AI-enabled social support chatbot usage: flowing ambivalence and liminalities
Hajer Kéfi, Insaf Khelladi, Zied Mani, Nathalie Veg-Sala
Abstract
Interest in social and emotional support chatbots has recently surged, making human – chatbot relationships increasingly common. However, users’ subjective experiences with these chatbots often extend beyond simple interactions, reflecting the complex dynamics of liminality and ambivalence. Through a netnographic study of the chatbot Replika, we explore how users experience relational liminality, and control and agency liminality. These dynamics contribute to what we term flowing ambivalence, where users feel both comforted and unsettled, fostering dependency on chatbots despite an awareness of their artificial empathy. Our findings suggest that emotional support chatbots provoke complex emotional states that fluctuate and adapt, underscoring the need for nuanced frameworks to understand how users relate to AI tools.