汽车在全自动驾驶状态下的损害赔偿规则与安全性能的选择

作者 日引聪(顾问研究员)、新熊隆嘉(关西大学)、吉田惇(九州大学)
发表日期/编号 2022年10月 22-J-035
研究课题 使人工智能更为社会接受的制度设计
下载/链接

概要

When autonomous driving vehicles, which are currently under development, become widespread in society, liability for accidents and the amount of damage caused will be attributed solely to vehicle performance. For this reason, there are two possible solutions to accidents: a damage liability rule and a product liability rule. Shavell (2020) focuses on the decision-making of automobile drivers under a damage liability rule and found that strict liability rules do not optimize car use and development of the technology for safety performance and thus in order to achieve the first best solution, he proposes a new liability rule (hereafter referred to as the Shavell rule) where parties bear their own damages and pay the government for the victim’s damage. This study extended his model to take into account the development of the technology for safety performance by firms. Our main findings are (1) when the utility function is the same across individuals, the safety performance that is developed under the Shavell rule would be excessive, and a vehicle purchase tax based on safety performance or a technology development tax should be introduced to discourage the excessive incentive to develop the technology, (2) in the case where fair premium damage insurance is available and the premium is determined based on automobile use, the same policy as is in (1) needs to be introduced, (3) when the utility function is different among individuals, the incentives to develop the technology may need to be strengthened by a vehicle purchase subsidy based on safety performance or a technology development subsidy in some cases, (4) Under a product liability rule, the vehicle-use tax and a technology development subsidy need to be implemented.