Machines of Comfort: AI-Companions as a Tool to Combat Loneliness

Abstract

AI companions are increasingly promoted as solutions to loneliness, yet their use raises important ethical questions. This thesis critically examines these questions through four key lenses: well-being, authenticity, informed consent, and commercialization. While these technologies may offer momentary comfort, they fall short of supporting the deeper human capabilities needed for long-term well-being and meaningful connection. Their interactions lack the mutual emotional investment that defines real companionship, and may, in the long run, intensify the very isolation they claim to address. The thesis also explores how emotional vulnerability complicates the notion of informed consent, especially when users are in distress and less able to assess long-term consequences. Finally, it highlights the troubling role of commercialization. Specifically, how these tools are driven by profit models that exploit emotional data with little ethical oversight. By drawing these concerns together, the thesis argues that AI companionship is not ethically justifiable as a solution to loneliness.