Social Without Society: Parasocial Relationships in Human–Generative AI Agent Interaction

Generative artificial intelligence (AI) is increasingly used in everyday conversations, allowing users to interact with AI in ways that feel social rather than purely functional. This study examines how users develop relationships with generative AI and how these relationships influence their future use and well-being. Drawing on social support and parasocial relationship theories, we propose that generative AI can provide different types of social support, including emotional support, informational support, and companionship. These supportive interactions encourage users to experience parasocial interaction, which may gradually develop into a perceived relationship with AI. We further examine how such relationships affect users’ intention to continue using AI and their subjective well-being. By viewing generative AI as a social actor instead of only a technological tool, this research explains why users become emotionally engaged with AI systems and highlights the psychological and behavioral outcomes of human–AI interaction.