This study aims to examine users ' perceptions of reliance on generative AI and to explore which configurations of social representations shape such reliance. To this end, the study integrates Social Representation Theory with fuzzy-set Qualitative Co...
This study aims to examine users ' perceptions of reliance on generative AI and to explore which configurations of social representations shape such reliance. To this end, the study integrates Social Representation Theory with fuzzy-set Qualitative Comparative Analysis (fsQCA) to identify configurations of perceptions underlying cognitive and emotional reliance on generative AI.
In Study 1, semi-structured interviews were conducted with users who had experience using generative AI. The interview data were analyzed through open coding and network analysis, which identified the core social representations constituting generative AI reliance.
In Study 2, survey items were developed based on the core perceptions identified in Study 1, and a survey was administered to users who self-reported reliance on generative AI. The fsQCA results revealed that cognitive reliance is primarily formed around task convenience and promptness, with psychological comfort, trust, and emotional venting constituting three alternative configurations. In contrast, emotional reliance was found to be driven by emotional venting and psychological comfort as core conditions. Notably, a distinct configuration was identified in which emotional reliance emerged even at low levels of trust when functional efficiency was sufficiently perceived and psychological interactions, such as emotional venting, were present.
Overall, these findings indicate that generative AI reliance shares a common foundation in task convenience, with cognitive reliance driven by function-oriented mechanisms and emotional reliance shaped by emotion-centered experiential mechanisms. Moreover, by adopting a mixed-methods approach that integrates Social Representation Theory with fsQCA, this study empirically demonstrates the multiple pathways and asymmetric causal structures underlying the reliance on generative AI.