GAI identity threat: When and why do individuals feel threatened?

Abstract

The widespread application of generative artificial intelligence (GAI) and its ability to perform various tasks have raised concerns about identity threats. This study aims to investigate antecedents, behavioral consequences, and boundary conditions of GAI identity threat. We conduct a two-phase study using a mixed-methods approach. The qualitative phase aims to identify the critical factors that influence GAI identity threat. After coding 30 semi-structured interview transcripts, we find that GAI's high intelligence and human-like abilities facilitate three kinds of affordance that influence GAI identity threat: creative affordance, analysis affordance, and communication affordance. The quantitative phase, based on IT identity threat theory, constructs a research model to examine how the three types of affordance influence GAI identity threat, as well as identifying the behavioral consequences. An examination of survey data from 405 users indicates that GAI's creative affordance, analysis affordance, and communication affordance are associated with GAI identity threat, which may lead to users’ resistance behavior. Furthermore, GAI autonomy and user self-identity moderate the relationship between GAI affordances and GAI identity threat, respectively. This study identifies users’ negative psychological mechanisms when using AI, develops IT identity threat theory in the context of GAI, and provides implications for practitioners to reduce identity threats.