Advertising with HAART: Introducing Human–AI Asymmetric Relationship Theory
Veronica L. Thomas, Heather Shoenberger, Andrés Gvirtz, Sean Sands, Lotte M. Willemsen · Journal of Advertising · 2026
Generative artificial intelligence (GenAI)is rapidly reconfiguring advertising from message delivery and optimization to relationship-like exchanges between consumers and brands. While existing work has explained these interactions through social response accounts and interpersonal relationship theories, this lens risks overlooking a defining structural mismatch: AI agents convincingly enact relational characteristics but cannot experience relational benefits, costs, or commitment. Perceived mutuality can be high even when the underlying relational capacity for reciprocity is absent. We introduce human–AI asymmetric relationship theory (HAART), which conceptualizes AI-mediated relationships as relationally expressive yet structurally one-sided. HAART advances three tenets: (1) inherent AI agent characteristics limit relational reciprocity; (2) AI agents mask these limits through performative relationship actions, defined as linguistic, paralinguistic, and interactional behaviors that simulate relational engagement; and (3) when this masking succeeds, an asymmetric relationship emerges, characterized by asymmetries in dependence, risk, and control. We develop six propositions that detail the positive (e.g., enhanced support) and negative (e.g., heightened vulnerability) consequences for consumers and brands, and provide implications for theory, practice, and policy.