When Does an AI Interviewer Work? How Task Type and Interview Outcome Shape Procedural Justice, Source Credibility, and Negative WOM

This study examines how interviewer type (AI vs. human), task type (factual vs. evaluative), and interview outcome (success vs. failure) jointly shape applicants' perceptions of procedural justice, source credibility, and negative word-of-mouth (WOM) intentions. Using a 2 × 2 × 2 between-subjects experimental design with 199 Korean adults, the study tested (a) the sequential mediating roles of procedural justice and source credibility (trustworthiness and expertise) in the relationship between interviewer type and negative WOM intentions, and (b) the moderating roles of task type and interview outcome on these indirect pathways. Results showed that AI interviewers, compared to human interviewers, generated lower procedural justice perceptions, which in turn reduced source trustworthiness and ultimately increased negative WOM intentions. A significant three-way interaction among interviewer type, task type, and interview outcome further emerged for source trustworthiness. In the success condition, AI interviewers were perceived as more trustworthy on factual than evaluative tasks. In the failure condition, however, this pattern reversed: AI interviewers were perceived as substantially less trustworthy than human interviewers specifically when the task was factual. Moderated mediation analyses further showed that the sequential indirect effect on negative WOM was significant only when applicants failed a factual task. These findings highlight that both task type and interview outcome are critical boundary conditions shaping applicants' responses to AI-mediated evaluation.