Organizational moderators of the effect of autonomous technology on organizational error prevention

Abstract

One of the earliest forms of artificial intelligence to be widely adopted in organizations is autonomous decision-making systems (ADS)—advanced automation designed to make human-like decisions that lead to real-world action. ADS technologies are designed with the goal of reducing error, but in complex systems, organizational factors will contribute to the ability of the ADS to achieve this goal. Theories of organizational error have not yet considered how organizational factors will moderate the benefit of ADS use for organizational error avoidance. The present work aims to fill this gap in the literature by integrating theories of organizational error with theory on human-automation interaction to theorize the effects of organizational factors on ADS use and the corresponding relevance to organizational error commission. We hypothesize that ADS use generally decreases error rates, but that some organizational features moderate the value of ADS for error prevention. Among the organizational characteristics that we hypothesize will reduce the error-prevention benefits of ADS are operator experience, task complexity, and interpersonal interdependence. We test these hypotheses on a unique data set from a large, European rail system operator that has recently adopted ADS technology, finding broad support for our theoretical framework.