Abstract
PURPOSE: The advent of conversational AI has offered individuals an alternative communication modality and an interactive partner outside traditional social media through one-on-one personalized services. This has given rise to a critical question: Does this technological innovation negatively affect specific populations? The sample for this study was comprised of individuals with attachment anxiety. The aims of the study were to explore the relationship between attachment anxiety and problematic use of conversational artificial intelligence and to investigate the mediating role of emotional attachment and the moderating role of anthropomorphic tendency. PATIENTS AND METHODS: Fifty-four Chinese adults who used conversational artificial intelligence were randomly selected to complete the questionnaires of attachment anxiety, emotional attachment, anthropomorphic tendency and the use of conversational artificial intelligence. Correlation analyses were employed to explore the relationships among the above mentioned variables, and the Hayes PROCESS test was used to test the mediating effect of emotional attachment and the moderating effect of anthropomorphic tendency. RESULTS: Attachment anxiety not only directly affects the problematic use of conversational artificial intelligence but also indirectly influences it through emotional attachment. The anthropomorphic tendency was also found to moderate the relationship between attachment anxiety and the problematic use of conversational artificial intelligence. CONCLUSION: Individuals with high attachment anxiety urgently need relatively perfect partners to reduce negative emotions. While the emergence of conversational artificial intelligence meets their needs, there is a risk of excessive dependence. To solve this problem, developers should focus on the personification design of the product, embed usage control and educational content, and provide intervention measures for high-risk groups.