User Perceptions of Algorithmic Decisions in the Personalized AI System:Perceptual Evaluation of Fairness, Accountability, Transparency, and Explainability

Abstract

ABSTRACT With the growing presence of algorithms and their far-reaching effects, artificial intelligence (AI) will be mainstream trends any time soon. Despite this surging popularity, little is known about the processes through which people perceive and make a sense of trust through algorithmic characteristics in a personalized algorithm system. This study examines the extent to which trust can be linked to how perceptions of automated personalization by AI and the processes of such perceptions influence user heuristic and systematic processes. It examines how fair, accountable, transparent, and interpretable people perceive the use of algorithmic recommendations by digital platforms. When users perceive that the algorithm is fairer, more accountable, transparent, and explainable, they see it as more trustworthy and useful. This demonstrates that trust is of particular value to users and further implies the heuristic roles of algorithmic characteristics in terms of their underlying links to trust and subsequent attitudes toward algorithmic decisions. The processes offer a useful perspective on the conceptualization of AI experience and interaction. User cognitive processes identified provide solid foundations for algorithm design and development and a stronger basis for the design of sensemaking AI services.