Ghost in the Machine: On Organizational Theory in the Age of Machine Learning

Abstract

With rapid advancements in machine learning, we consider the epistemological opportunities presented by this novel tool for promoting organizational theory. Our paper unfolds in three sections. We begin with an overview of the three forms of machine learning (supervised, reinforcement, and unsupervised), translating these onto our common modes of research (deductive, abductive, inductive, respectively). Next, we present frank critiques of machine learning applications for science as well as of the state of organizational scholarship writ large, highlighting contemporary challenges in both domains. We do so to make the case that machine learning and theory are not in competition but have the potential to play complementary roles in moving our field beyond siloed domains and incremental theory. Our final section speaks to this synergy. We propose that machine learning can act as a tool to test and prune mid-range theory and as a catalyst to expand the explanatory spectrum that theory can inhabit. Specifically, we outline how machine learning can support local but perishable theory targeting pragmatic problems in the here and now, and grand theory that is sufficiently bold and generalizable across contexts and time to serve the social-functional purposes of inspiring and facilitating long-term epistemological progress across domains.