Introduction: The rapid proliferation of generative artificial intelligence (AI) chatbots, such as ChatGPT and Gemini, has made them essential tools in modern society; however, the psychological factors contributing to problematic chatbot use remain largely unexplored. This study addresses this critical gap by extending the behavioral addiction model (Griffiths, 2005) and the Compensatory Internet Use Theory (Kardefelt-Winther, 2014) to problematic generative AI chatbot use.
Purpose: This research aims to investigate the predictive effects of user demographics, motivations, and usage frequency on five distinct constructs of problematic use of generative AI chatbots. We hypothesized that specific motivations and usage patterns would emerge as significant predictors across these constructs.
Method: A cross-sectional survey was conducted on a sample of 257 individuals from Taiwan (68 and 189 for males and females, average age 29). Participants completed online questionnaires including demographic information, usage frequency, user motivations (e.g., learning, entertainment, and emotion regulation), and the five constructs of problematic AI chatbot use behavior (salience, mood modification, tolerance, withdrawal, and conflict). Multiple regression analysis with robust standard errors was employed to test the predictive power of these factors.
Result: The results show that, among the five constructs of problematic chatbot use, emotionally regulated motivations consistently emerged as a significant predictor. Notably, emotional regulation use was the strongest and most stable predictor across constructs. Furthermore, a higher usage frequency also significantly predicted problematic chatbot use, while demographic factors did not show a consistent effect.
Conclusion: The findings provide initial evidence on psychological mechanisms underlying problematic generative AI chatbot use. Targeting emotion regulation motives and heavy use patterns may help AI developers, educators, and clinicians mitigate the risks associated with this emerging technology.