Abstract:

As the performance of artificial intelligence (AI) enabled algorithms to improve, so too increases the potential for them to be used to increase the efficiency and effectiveness of human resource management (HRM) decisions. Yet, public distrust in AI algorithms could keep organizations from using this technology to improve HRM decision-making. Here, we examine one factor that may influence the perceived trustworthiness of AI algorithms used in HRM, specifically those used in personnel selection decisions. Drawing from organizational justice and trust theories, we posit that knowledge of how the algorithm compares with human recruiters in terms of hiring members of traditionally discriminated demographic groups serves as a fairness heuristic that affects the algorithm's perceived trustworthiness by increasing its perceived ability, benevolence, and integrity. In three experimental studies (N = 1382), we show that when people are informed that an algorithm used in personnel selection results in more women or racial minorities being hired, as compared to selection decisions made by human recruiters, they perceive it as having higher ability, benevolence and integrity, and are more willing to adopt it and to follow its recommendations. The opposite is true when the algorithm is said to decrease the number of women and racial minorities being hired. Our research suggests that auditing personnel selection decisions made using AI algorithms and communicating how they compare with human recruiters in terms of their diversity, equity, and inclusion outcomes is important for the perceived trustworthiness and public acceptance of this technology.

Publisher's Version