AI and public contests: a model to improve the evaluation and selection of public contest candidates in the Police Force

Abstract

Purpose The number of candidates applying to public contests (PC) is increasing compared to the number of human resources employees required for selecting them for the Police Force (PF). This work intends to perceive how those public institutions can evaluate and select their candidates efficiently during the different phases of the recruitment process. To achieve this purpose, artificial intelligence (AI) was studied. This paper aims to focus on analysing the AI technologies most used and appropriate to the PF as a complementary recruitment strategy of the National Criminal Investigation police agency of Portugal – Polícia Judiciária. Design/methodology/approach Using design science research as a methodological approach, the authors suggest a theoretical framework in pair with the segmentation of the candidates and comprehend the most important facts facing public institutions regarding the usage of AI technologies to make decisions about evaluating and selecting candidates. Following the preferred reporting items for systematic reviews and meta-analyses methodology guidelines, a systematic literature review and meta-analyses method was adopted to identify how the usage and exploitation of transparent AI positively impact the recruitment process of a public institution, resulting in an analysis of 34 papers between 2017 and 2021. Findings Results suggest that the conceptual pairing of evaluation and selection problems of candidates who apply to PC with applicable AI technology such as K-means, hierarchical clustering, artificial neural network and convolutional neural network algorithms can support the recruitment process and could help reduce the workload in the entire process while maintaining the standard of responsibility. The combination of AI and human decision-making is a fair, objective and unbiased process emphasising a decision-making process free of nepotism and favouritism when carefully developed. Innovative and modern as a category, group the statements that emphasise the innovative and contemporary nature of the process. Research limitations/implications There are two main limitations in this study that should be considered. Firstly, the difficulty regarding the timetable, privacy and legal issues associated with public institutions. Secondly, a small group of experts served as the validation group for the new framework. Individual semi-structured interviews were conducted to alleviate this constraint. They provide additional insights into an interviewee’s opinions and beliefs. Social implications Ensure that the system is fair, transparent and facilitates their application process. Originality/value The main contribution is the AI-based theoretical framework, applicable within the analysis of literature papers, focusing on the problem of how the institutions can gain insights about their candidates while profiling them, how to obtain more accurate information from the interview phase and how to reach a more rigorous assessment of their emotional intelligence providing a better alignment of moral values. This work aims to improve the decision-making process of a PF institution recruiter by turning it into a more automated and evidence-based decision when recruiting an adequate candidate for the job vacancy.

Artificial Intelligence Adoption in Human Resources Management

“…Unbiased, objective machine learning software may detect facial expressions of candidates to evaluate their motivation levels (Kappen & Naber, 2021). Subjective criteria such as nepotism and favouritism are less likely to be observed and the whole process of recruitment is more effective (Bailao et al, 2022;Kshetri, 2021).…”