Abstract
Promotion decisions depend fundamentally on subjective assessments of competence and reliability, making them vulnerable to strategic manipulation in competitive organizational environments. As human resource management increasingly delegates evaluation authority to artificial intelligence systems, a critical question emerges about whether algorithmic mediation transforms not merely decision accuracy but the ethical conduct of employees who provide evaluative input within these systems. Algorithmic processing may enhance accountability through systematic record-keeping and pattern detection, yet simultaneously erode the interpersonal scrutiny that constrains credibility-undermining behavior. We examine how evaluation architecture shapes managers' deployment of subtle gaslighting tactics in competitive promotion contexts. In an experimental study, mid-level managers completed performance appraisals of an employee with strategically ambiguous performance competing for limited bonus allocation. Managers were randomly assigned to evaluation conditions where their assessments would be processed through an autonomous AI system applying predefined algorithmic criteria or reviewed by a senior management compensation committee exercising human judgment. Independent raters blind to experimental conditions assessed evaluations for gaslighting tactics, overall negativity, and factual grounding, while participants reported perceived accountability and psychological distance from consequences. This research interrogates whether algorithmic mediation inadvertently undermines the relational foundations necessary for organizations and individuals to thrive together in competitive workplace environments, with implications for responsible AI deployment in human resource management.