Consumer lying behavior to AI agent is widely prevalent in the marketplace and leads to tremendous costs for businesses. Specifically, one stream of relevant research shows that  individuals tend to be more honest with robots than with humans (Kim et al., 2022; Lucas et  al., 2014); yet another stream of research demonstrates that people are more inclined to cheat  when interacting with robots than with humans (Cohn et al., 2022; Kim et al., 2022). We  reconcile prior inconsistent findings by examining the underlying motivations of lie. In particular, we demonstrate that consumers lying behavior in front of AI agent is moderated  by the reward of lie.

The terms material and social have long been used in the motivation literature to distinguish between general classes of outcomes (Eisenberger et al., 1986; Elizur, 1984；Gould, 1979). With respect to definition, material reward involves money, tangible goods that can be directly exchanged in the marketplace (Jenkins and Delbridge, 2020). Social reward, on the other side, involves positive impression, desirable social feedback such as praise, appreciation, and affection (Buss, 1983). Although motivations of lying behavior involves both material and social reward (Hasford et al., 2022), people tend to view lying behavior as either predominantly benefit them in either a material or a social manner.

We propose that the effect of the use of frontline AI agent (as opposed to using humans) on consumer lying behavior is contingent upon the type of lie (i.e., material egoistic lie vs. social egoistic lie). Material egoistic lie depicts situations where people’ lying behavior is chiefly motivated by material reward such as material gains and economic benefits. In contrast, social egoistic lie refers to situations where people’s lying behavior is predominantly driven by social reward such as desired social feedbacks and positive impression. Specifically, we predict that consumers are more likely to tell material egoistic lie to AI agent (vs. human) due to increased moral disengagement; in contrast, consumers are less likely to tell social egoistic lie to AI agent (vs. human) because they perceive AI agent as having a lower level of emotional ability.

Our first set of theoretical contribution is to present an integrative framework to analyze consumer lying behavior towards AI agent by classifying lie into material egoistic lie and social egoistic lie. Previous research shows contradictory findings about lying behavior to AI agent (vs. human) in different scenarios (Kim et al., 2022; Lucas et al., 2014). With the aim of extending prior literature, we are the first to find that the type of lie (material egoistic lie vs. social egoistic lie) affects consumer dishonesty to AI agent (vs. human). In this research, we test the interaction between the type of lie and entity and thus provide novel insight into consumer dishonesty.

Our findings provide vital guidance for companies and organizations to deploy AI agent more effectively. For instance, the implementation of AI agent within the transaction’s procedure is suggested to be augmented with the assistance of human personnel, which lessens consumers’ lying for material rewards. However, in the context of social egoistic lie, such as healthcare diagnosis service, performance-related tasks, and private data collection, it is recommended that the presence of AI agent shall be strengthened, leading to a more frank and upright disclosure from consumers. Therefore, our research helps to mitigate consumer lying in various contexts and thus improve the operative efficiency for current companies and organizations, managers that simultaneously face the challenges that AI agent brings in.