How and when AI companionship quality enhances user mental well-being

Abstract

With the rapid advancement of artificial intelligence technology, AI companions have gradually become important tools for alleviating mental health issues such as loneliness and anxiety. Grounded in social support theory, this paper employs a combination of multi-wave questionnaire surveys and experimental methods to comprehensively investigate the mechanisms and boundary conditions by which AI companionship quality affects users' mental well-being. The results reveal that the use of AI companions is associated with enhanced mental well-being, which is mediated by increased emotional clarity and reduced self-presentation concerns among users. Furthermore, mindfulness level plays a significant moderating role: specifically, high-mindfulness users are better able to enhance their emotional awareness and reduce self-evaluative burden through AI companionship, whereas low-mindfulness users, due to cognitive constraints, struggle to translate external support into effective psychological protective factors. In addition, AI anthropomorphism conditions positively moderate the pathway through which AI companionship quality enhances mental well-being via emotional clarity, but its moderating effect on the pathway via reduced self-presentation concerns is not significant, suggesting that the “non-judgmental” nature of AI companions plays a more fundamental role in providing emotional safety. This paper offers managerial implications for the design, development, marketing, and deployment of AI companion products.