Recommending AI based on Quantified Self: Investigating the mechanism of consumer acceptance of AI recommendations

Rapid advancement and widespread use of Quantified Self technology have contributed to the development of artificial intelligence (AI) recommendations. This study is aimed at exploring how users’ perception of the technology affordance of Quantified Self is associated with their intention to accept AI-recommended products. We distinguish four key technology affordances (information richness, personalization, visibility, and metavoicing) of the Quantified Self platform on affordance theory and explore their impact on consumer responses. A one-group design experiment with a pretest was conducted, and 360 participants were recruited to participate in the experiment and fill in the questionnaires. Finally, 344 valid questionnaires were obtained. The findings indicate the following: (1) information richness, personalization, visibility, and metavoicing positively impact purchase intention through trust and behavioral control. (2) Appearance concern positively moderated the relationships between metavoicing and trust and between metavoicing and behavioral control and negatively moderated the relationships between information richness and behavioral control and between personalization and behavioral control. Hence, this study has theoretical significance and practical implications for enriching the knowledge on how to introduce Quantified Self to improve the conversion rate of AI recommendations.