A double-edged sword of AI in agri-food production: how production methods shape consumers' purchase intention

Purpose Drawing on technology acceptance theory and innovation resistance theory, this study systematically investigates how production methods (artificial intelligence (AI) vs traditional) in agri-food production influence consumers' purchase intention. Design/methodology/approach Three experimental studies were conducted. Study 1 validated the main effect of production methods on purchase intention. Study 2 revealed the dual mediating roles of perceived novelty and perceived naturalness in this relationship. Study 3 examined the moderating role of food atypicality. Findings AI (vs traditional) production methods increase perceived novelty but undermine perceived naturalness of agri-food. Due to negativity bias, the negative path dominates the positive path, leading to an overall negative main effect on purchase intention. Moreover, these effects are moderated by food atypicality: for atypical food, purchase intention for AI-produced agri-food is enhanced via perceived novelty; while for typical food, purchase intention for traditionally produced options is enhanced via perceived naturalness. Originality/value First, this research expands the AI marketing literature by shifting the predominant research perspective from human–AI social interactions to customers' inferences about food's natural attributes. Second, it integrates technology acceptance theories to identify the dual roles of perceived naturalness and perceived novelty in the evaluation of AI-produced agri-food, as well as the dominance of negative perceptions in this evaluation process. Third, this study introduces food atypicality as a boundary condition to reveal differing consumer response patterns.