Drivers of Overtrust in AI among Gen-Z

The increasing use of AI presents many opportunities for Gen Z as they prepare to enter the workforce. In the use of an AI, the Principal (AI user) creates an agentic relationship with the AI. As users employ an AI, they effectively use the AI as an Agent to consider a matter on behalf of the user (Principal). In this process, there is an asymmetry of information that occurs when the Agent, who is closer to the transaction, has more information than the Principal (Eisenhardt, 1989). This information asymmetry allows the Agent to act opportunistically. In the case of a user employing an AI, the Principal expresses their objectives in the prompt, while the Agent has a different set of objectives that are not explicitly stated. These objectives are contained in the reward function for the AI.As AI becomes more capable, Gen Z users are placing more trust in these systems and their output. This increases the potential impact of opportunistic behavior by the AI. Individuals interacting with computer agents tend to form trust faster and trust these agents more than human agents, a phenomenon known as automation bias (Dzindolet et al., 2003; Parasuraman & Manzey, 2010). How much of this trust is warranted? Microsoft researchers implied that we may be trusting in AI generated output too much (Passi, et al., 2025). Goel and his colleagues (2005) defined over-trust as a condition occurring when trust exceeds the level commensurate with the conditions.In this study, we will use the over-trust model (Goel, et al., 2005) to uncover the factors that drive Gen Z’s tendencies to over-trust in AI. We adopt an experimental approach to complete our study. Based on our findings, we will provide insights toward designing AI systems that mitigate over-trust among Gen-Z users.