AI 지원 성과평가 유형이 리더-구성원 교환관계(LMX)에 미치는 영향 : 알고리즘 환원주의와 리더 개입의 매개효과
Impact of AI-supported Performance Evaluation Types on Leader-Member Exchange : Mediating Roles of Algorithmic Reductionism and Leader Intervention
As artificial intelligence (AI) is increasingly integrated into human resource management (HRM), organizations are adopting AI-supported evaluation systems to enhance the efficiency and objectivity of assessments. However, little is known about how different types of AI-supported performance evaluations affect employees’ perceptions of leader-member exchange (LMX). Drawing on LMX and attribution theory, this study analyzed the impact of AI-supported evaluation types on employees’ LMX perceptions and examined the mediating roles of “algorithmic reductionism” and “leader intervention.” To this end, a scenario-based experiment using a 2×2 factorial design was conducted, manipulating AI transparency (high vs. low) and responsibility attribution (leader vs. AI). Participants were assigned to one of four evaluation scenarios and subsequently responded to items regarding algorithmic reductionism, leader intervention, and LMX perceptions. The study results indicate that perceptions of LMX vary significantly depending on evaluation conditions. Members perceived the highest levels of LMX when leaders actively reviewed AI-generated evaluation results and took responsibility for the final decisions. Additionally, algorithmic reductionism was found to significantly influence LMX perceptions. These findings underscore the critical role of human responsibility in AI-driven HR decision-making processes. Academically, this study contributes to expanding research on AI-based HRM; practically, it suggests that organizations must ensure active leader involvement when implementing AI-driven performance evaluation systems.
목차
Abstract 1. 서론 2. 이론적 배경 2.1 알고리즘 통제와 조직 권력, 공정성 및 신뢰의 변화 2.2 AI 지원 평가유형이 LMX에 미치는 영향 메커니즘과 연구가설 3. 연구방법 3.1 연구모델 3.2 측정도구 3.3 실험 조작 점검 4. 연구결과 4.1 인구통계학적 정보 4.2 측정모델 분석 4.3 경로분석에 기반한 평가 유형별 LMX 영향 4.4 이원분산분석에 기반한 평가 유형별 LMX 영향 차이 5. 결론 5.1 연구요약 5.2 시사점 및 향후 연구계획 References Appendix