Augmenting or Replacing Human Connection? An Ethical Analysis of AI Companions Across the Life Course

Artificial intelligence (AI) companions are entering domains once reserved for human relationships, including infant care, adult friendship, and the care of older people. This study examines whether such systems augment or displace human connection, and what ethical consequences they have for autonomy, emotional well-being, and privacy. Using a qualitative multiple-case design, the study analyses three purposively selected applications that span the life course. First, an evaluation of Nanni AI, an infant-cry interpretation app for parents, is conducted. Second, Replika, a generative AI companion for adults, is evaluated; finally, ElliQ, a proactive social robot for older adults, is evaluated. A corpus of 49 publicly available documents (company materials, media and expert reviews, app-store listings, and user-generated posts) was examined through directed content analysis and triangulated with peer-reviewed literature, using a five-dimensional framework derived from principlist AI ethics, Rawlsian justice, relational autonomy, and attachment theory. All three systems deliver tangible functional support, but their ethical risk profiles differ with the depth of emotional engagement they are designed to elicit and the vulnerability of their users. Replika shows the most pronounced concerns: engagement-oriented design, monetized intimacy, emotional dependence, and weak age verification. ElliQ’s reported benefits rest largely on manufacturer-sponsored surveys, while Nanni AI’s concerns center on accuracy and parental confidence. The study contributes a life-course framework for assessing AI companions and argues that they should be designed and governed as tools that augment, rather than substitute for, human relationships.