ECIS 2020 Research Papers

Abstract

With technological developments in artificial intelligence, algorithms are increasingly capable to perform tasks that were considered to be unique for humans. However, literature suggests that although algorithms are often superior in performance, users are reluctant to interact with algorithms instead of human agents – a phenomenon known as algorithm aversion. But, as algorithm aversion is attracting scientific attention, empirical findings are inconclusive and papers find the opposite effect of algorithm appreciation. With this literature review, we synthesize evidence from 29 publications with 84 distinct experimental studies to investigate how algorithm characteristics and human agents’ characteristics influence algorithm aversion. We show how algorithm agency, performance, perceived capabilities and human involvement as well as human agents’ expertise and social distance, influence whether users develop algorithm aversion, i.e., choose humans over algorithms, utilize humans’ support more often and evaluate humans’ actions more favourable. Furthermore, we provide a systematic conceptualization of aversion as a biased assessment and develop propositions for future research. With our work, we contribute to algorithm aversion literature and the contemporary discussion on the impact of algorithmic agents on the future of work. We indicate that the emerging literature stream on algorithm aversion is worth considering for human-computer interaction researchers.

Share

COinS
 

When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.