''If it helps then it doesn't matter'': Exploring Non-User Views on Chatbot Companionship through Story Completion

Companion chatbots are conversational AI tools used or designed for emotional and relational support. With the potential for significant benefits as well as risks, they have been uniquely positioned as human surrogates and technical innovations. In this study, we explore perceptions of possible futures for this technology from the perspective of non-users; one goal of examining non-use is to reveal anxieties and assumptions about technology and its role in society, complementing user studies. We conducted a story completion study with 163 English-speaking non-user participants across age groups in which they finished story stems about a socially vulnerable user interacting with a companion chatbot. Despite participants' own disinterest, resistance, or disenchantment to the technology, they largely imagined positive or neutral outcomes for a protagonist they sympathized with and wished well. Overall, participants' stories and reflections were marked by ambivalence about the technology itself and doubt about whether it could be human enough to provide true companionship, and took a practical tone, focusing on the user's needs and various ways the chatbot might meet them. We discuss contributions to scholarship on both non-use and attitudes about chatbot companionship, highlighting the importance of considering moderate, excluded perspectives on controversial emergent technologies.