This study investigates the impact of three Auditory User Interfaces (AUIs) on takeover performance and mental workload in autonomous vehicles (AVs). As AVs reduce the need for driver engagement, users can focus on non-driving-related activities (NDRAs), which often require visual attention. When a takeover is needed, auditory communication plays a crucial role in ensuring effective information delivery. This research examines how different levels of auditory explanations for takeover requests influence driver response, considering initial trust in automation (TiA). Twenty participants (8 females, 12 males) were categorized into high- and low-TiA groups. Each experienced different AUIs across eight driving trials while performing a typing task as an NDRA in a simulator. Takeover reaction time, glance frequency, and perceived mental workload were measured to evaluate takeover performance and cognitive demand. Results revealed performance differences between user groups, offering insights into trust-aware auditory interaction design for AVs in NDRA scenarios.