Autonomous driving technologies are expected to affect shared mobility; therefore, the discussions on new business models using shared autonomous vehicles (SAVs) and their feasibilities are underway. This study analyzes consumer preference and willingness to pay for an SAV service, focusing on the automation level of the shared vehicle and liability for an accident, to investigate who will use the SAV service and how much. For this, we conducted a conjoint survey with 1,000 South Koreans and estimated the consumer utility function for an SAV service using a multiple discrete–continuous extreme value (MDCEV) model. The results show that consumer utility increases when the SAV automation level is higher and the liability for an accident lies with the SAV service provider. Moreover, we find that potential consumers have different preferences regarding the liability of a car accident according to their sociodemographic characteristics. We also analyze choice probability and usage of SAV services according to autonomous driving technology levels and who is held liable for an accident. The results show that people with a higher income prefer to use SAV service if service providers are responsible for an accident. Contrarily, older people, drivers, and people with a low income prefer to use SAV service if the accident’s liability lies with the automobile manufacturer. We also find that SAV service usage increases when the level of autonomous driving technology is high or liability for an accident lies with the SAV service provider or automobile manufacturer.

预计自动驾驶技术将影响共享出行；因此，关于使用共享自动驾驶汽车 (SAV) 的新商业模式及其可行性的讨论正在进行中。本研究分析了消费者对 SAV 服务的偏好和支付意愿，重点关注共享车辆的自动化水平和事故责任，以调查谁将使用 SAV 服务以及使用多少。为此，我们对 1,000 名韩国人进行了联合调查，并使用多重离散-连续极值 (MDCEV) 模型估计了 SAV 服务的消费者效用函数。结果表明，当 SAV 自动化水平较高且事故责任由 SAV 服务提供商承担时，消费者效用增加。而且，我们发现潜在消费者根据其社会人口学特征对车祸责任有不同的偏好。我们还根据自动驾驶技术水平和事故责任来分析 SAV 服务的选择概率和使用情况。结果表明，如果服务提供商对事故负责，收入较高的人更愿意使用 SAV 服务。相反，如果事故的责任在于汽车制造商，老年人、司机和低收入者更喜欢使用 SAV 服务。我们还发现，当自动驾驶技术水平较高或事故责任在于 SAV 服务提供商或汽车制造商时，SAV 服务的使用会增加。