The algorithmic mirror: AI disinhibition and anxiety-driven compensatory consumption

Purpose Consumers are increasingly exposed to AI influencers across digital marketing environments, yet little is known about how these virtual agents shape anxiety-driven purchasing behavior. This study aims to propose the AI disinhibition effect, a dual pathway model in which appearance anxiety drives consumption through a fear of missing out (FOMO)-mediated compensatory route common to all influencer types and a direct disinhibitory route unique to AI contexts where social evaluation is absent. Design/methodology/approach Four experimental studies expose participants to human, realistic AI or animated AI influencers presenting products at varying risk levels. Conditional process analyses examine the mediating role of FOMO and the moderating roles of influencer type, visual realism and product risk. Findings Appearance anxiety increases FOMO, which drives compensatory consumption across all influencer types. A direct anxiety to impulsive and compulsive buying pathways additionally emerges in AI contexts, where the absence of social evaluation reduces self-regulatory restraint. This disinhibitory effect is stronger for animated than realistic AI influencers and diminishes as product risk increases, disappearing entirely for high-risk categories where safety concerns restore deliberative processing. Research limitations/implications The studies rely on experimental stimuli and short-term behavioral intentions, suggesting opportunities for longitudinal and field-based extensions. Practical implications AI influencers are most effective for low-risk, identity-expressive categories such as fashion and beauty and considerably less so in high-risk categories such as health or financial products, where consumers engage greater scrutiny. Social implications By identifying conditions under which AI influencer environments amplify anxiety-driven consumption while bypassing deliberative control, this research raises important concerns for consumer well-being and highlights the ethical responsibilities of firms deploying virtual agents in marketing practice. Originality/value This research introduces the AI disinhibition effect and integrates it with compensatory consumption theory. It advances understanding of how influencer type, visual realism and product risk jointly shape anxiety-driven purchasing in AI-mediated marketing environments.