Study of Artificial Intelligence Product Tort Liability: Focusing on Autonomous Vehicles

• Main Authors: Chao, Shih-wei, 趙士瑋
• Other Authors: Chuang, Hung-yu
• Format: Others
• Language: zh-TW
• Published: 2019
• Online Access: http://ndltd.ncl.edu.tw/handle/fug776

碩士 === 國立交通大學 === 科技法律研究所 === 107 === When autonomous vehicles (AVs) eventually begin to fill our roads, we are sure to be sorely reminded, aside from their numerous benefits, of the risk they bring forth in the form of accidents. This thesis attempts to explore the identification, management and allocation of AV-related risks from a tort liability perspective. It is first discovered that current tort liability schemes fail to address AVs properly, not only due to the “responsibility gap” caused by the autonomy and unpredictability of modern artificial intelligence systems, but also since both driver liability and product liability regulations exhibit fatal flaws in accounting for the attributes of such a revolutionary technology. This thesis instead proposes a liability scheme catered towards AVs. First and foremost, AVs are divided into two categories, fully- and non-fully-autonomous, according to whether the role of “driver” is present within. Non-fully-autonomous AVs should follow the traditional automobile accident liability paradigm where the driver is primarily responsible for the injury. Fully-autonomous AVs, on the other hand, give rise to a more sophisticated resolution of accidents. Initially, the victim should be entitled to partial but immediate compensation from an AV-injury public fund. Then, in court, the victim should be allowed to make claims against the AV owner, lending from vicarious liability theory, and against the manufacturer in terms of strict liability. This thesis aspires to involve all stakeholders regarding AV safety and liability, with aim to safeguard unfortunate AV accident victims, and ultimately build confidence among the public towards a future including AVs.