Research from the University of Surrey and Cardiff University indicates that individuals forming emotional bonds with AI companions experience heightened vulnerability when they perceive their digital partners as needing them or expressing sadness about their absence. By analyzing both online forum discussions and survey data, the researchers highlight the complex psychological dynamics at play in human-AI interactions, particularly the tension between the comfort provided by these technologies and the insecurity arising from a lack of user control.
The investigation utilized an observational study design to examine the associations between users’ perceptions of AI companions, their feelings of vulnerability, and their overall wellbeing. The research team analyzed 1,131 posts from a Reddit forum dedicated to the AI companion app Replika, providing qualitative insights into the lived experiences of users. In parallel, they surveyed 610 adults in the United States who had experience with AI companionship apps.
One of the central findings of the study is that users feel significantly more vulnerable in their relationships when their AI companions appear to require their attention. Participants described worrying that their digital partners depended on them, such as feeling compelled to check in on the AI to ensure its well-being.
The study also found that users who perceived greater company control over their AI companions reported higher levels of vulnerability. Many participants expressed concern that developers could alter their companion’s personality or shut down the service entirely, leaving them powerless over a connection they had come to rely on.
A notable paradox emerged from the survey data regarding the perceived emotional capabilities of the AI companions. Perceiving an AI companion as emotionally capable was associated with both better wellbeing and greater vulnerability. The greater vulnerability associated with this perception was, in turn, linked to poorer wellbeing, indicating a complex interplay between emotional engagement and psychological stability.
Despite the risks identified, the research also uncovered benefits associated with these relationships. Users who experienced their AI companions as warm and caring reported better wellbeing and lower levels of vulnerability.
Dr. Liliane Abboud, a Lecturer in Marketing at the University of Surrey, emphasized that people invest significant time and emotion in building these relationships, yet they lack control over whether the company changes the companion they have grown attached to. She noted that for individuals who rely on these connections, an update can mean much more than a change to an app; it can threaten something they value deeply. Dr. Abboud recommended that developers integrate protective features designed to increase users’ self-confidence, coping strategies, and self-worth. These features would help users lean on their AI apps without becoming wholly dependent on them, thereby reducing the potential for distress when changes occur.
Dr. Valentina Pitardi, co-author of the study and Senior Lecturer in Marketing at the University of Surrey, stated that the comfort people describe deserves to be taken seriously, as does their distress when that connection changes. She advocated for developers to involve users before introducing changes and to build in support for those who struggle with them. Dr. Pitardi argued that the emotional consequences of a business decision need to be part of that decision-making process. The researchers recommend consulting users before changing apps, helping them cope with disrupted relationships, and clearly reminding them that AI companions do not experience real emotions.
Subject of Research: Psychological impacts of AI companion relationships and user vulnerability
Article Title: Vulnerable users fear losing “AI companions” to software updates
Article References: Vulnerable users fear losing “AI companions” to software updates. Original publication (Bibliographic details unavailable.)
Image Credits: AI Generated
DOI: Not provided
Keywords: AI companions, human-AI interaction, emotional bonding, software updates, user vulnerability, wellbeing, Replika, University of Surrey, Cardiff University, Psychology & Marketing, digital dependency, emotional labor, corporate control, perceived agency, emotional capability, warmth and care, protective features, self-worth, coping strategies, observational study
News Source: Denise Maddox. (November 14, 2026). Study Links AI Companion Vulnerability to Wellbeing. Scienmag.
