The Impact of Verbalized AI on Human Cognition and Decision-Making: A Driving-Simulation Experiment with NeuroIS Techniques

This study investigated how the characteristics of AI instructors, specifically voice gender (male vs. female) and audio fidelity (robotic vs. humanoid), affect human cognition, emotional response, and decision-making in a driving simulation. Using a 2×2 factorial experiment with 100 university students, we examined how these variables influence immediate driving responses and subsequent driving performance. The results showed that male synthetic and humanoid voices lead to faster responses and better driving outcomes. Cognitive effort and confused emotions were found to mediate these effects, with robotic voices triggering faster reactions, but poorer long-term retention. This study suggests aligning voice characteristics with user expectations enhances trust, attention, and learning. Practical implications include using male humanoid voices for tasks requiring deep comprehension, and female robotic voices for fast-response situations.