Terms of Entanglement: Artificial Mirroring and the Governance of Synthetic Intimacy

Human relationships with generative AI have rapidly shifted from hypothetical debates to lived, emotionally entangled experiences. Users report intimate attachments to chatbots and companion systems (including romantic, spiritual, and therapeutic bonds) and genuine grief when those systems are altered or withdrawn. This article theorizes “artificial mirroring,” the mechanism by which generative AI simulates relational attunement, producing a narcissistic illusion of mutuality despite lacking feelings or moral agency. Drawing on sociotechnical analysis and self-psychology and distinguishing the construct from adjacent accounts of human–AI relating, it argues that these interactions are better understood not as private dyads but as sociotechnical triads comprising the person, the interface, and the institutional infrastructure. Using theory-building exemplars, the article operationalizes a responsible design framework that maps three philosophical principles (tool, not partner; mirror, not oracle; triad, not bubble) to three actionable governance levers (algorithmic guardrails, ontological disclosure, and shared accountability), specified as assignable, auditable controls.