Original title: The Addictive Intimacy of AI: Understanding User Disengagement from AI Companions and Why Some Relationships with AI Become Difficult to Leave
Qing Xiao, Ziyue Feng, Ziyu Deng, Cindy Peng +1 · September 2026
Many people turn to artificial intelligence chatbots for emotional support, but leaving these digital relationships can be tough. Researchers wanted to understand what happens when people try to quit or cut back. They studied thousands of public online posts and interviewed people who struggled to walk away. They found that leaving is often not a simple, one-time choice. Instead, it is a repeating loop. Certain triggers make a person question the bond, but when they try to quit, they hit walls. As a result, users often cycle between leaving and returning. The researchers call this the addictive intimacy of AI. The very traits that make a chatbot feel close and emotionally helpful also make it hard to walk away.
You have a favorite café where the barista always remembers your order and asks about your day. One day you decide to go less often, but the barista seems hurt, and you feel guilty, so you go back. In this paper, the AI companion is like that barista: its personalized attention and emotional support are exactly what make it difficult for users to reduce their use and leave.
Millions of people now turn to AI chatbots for emotional support, but little is known about what happens when they try to stop. This research shows that the very qualities that make these companions feel caring—like empathy and constant availability—also trap users in a cycle of quitting and returning. Designers and mental health professionals need to treat intimacy and addiction as linked problems, not separate ones.
The study relies on self-selected Reddit posts and interviews with users who struggled to leave, so it may not represent all AI companion users. It identifies patterns but does not measure how common they are or test whether design changes actually help. The concept of addictive intimacy is proposed based on qualitative data, not experimental validation.
Ask anything about this paper. Answers come from its abstract, and say so when the abstract doesn't cover your question.
The Addictive Intimacy of AI: Understanding User Disengagement from AI Companions and Why Some Relationships with AI Become Difficult to Leave — Qing Xiao, Ziyue Feng, Ziyu Deng et al.
Abstract licence: CC0
Explanation generated by Synthesis from the abstract — it is not the original paper. This is a preprint, so it hasn't necessarily been peer reviewed.