•  
  •  
 

AIS Transactions on Human-Computer Interaction

Abstract

As algorithmic systems increasingly provide advice across a wide range of decision-making contexts, understanding when and why people rely on algorithmic recommendations has become critical. This study examines how individual differences in uncertainty beliefs shape whether they accept algorithmic versus human advice. Drawing on decision making under uncertainty theories, we integrated an economic lottery choice task with three preregistered judge–advisor system experiments that varied in task objectivity. Across tasks, we observed a robust algorithmic appreciation effect that is systematically moderated by individual uncertainty preferences. Higher uncertainty tolerance amplified algorithmic advice taking in more objectively verifiable tasks, whereas preferences for non-social versus social sources of uncertainty predicted advice acceptance in less objectively verifiable tasks. These patterns reveal substantial heterogeneity in algorithmic reliance that task characteristics alone cannot explain. Moreover, uncertainty tolerance and uncertainty source preference significantly improve prediction of advice-taking behavior beyond standard task and behavioral factors. Together, our findings demonstrate that behaviorally revealed uncertainty preferences partly shape algorithmic aversion and appreciation. This work contributes to human–computer interaction and information systems research by identifying person-centric mechanisms that explain who use algorithmic advice and when and why they do so and, thus, informing efforts to design more adaptive and human-centered decision-support systems.

DOI

10.17705/1thci.00247

Download

Share

COinS
 

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