AI Cracks Jokes: Building Consumer–Brand Closeness Through Machine-Generated Humor

Integrating AI led social cues into service encounters has become a key strategy for humanizing technology. Grounded in Social Penetration Theory, this research examines how AI-generated humor styles (affiliative vs. self-defeating) influence consumer brand identification and proposes consumer personality as a key moderator. To ensure methodological rigor, we employ a multi-study experimental design across four independent studies, each using scenario-based experiments with controlled manipulations and separate participant samples. Results show that AI humor operates as socio-emotional disclosure that strengthens brand connections. Both humor styles significantly enhance brand identification relative to neutral interactions, through the parallel mediating roles of psychological closeness and customer trust. Drawing on Benign Violation Theory to clarify boundary conditions, the findings reveal a moderating effect of consumer personality. Extraverts respond positively to both humor types due to higher stimulation tolerance, whereas introverts show a clear preference for affiliative humor and view self-defeating humor as risky or inappropriate. These results reconcile inconsistencies in human AI interaction research and offer practical insights for leveraging person AI fit to build authentic digital relationships.