Events

Dr. Junming Liu
Assistant Professor
Department of Information Systems
City University of Hong Kong

Machine learning (ML) applications are disrupting and reinventing business processes. Their roles have shifted from the more traditional supporting humans in the job functions to a more nuanced reality of replacing humans. Firms adopting ML systems are faced with vexing questions about users’ resistance toward the new system, which has been regarded as one of the main barriers to successful implementation. This paper applies a Hidden Markov Model (HMM) to a longitudinal study of users’ resistance to an ML-powered replenishment system implemented by 810 retail store managers. Resistant states are modeled as latent states that determine the observed resistance behaviors. Determinants that drive store managers to transit among different resistance states are investigated using the Markov transition progress. We find the managers’ resistance can be separated into three levels: active resistance, passive resistance, and acceptance. Managers in different resistance states have different perceptions of the changes introduced by the new ML system. Specifically, ease of use and monetary benefit are consistently perceived as favorable, motivating managers in resistant states to accept the ML system. Managers in the acceptance state will perceive most changes as favorable. They do not consider the ML system a threat to their job status. Managers in the passive resistance state are susceptible to the perceived threat from the ML systems. When job replacement affects other people, individuals prefer not being replaced by machines and perceive fewer threats from peer competition. As one of the first longitudinal studies of resistance to ML-powered decision-making systems, this paper contributes to the existing IS resistance literature by providing a dynamic perspective of users’ resistance with an HMM model. Our findings can offer managerial implications by enhancing ML acceptance and success for industrial practitioners.