External Speech Interface: Effects of Gendered and Aged Voices on Pedestrians' Acceptance of Autonomous Vehicles in Shared Spaces

Despite their potential to support multimodal communication between autonomous vehicles (AVs) and pedestrians, speech interfaces are relatively underexplored. Important voice characteristics such as gender and age have been investigated in robots and in-vehicle systems; however, no study has yet compared their impact on pedestrians’ acceptance of AVs through external speech interfaces. Through a within-subjects virtual reality experiment (n=10), we examine five conditions, including synthesised feminine and masculine voices from younger and older adults, alongside a baseline condition without sound. Results show that speech interfaces complement visual cues, generally enhancing pedestrians’ acceptance. Further, acceptance was influenced by gendered and aged voices, with female voices being favoured for their perceived likability, intelligence, and trustworthiness. This research enhances understanding of AV external speech interfaces and aids in developing socially acceptable robotic speech designs. Our future work includes involving more participants from a broader age spectrum and cultural backgrounds.