This study, using an online survey with large samples, analyzes the latent demand for autonomous vehicles in Japan. The analysis is twofold.
First, we applied the choice-based conjoint analysis to estimate the respondents' willingness to pay (WTP) for the auto driving system (conditional automation and full automation) as well as the fuel types (hybrid and electricity) of a car that respondents would buy. We also estimate the factors affecting each of the five respondents' classes grouped by the latent class conditional logit, to elicit the consumer heterogeneity. We find that those who do not favor driving and those who trust the safeness of autonomous driving tend to have higher WTP for automation. Contrast to the preferences to fuel choice, the environmental concern and altruism of the respondents did not affect the selection of automation.
Second, we deal with consumers' attitudes toward the moral dilemma that artificial intelligence (AI) armed in vehicles should face: "the trolley problem" of choosing between two evils, such as running over pedestrians or sacrificing themselves and their passenger to save the pedestrians. We find that, like in the United States, there exists a particular gap between the Japanese consumers' morality and their expected purchasing behavior. Considering it, we alert that autonomous vehicles may cause the social dilemma, and insist the need to pay more attention to this social dilemma when we design the AI algorithm or traffic laws.