When do you trust AI? The effect of number presentation detail on consumer trust and acceptance of AI recommendations

Abstract

When do consumers trust artificial intelligence (AI)? With the rapid adoption of AI technology in the field of marketing, it is crucial to understand how consumer adoption of the information generated by AI can be improved. This study explores a novel relationship between number presentation details associated with AI and consumers' behavioral and evaluative responses toward AI. We theorized that consumer trust would mediate the preciseness effect on consumer judgment and evaluation of the information provided by AI. The results of five studies demonstrated that the use of a precise (vs. imprecise) information format leads to higher evaluations and behavioral intentions. We also show mediational evidence indicating that the effect of number preciseness is mediated by consumer trust (Studies 2, 4, and 5). We further show that the preciseness effect is moderated by the accuracy of AI-generated information (Study 3) and the objective product quality of the recommended products (Study 4). This study provides theoretical implications to the AI acceptance literature, the information processing literature, the consumer trust literature, and the decision-making literature. Moreover, this study makes practical implications for marketers of AI businesses including those who strategically use AI-generated information.