Human vs AI: how applicants perceive the fairness of different evaluators during shortlisting.

This study explored how applicants perceive the fairness of artificial intelligence (AI) compared to human evaluators during shortlisting, as well as how selection outcomes influence fairness perceptions. A total of 170 participants completed questionnaires immediately after a simulated job application that included a situational judgment test (T1). Twenty-four hours later, they were randomly assigned to a fictional evaluator and outcome, forming four experimental groups: AI-Successful, AI-Unsuccessful, Human-Successful, and Human-Unsuccessful, and were invited to complete a second questionnaire (T2). The changes in procedural justice between T1 and T2, along with differences in distributive justice, were analysed. Results showed that procedural justice perceptions among unsuccessful participants significantly declined, especially when evaluated by AI. Distributive justice scores were significantly lower for unsuccessful candidates and were unaffected by evaluator type. These findings suggest that while negative outcomes diminish fairness perceptions, AI evaluators are viewed as less procedurally fair than humans in this context. The study highlights the importance of careful and transparent AI implementation in selection processes.