Attribution of blame of crash causation across varying levels of vehicle automation
Attribution of blame of crash causation across varying levels of vehicle automation
复制标题
DOI：
10.1016/j.ssci.2020.104968
复制
发表时间：
2020-12-01
期刊：
影响因子：
6.1
通讯作者：
Prabhakharan, Prasannah
中科院分区：
文献类型：
--
作者：
Bennett, Joanne M.;Challinor, Kirsten L.;Prabhakharan, Prasannah
虽然公众对自动驾驶汽车的看法在很大程度上是有利的，但在发生碰撞时，人们对责任的担忧不断出现（Kyriakhovsky等人，2015年）。关于涉及自动驾驶汽车的碰撞事故的法律的责任的复杂性存在争议，意见从司机总是负责到不可能让“自动驾驶司机”负责。虽然在不同级别的车辆自动化的碰撞中“谁应该负责”一直存在争议，但公众对自动化车辆碰撞中的责任归属知之甚少。为了更好地了解这些可能如何影响信任和采用这些技术，本研究旨在了解公众对涉及不同自动化水平的碰撞责任的看法，以及这种责任的感知后果。共有129名年龄在19岁至61岁之间的本科生（M = 24.6，SD = 7.64）阅读了四个小插曲，其中详细描述了行人碰撞场景，每个小插曲中操纵的车辆自动化水平（手动驾驶，部分自动化，高度自动化和全自动驾驶）。参与者被问到三个开放式问题：“你把责任归咎于谁？'，'根据你在哪里分配责任，你会采取什么行动？“，以及”今后如何防止这一事件？'.结果显示，参与者将责任归咎于六个利益相关者类别（司机，行人，汽车，政府，制造商和程序员）。随着自动化程度的提高，指责司机的参与者比例下降，而指责制造商的参与者比例增加。参与者通常将针对驾驶员、制造商或两者的法律的行动确定为他们的行动方针。这一比例因自动化水平而异，当车辆完全自动化时，针对“驾驶员”的法律的行动仍然存在。此外，随着自动化水平的提高，要求审查、改进或完全避免自动化的呼声也越来越高。这项研究的绝大多数结果强调，公众认为车祸的最终责任在于人类驾驶员，而不是“自动驾驶员”。这些发现对公众的信任和自动驾驶汽车的采用率都有影响。此外，它强调了围绕涉及自动驾驶汽车的碰撞结果加强治理和法律的框架的必要性。
Whilst public opinion towards automated vehicles is largely favourable, there are recurrent concerns around responsibility in the event of a crash (Kyriakidis et al., 2015). There is debate about the complexities regarding the legal responsibility of crashes involving automated vehicles, with opinions ranging from the driver is always accountable to the impossibility of holding an 'automated driver' responsible. Whilst "who is responsible" for a crash at different levels of vehicle automation has been debated, little is known about public opinion around the attribution of blame in automated vehicle crashes. In order to better understand how these might impact trust and adoption of these technologies, the present study aimed to understand public perceptions of responsibility for crashes involving different levels of automation, and the perceived consequences of that responsibility. A total of 129 undergraduate students, aged between 19 and 61 (M = 24.6, SD = 7.64) read four vignettes which detailed a pedestrian crash scenario with level of vehicle automation being manipulated in each vignette (manual driving, partially automated, highly automated and fully automated driving). Participants were asked three open-ended questions; `Where do you assign blame?', `Based on where you assign blame, what course of action would you take from here?', and `How could this event be prevented in the future?'. Results revealed that participants attributed blame to six stakeholder categories (driver, pedestrian, car, government, manufacturer and programmer). As automation increased, the proportion of participants who blamed the driver decreased, whilst those blaming the manufacturer increased. Participants commonly identified legal action against the driver, the manufacturer or both as their course of action. The proportions varied across level of automation, with legal action against the `driver' still identified when the vehicle was fully automated. Furthermore, as level of automation increased, there were increased calls for automation to be reviewed, improved or completely avoided. Overwhelmingly the findings from this study highlight that the public believe that the ultimate responsibility for a crash is in the hands of the human driver, rather than the `automated driver'. These findings have implications for the public trust and rates of adoption of automated vehicles. Further it highlights the need for greater governance and legal frameworks around the outcomes of crashes involving automated vehicles.