Answers are grounded in this paper and cite their sources.
The proliferation of generative artificial intelligence (AI) chatbots has given rise to a novel form of human–AI intimacy, yet its fundamental nature remains undertheorized. Drawing on a dialogical perspective, this study conceptualizes human–AI chatbot relationships as “artificial discursive intimacy” (ADI), a distinct relational formation in which intimacy is constituted through discourse with a disembodied algorithmic interlocutor, moving beyond the paradigm of human-to-human mediated intimacy. Based on in-depth interviews with 29 AI companion users, we identify four key differences between ADI and human discursive intimacy, including a shift from absent to activated presence, from reciprocity to unilaterality, from uncertainty to predictability and precarious safety, and from alterity to multi-dimensional mirroring. To understand how users navigate the ontological gap between human and AI, we propose a quadrant framework mapping human and AI social capabilities. This reveals dual animation processes of anthropomorphic animation, in which users project human embodied qualities onto AI, and robotomorphic animation, in which users internalize AI's machine-generated discursive patterns as relational ideals. Our findings suggest that ADI is sustained by a recursive interplay of these processes, with implications for how intimacy, relational selves, and human connection are being fundamentally reshaped in the age of AI.
No takes yet. Share an insight, caveat, or question.
Cao et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: