When Algorithms Fail: Consumers’ Responses to Brand Harm Crises Caused by Algorithm Errors

Abstract

Algorithms, increasingly used by brands, sometimes fail to perform as expected or, even worse, cause harm, leading to brand harm crises. Algorithm failures are unfortunately increasing in frequency, yet little is known about consumers’ responses to brands following such crises. Extending developments in the theory of mind perception, the authors hypothesize that, following a brand harm crisis, consumers respond less negatively if the error was caused by an algorithm (vs. a human). The authors further hypothesize that consumers’ lower mind perception of agency of the algorithm (vs. a human), which lowers their perceptions of the algorithm’s responsibility for the harm, mediates this relationship. Four moderators of this relationship are hypothesized: two algorithm characteristics (whether the algorithm is anthropomorphized and whether it involves machine learning) and two characteristics of the task for which the algorithm is deployed (whether the task is subjective [vs. objective] and whether it is interactive [vs. noninteractive]). The authors find support for the hypotheses in eight experimental studies. The effects of two managerial interventions to manage brand harm crises caused by algorithm errors are examined. This research advances the literature on brand harm crises, algorithm usage, and algorithmic marketing and generates managerial guidelines to address such crises.