Problematic ChatGPT use: Manifestations, etiologies, and evaluation

ChatGPT has rapidly permeated various domains, including education, industry, creativity, and daily life. While offering efficiency and emotional engagement, excessive reliance may lead to problematic use with adverse effects on mental health, cognition, and social functioning. This review synthesizes current evidence on manifestations, etiologies, assessment, and neurophysiological correlates of problematic ChatGPT use.

A narrative review of empirical studies, psychometric developments, theoretical models, and preliminary electroencephalogram findings was conducted, with cross-cultural comparisons across Taiwan, Turkey, China, Korea, Singapore, and the United States.

Problematic ChatGPT use is prevalent among young adults and associated with depression, stress, impaired self-control, and functional difficulties. Psychometric tools such as the Problematic ChatGPT Use Scale, Conversational AI Dependence Scale, and Generative AI Dependency Scale provide initial frameworks but require further validation. Theoretical models (Interaction of Person-Affect-Cognition-Execution model, compensatory Internet use, dual-system theory) have been used to explain the roles of emotional regulation deficits, cognitive biases, and intuitive processing for the development of problematic ChatGPT use. Preliminary electroencephalogram studies suggest altered functional connectivity, paralleling digital addiction patterns.

Problematic ChatGPT use involves psychological, behavioral, and neurobiological mechanisms. Future research should strengthen psychometric validation, expand electroencephalogram investigations, and evaluate interventions such as digital literacy education, behavioral self-regulation, cognitive-behavioral therapy, and neurofeedback to promote healthier AI engagement.