Decoding algorithm appreciation: Unveiling the impact of familiarity with algorithms, tasks, and algorithm performance

Abstract

Algorithm appreciation, defined as an individual's reliance or tendency to rely on algorithms in decision-making, has emerged as a subject of growing scholarly interest. Inquiries into this subject are crucial to understanding human decision-making processes as in the era of artificial intelligence, algorithms are increasingly being integrated into decision-making. To contribute to this evolving field, this study examines three factors that might play significant roles in enhancing trust in algorithms: familiarity with algorithms, familiarity with tasks, and familiarity with algorithm performance. Drawing upon prior studies, a conceptual model was developed and empirically tested using a scenario study. Data on 327 individuals showed a strong positive association between familiarity with algorithms and trust in algorithms. In contrast, task familiarity appeared to have no significant influence on trust. Trust, in turn, was identified as a key driver of algorithm appreciation. The study also revealed the moderating role of familiarity with algorithm performance in the relationship between familiarity with algorithms and trust in algorithms. Post hoc analysis highlighted that trust fully mediates the relationship between algorithm familiarity and algorithm appreciation. The study underscores the significance of algorithm familiarity and performance transparency in shaping trust in algorithms. The study contributes theoretically by offering important insights about the influences of different forms of familiarity on trust and practically by prescribing practical guidelines to enhance algorithm appreciation.