Augmenting diversity in hiring decisions with artificial intelligence tools

Abstract

Existing research draws an ambiguous picture about the impact of artificial intelligence (AI) in hiring decisions, particularly in augmenting diversity. This study moves beyond the common question of whether humans or machines make better decisions by investigating the conditions under which AI-supported tools can enhance diversity outcomes and augment the individual decision maker. We conduct an experiment with a simulated AI tool with a sample of 139 participants making 278 hiring decisions in the medical field. Our study elaborates on sociomaterial theorizing while exploring how characteristics of the AI tool, organizational factors, and user backgrounds influence decision-making. Results indicate that providing a reliable AI-based decision support tool alone cannot enhance diversity. Instead, diversity-related explanations provided by the AI tool, hiring contexts emphasizing qualitative goals, and explicit diversity guidelines constitute diversity-enhancing conditions. These conditions go hand-in-hand with decision makers' reflection of the hiring decision. Their AI literacy and prior HR experience did not show any significant effects.