How do consumers evaluate AI-generated (vs. human-generated) recommendations?

Team: Hyunjoo Im & Garim Lee

Program: Apparel Studies - Retail and Consumer Studies

Do consumers always believe humans create better boxes than AI? The context-dependent role of creativity in fashion and meal subscriptions.

Purpose – The recent surge of subscription box services calls for research to understand how consumers respond to curation services. This study aims to develop and test a theoretical model to predict consumer response to AI (vs human). Particularly, we tested the role of stereotyping in shaping consumer perception of creativity in this context while considering the contextual moderators, shopping goals (hedonic vs utilitarian) and product category (fashion vs meal).

Design/methodology/approach – Across three studies (total n = 761), we tested assumptions and hypotheses of the study. Preliminary study 1 (n = 511 Amazon mTurk, online survey) confirmed consumer stereotypes of humans and machines. Preliminary study 2, a single-factor between-subjects online experiment (recommender: human vs AI), was conducted at the University of Minnesota (n = 56). The main study was a 2(recommender: human vs AI) x 2(product: fashion vs meal) x 2(goal: utilitarian vs hedonic) between-subjects online experiment (n = 194, Amazon mTurk).

Findings – The results confirmed that consumers are more likely to follow recommendations made by a human more than recommendations made by AI and the perceived creativity of the recommender explained the effect. Significant differences across product categories and shopping goals of the consumers were observed, calling for attention to the context of consumption.

Originality/value – This study extends the understanding of consumers’ responses to recommendations in curation subscription services by highlighting the role of perceived creativity of humans versus AI.