29, 2022, from https://arxiv.org/abs/1905.13136

Pillai, & Sivathanu, B. (2020). Adoption of artificial intelligence (AI) for talent acquisition in IT/ITeS organizations. Benchmarking : an International Journal, 27(9), 2599–2629. https://doi.org/10.1108/BIJ-04-2020-0186

Qin, Zhu, H., Xu, T., Zhu, C., Ma, C., Chen, E., & Xiong, H. (2020). An Enhanced Neural Network Approach to Person-Job Fit in Talent Recruitment. ACM Transactions on Information Systems, 38(2), 1–33. https://doi.org/10.1145/3376927

Ritter, & Pedersen, C. L. (2020). Digitization capability and the Digitalization of business models in business-to-business firms: Past, present, and future. Industrial Marketing Management, 86, 180–190. https://doi.org/10.1016/j.indmarman.2019.11.019