Emotional attachment to AI chatbots: Evidence from Germany, China, South Africa, and the United States

Artificial intelligence (AI) chatbots are increasingly embedded in everyday life, yet their role extends beyond functional assistance to include affective human–machine interaction. While prior research has focused on social chatbots or niche user groups, little is known about the prevalence and consequences of emotional attachment to widely used, general-purpose systems. This study addresses this gap using cross-national survey data from Germany, China, South Africa, and the United States (N = 7027). We develop a dynamic interaction framework in which emotional attachment emerges from the interplay between perceived emotional support, social connectedness, and interaction patterns. Empirically, we show that emotional attachment is widespread: at least one third of users report attachment-related behaviors. Perceived emotional support—particularly reduced loneliness, freedom from judgment, and a sense of privacy—is the strongest predictor. Contrary to expectations, larger social networks are associated with stronger attachment. Emotional attachment is, in turn, strongly associated with indicators of user dependence. These findings highlight the growing affective integration of AI into social life and raise important questions for digital well-being and AI governance.