Key points are not available for this paper at this time.
Purpose Algorithmic decision-making is changing the dynamics of operability. It has a profound impact on decision-making across a broad spectrum of domains. The study adopts a generalized approach, examining the algorithm aversion without limiting it to a specific sector, to uncover, validate and prioritize factors that are broadly applicable across diverse domains.Research Methodology A semi-structured personal investigation was conducted to collect data from 184 respondents. A fuzzy analytic hierarchy process (F-AHP) was performed to assign individual weights to each criterion. The criteria were ranked based on their global weights, and the individual impacts of each criterion and sub-criterion were tested and confirmed.Findings The literature screening revealed six primary criteria and thirty-one sub-criteria influencing algorithm aversion, with personality factors emerging as the most significant. Attributes such as self-esteem, self-efficacy, neuroticism, locus of control and extraversion significantly shape individual responses to algorithmic decision-making. Algorithm-specific characteristics, including transparency, reliability, and explainability, along with the psychological aspects also play prominent roles.Implications The study provides practical insights for algorithm developers to improve user trust, such as developing user-friendly interfaces, error feedback mechanism and promoting human-algorithm collaboration, while also urging policymakers to create frameworks that guarantee fairness, transparency, and accountability in automated systems.Originality The present study is the first to use a multi-criteria decision-making (MCDM) technique to demystify the drivers of algorithm aversion.
Gupta et al. (Mon,) studied this question.