Applicants’ Fairness Perception of Human and AI Collaboration in Resume Screening

Abstract

Artificial intelligence (AI) is increasingly used in resume screening for its efficiency and cost-effectiveness. However, researchers have noted concerns about applicants’ perceptions of fairness in AI-based decision-making, suggesting that human intervention remains necessary. Our study examines three human-AI partnerships: human-led, AI-led, and balanced structures, and explores how these human-AI aggregated decision-making structures affect applicants’ perceptions of fairness through the lens of responsibility attribution. In the context of resume screening, our experiment (N = 283) reveals that human-led and balanced structures lead to higher perceptions of fairness than AI-led structures, particularly when the outcome is rejection. Rejected applicants in a balanced human-AI structure tend to attribute less responsibility to the employer and perceive greater fairness than those rejected in an AI-led structure. Our research integrates applicant attribution-reaction theory (AART) into the field of AI management and suggests collaborative roles of humans and AI in managing applicants’ responsibility attributions and fairness perceptions.