Chatting with confidants or corporations? Privacy management with AI companions

AI companion chatbots feel like intimate confidants but exist as part of corporate data systems. Drawing on Communication Privacy Management (CPM) theory and Masur’s dimensional privacy framework, we interviewed users of platforms like Replika and Character.AI (N = 15) to understand their privacy decisions in AI companion interactions. We find that anthropomorphic design lowers disclosure thresholds, encouraging increasingly intimate self-disclosure. As relationships develop, some users engage in simulated co-ownership, treating shared information as relationally held despite recognizing corporate control. Because post-disclosure correction feels futile, privacy management becomes anticipatory and front-loaded—users rely on distancing strategies such as using fake names rather than post-hoc deletion. For some users, privacy turbulence centers less on fears of exposure and more on fears of losing conversational memory—a relational reset feels more threatening than a data breach. These findings extend CPM theory by showing that in human–AI companionship contexts, relational and institutional privacy concerns become inextricably entangled.