Becoming a cognitive miser? Antecedents and consequences of addictive ChatGPT use

Generative Artificial Intelligence (GenAI) has significantly enhanced productivity across diverse domains, however, growing concerns persist regarding its addictive use and potential cognitive ramifications. This study employed dual-system theory to examine the cognitive-behavioral mechanisms underpinning addictive ChatGPT use, positioning self-construal as a key antecedent and cognitive miserliness as a critical consequence. We adopted a mixed-methods approach for both exploratory and confirmatory analysis. Three cross-sectional surveys revealed that heightened activity in the impulsive cognitive system was a significant predictor of addictive ChatGPT use. Moreover, a dysregulation between the impulsive and reflective cognitive systems exacerbated compulsive tendencies and facilitated cognitive miserliness. Self-construal also affected these effects: individuals with an independent orientation were more likely to rely on the impulsive cognitive system, while those with an interdependent orientation engaged the reflective cognitive system more frequently. Notably, the reflective cognitive system often failed to adequately inhibit addictive ChatGPT use. Qualitative insights supported these findings and offered a deeper understanding of users' behavioral patterns. This study contributes to theoretical discourse on the adverse cognitive impact of GenAI use, and informs managers seeking to design interventions that promote healthier digital engagement.