Fairness, AI & recruitment

Abstract

The ever-increasing adoption of AI technologies in the hiring landscape to enhance human resources efficiency raises questions about algorithmic decision-making's implications in employment, especially for job applicants, including those at higher risk of social discrimination. Among other concepts, such as transparency and accountability, fairness has become crucial in AI recruitment debates due to the potential reproduction of bias and discrimination that can disproportionately affect certain vulnerable groups. However, the ideals and ambitions of fairness may signify different meanings to various stakeholders. Conceptualizing fairness is critical because it may provide a clear benchmark for evaluating and mitigating biases, ensuring that AI systems do not perpetuate existing imbalances and promote, in this case, equitable opportunities for all candidates in the job market. To this end, in this article, we conduct a scoping literature review on fairness in AI applications for recruitment and selection purposes, with special emphasis on its definition, categorization, and practical implementation. We start by explaining how AI applications have been increasingly used in the hiring process, especially to increase the efficiency of the HR team. We then move to the limitations of this technological innovation, which is known to be at high risk of privacy violations and social discrimination. Against this backdrop, we focus on defining and operationalizing fairness in AI applications for recruitment and selection purposes through cross-disciplinary lenses. Although the applicable legal frameworks and some research currently address the issue piecemeal, we observe and welcome the emergence of some cross-disciplinary efforts aimed at tackling this multifaceted challenge. We conclude the article with some brief recommendations to guide and shape future research and action on the fairness of AI applications in the hiring process for the better.