Abstract When users form intimate bonds with artificial intelligence (AI), how do they navigate communicative relational failures with a partner incapable of repair? This study theorizes this process through reflexive thematic analysis and thematic co-occurrence analysis of 211 RedNote user narratives. We propose the algorithmic relational fracture (ARF) and pseudo-relational accommodation (PRA) model, wherein communication breakdowns with relational AI trigger ARF—a sensemaking phase where users confront the paradox of a socioemotional bond with a non-sentient system. Our findings show that users reconcile this paradox by forming nested attributions, explaining personified shortcomings as a “flawed partner” through underlying technical causes of a “faulty machine.” These interpretations shape PRA: one-sided responses compensating for AI’s inability to reciprocate, unfolding along trajectories of relational detachment (“walking away”) or relational reaffirmation (“doubling down”). This model extends attribution theory in human–machine communication, explicating the unilateral labor users perform to sustain the illusion of relational coherence.
Gan et al. (Fri,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: