Can I trust my AI friend? The role of emotions, feelings of friendship and trust for consumers' information-sharing behavior toward AI

Abstract

Research background: AI devices and robots play an increasingly important role in consumers’ everyday life, by accompanying the consumer all day long. This presence has several utilitarian and social benefits, but at the same time the optimal functioning of AI requires personal information from the consumer. Purpose of the article: Starting from the premise that people share more information with friends, we have tested empirically whether an emotional behavior of AI can evoke the same emotions in the relationship between consumers and their AI devices, leading to a higher self-disclosing behavior. Methods: To validate the proposed hypotheses, three mediation models were tested using structural equation modelling in Smart-PLS 3.3.3, based on data collected with the help of an online survey.  Findings & value added: We prove empirically that AI’s emotional behavior can increase consumers’ trust, it can evoke feelings of friendship and it can determine a higher perceived control over the shared private information, thus leading to lower perceived threats regarding the consumers’ vulnerability and exposure related to sharing of private data. These results have important implications for designing consumer-AI interactions.