Employing communication for building learned trust in autonomous vehicles, A qualitative pilot study.

Abstract

As autonomous vehicles (AVs) advance towards full automation, trust between passengers and the automated systems becomes a critical factor to their existence. This research presents a pilot study on developing a design framework to help build and maintain a dynamic learned trust during interactions with fully autonomous vehicles by exploring various information types, structures, and communication modalities. The study draws on insights from previous research, semi-structured interviews with 6 users and 3 experts to identify key design elements that influence trust. Key findings suggest that minimizing driving-related information and instead focusing on journey-related details can prevent passengers from feeling the need to monitor the vehicle’s decisions, thereby fostering trust. The concept of "Information on Demand" emerged as a valuable approach to balance transparency and personalization, allowing passengers to request specific information whenever is needed. Additionally, "technical explanations" were identified as effective in restoring trust when errors occur, emphasizing the importance of timely and clear communication. The research also highlighted the limited impact of non-driving tasks, such as entertainment on trust. Furthermore, communication modalities should be tailored to the type of information being conveyed, taking into account various risks and the passengers’ ability to process different communication methods. This pilot study’s results lay the foundations for a larger scale study aiming to examine various factors that influence the dynamic learned trust during the interaction with the automated system in the vehicles.