Blaming automated vehicles in difficult situations

Abstract

Summary Automated vehicles (AVs) have made huge strides toward large-scale deployment. Despite this progress, AVs continue to make mistakes, some resulting in death. Although some mistakes are avoidable, others are hard to avoid even by highly skilled drivers. As these mistakes continue to shape attitudes toward AVs, we need to understand whether people differentiate between them. We ask the following two questions. When an AV makes a mistake, does the perceived difficulty or novelty of the situation predict blame attributed to it? How does that blame attribution compare to a human driving a car? Through two studies, we find that the amount of blame people attribute to AVs and human drivers is sensitive to situation difficulty. However, while some situations could be more difficult for AVs and others for human drivers, people blamed AVs more, regardless. Our results provide novel insights in understanding psychological barriers influencing the public's view of AVs.

Do We Blame it on the Machine? Task Outcome and Agency Attribution in Human-Technology Collaboration

“…The self-serving bias widely identified in these studies (e.g., [46], [47]), that is, to attribute more responsibility of a car accident to the AVs than to self, was only partially present in our findings. Especially in our loss scenario, participants assigned the same, or slightly higher, level of responsibility to themselves.…”

“…Some researchers have named the phenomenon "blame attribution asymmetry" [45], such that AVs are blamed more harshly, and blame and responsibility are more likely to be attributed to the AVs. Blaming the AVs is observed experimentally and in real life, for harms that are actual or imagined [46]. Our research shows the psychological mechanism through which users with certain personality characteristics (i.e.…”

Blaming Humans and Machines: What Shapes People’s Reactions to Algorithmic Harm

“…We hence do not generalize our results to other scenarios. For instance, scholars have investigated public expectations of self-driving cars [7], and recent work has inquired how laypeople react to crashes through the lens of blame [41]. Different research fronts are necessary to understand how people react to algorithmic harm.…”

Human injury-based safety decision of automated vehicles

“…However, the complete avoidance of crashes remains unattainable, confirmed by the hundreds of AV-involved crashes reported by the California Department of Motor Vehicles ( California DMV, 2020 ). Based on these valuable AV crash reports, many studies focused on analyzing AV collision frequencies, crash types, and associated contributing factors ( Xu et al., 2019 ; Boggs et al., 2020 ; Kutela et al., 2022 ), evidencing that traffic crashes can still occur in the future due to the perception failures of ego vehicles or the irrational actions of other traffic participants ( Bonnefon et al., 2016 ; Wang et al., 2019 ; Franklin et al., 2021 ). Yet, when confronting inevitable obstacles, little attention has been given to AV decision-making systems that manage to reduce the severity of an imminent collision.…”