Figure 1
The figure presents the conceptual model of the study. Four features of the human–AI decision arrangement—interaction quality, human agency, perceived humanness and voice—are positioned as antecedents of fairness, accountability and transparency. Fairness, accountability and transparency are subsequently linked to trust in AI-supported performance evaluation. Task complexity and perceived uncanniness are included as moderating variables in the model, with dashed lines indicating the proposed moderation relationships. Solid arrows represent the hypothesised direct relationships between the antecedents, fairness, accountability and transparency, and trust.Conceptual model. Source: Authors' own work