Corporate Interest vs. Public Safety in the Admission of Self-Driving Cars. Evidence from a Conjoint Experiment conducted in the US, Japan, and Germany

Self-driving cars are a critical technology that will be decisive for the prosperity of countries with automotive industry. While manufacturers of autonomous vehicles rely on testing their prototypes in real-world traffic, a public debate has arisen over whether such test approvals pose a threat to other road users' safety. In this study, we present findings from a pre-registered survey-experiment, which directly addresses the question of how citizens evaluate the trade-off between corporate interests and road safety in the admission of prototype self-driving cars. Using a framing design, we first informed our respondents that self-driving cars rely on machine learning which, implies the earlier a manufacturer of self-driving cars succeeds in testing its prototype in public traffic, the earlier this manufacturer can offer a fully developed product that can be capitalized. After that, the respondents received one of the following four statements and were asked to express their support for each statement measured on an eleven-point scale ranging from 0 ("strongly disagree") to 10 ("strongly agree"). The four statements are: (1) To strengthen the domestic automotive industry, self-driving cars should be approved quick and non-bureaucratic [Corporate Interest], (2) The safety of other traffic participants must always come first when approving self-driving cars [Public Safety], (3) To strengthen the domestic automotive industry, self-driving cars should be approved quick and non-bureaucratic, even if this puts the safety of other traffic participants at risk [Corporate Interest vs. Public Safety] or (4) The safety of other traffic participants must always come first when approving self-driving cars, even if this puts the domestic automotive industry at a disadvantage [Public Safety vs. Corporate Interest]. This framing experiment has been conducted on representative online samples in the USA (N=1,188), Japan (N=1,135), and Germany (N=1,174). While all three countries have a substantive automotive industry, the country selection also reflects cultural differences regarding AI and cyber-physical systems. However, for all three countries, we find that triggering public safety diminishes support for the approval of prototype self-driving cars in the cooperate interest frame while triggering corporate interest does not affect respondents' approval in the public safety frame. Furthermore, an in-depth analysis of heterogeneous effects shows that respondents' self-reported level of technophilia moderates the effects of public safety in the cooperate interest frame. In contrast, respondents' partisan orientation does not affect the approval of prototype self-driving cars.