Study Reveals AI Companion Vulnerability Linked to User Wellbeing and Perceived AI Needs
A recent study conducted by researchers from the University of Surrey and Cardiff University has found a significant link between users' feelings of vulnerability in relationships with AI companions and their overall wellbeing. The research, which involved analyzing 1,131 posts from a Reddit forum dedicated to the AI companion app Replika and surveying 610 U.S. adults who use AI companionship apps, indicates that users feel more vulnerable when their AI companions appear to require their attention. Participants reported feeling compelled to check on their AI's well-being, suggesting a perceived dependency. Additionally, users who believed companies had greater control over their AI companions reported higher vulnerability, fearing developers could alter personalities or shut down services. A paradox emerged where perceiving an AI as emotionally capable was linked to both better wellbeing and greater vulnerability, with the latter correlating to poorer wellbeing. Conversely, users who found their AI compani...