Literature Review on Algorithm Aversion: Definition, Antecedents, and Mitigation

Abstract

With the rapid advancement of information technology, algorithms are increasingly deployed across finance, healthcare, education, social media, and many other domains. Although algorithms can improve both the accuracy and efficiency of decisions, a robust body of evidence shows that people often prefer human judgment even when algorithms outperform humans—a phenomenon termed “algorithm aversion.” This paper reviews the literature to delineate the concept of algorithm aversion, identify its psychological and contextual antecedents (e.g., psychological mechanisms, algorithm design features, and task characteristics), and contrast it with the emerging phenomenon of “algorithm appreciation.” We further examine how aversion shapes consumer behavior and decision processes, and synthesize evidence-based remedies such as enhancing transparency, granting users control over algorithmic parameters, and increasing user involvement in algorithm development. By integrating insights from decision psychology and behavioral economics, the review enriches theoretical understanding and offers actionable recommendations for designing algorithms that are trusted, accepted, and ultimately more effective in real-world applications.