In Algorithms We Trust? Exploring the Role of Human-in-the-Loop and Social Presence
Abstract
Algorithmic decision-making (ADM) is widely applied in organizations and is associated with increased objectivity, consistency, and efficacy. However, when tasks are associated with social or emotional skills, as is often the case in personnel selection, humans commonly view its use as untrustworthy. We suggest that social presence accounts for negative perceptions of the use of ADM and propose the incorporation of a human-in-the-loop (HITL) approach mitigates perceptions of untrustworthiness. We conducted an online experiment (n = 416) to examine applicants’ perceptions of ADM and HITL in comparison to human decision-making (HDM). We found social presence is significantly associated with trust, although it did not differ significantly between decision entities. Further, our results contribute to the literature by identifying that a HITL design not only enhances trust in ADM systems but is also perceived as more trustworthy compared to HDM alone. Further, our findings demonstrate that trust mediates the relationship between decision entities and organizational attractiveness.
Keywords
Algorithmic Decision-Making; Trust Transparency; Human-in-the-Loop