商品製造人責任之研究-以自動駕駛車輛為中心

Abstract

新興科技從萌芽到成熟發展是一段長久的動態歷程，而AI人工智慧應用項目中，無需(或低度)人為介入之自駕車，無非是本世紀最熱門的科技議題之一。 自駕車，一種仰賴鏡頭及感測元件蒐集環境影像、數據，再經過軟體精密運算後給出之行進指令的車輛，駕駛人的角色逐漸淡化，車體硬件結構趨於簡化，電腦運算系統趨於複雜，並要搭配低遲延的網際網路，才能真正實現道路駕駛。 本文首先就自駕車之發展與國際上之分級標準進行介紹，目前產業界對自駕車的定義普遍採用國際車輛工程學會（SAE）J3016 標準，附帶介紹中國最新頒布之GB/T 40429-2021《汽車駕駛自動化分級》國家標準，並說明發展自動駕駛技術後，汽車產業供應鏈出現之變化，點出「軟體定義汽車」之時代來臨。而此種有別於往的交通模式，將對現有法律法規將產生或大或小之衝擊，本文中段圍繞於商品責任法制之檢視，有關商品製造人責任，民法第191條之1採推定過失責任制度，而消費者保護法則採無過失責任制度。然而，驅動自駕車之決策指令高度自動化，甚可能涉及駕駛人(使用人)本身有無定期更新程式，本文認為若僅基於風險分配之說法的要整車廠或系統設計者承擔無過失責任或推定過失責任，對於自駕車供應鏈相關業者而言，恐不夠具有說服力。另建議藉由觀察德國與中國在推動自駕車法案或相關安全標準，以作為台灣之借鏡。 最後，本文嘗試提出兩項建議，一是建立商品安全性評估基準，確保汽車和/或相應系統的整體安全，使製造人有空間主張自駕車符合行業標準規範，生產時已盡防止損害發生之注意；二是導入其他補償制度，例如：由政府主導產業聯盟設置救濟基金，作為特殊事故時之補償，以消弭業者與消費者間之擔憂，使自駕車有機會走向普及。

The development of emerging technologies from infancy to maturity is a long dynamic process. Among artificial intelligence (AI) applications, self-driving cars, also known as an autonomous vehicles (AV), with no (or minimal) human intervention, are nothing but one of the hottest tech topics of the century. The self-driving car is the vehicle in which cameras and sensing elements are applied to collect environmental images and data sending to software for precise calculations so as to give travel instructions. The key for self-driving cars to truly achieve road driving is that the role of the driver is gradually diminished, the hardware structure of the car body tends to be simplified, the computing system of the computer tends to be more complex, and it needs to be equipped with a low-latency Internet. In this article, the development of self-driving cars and the international classification standards are first introduced, including the J3016 standard of the Society of Automotive Engineers (SAE), which is currently widely used in the industry, and the GB/T 40429-2021 standard, which is the latest national standard in China, namely Taxonomy of Driving Automation for Vehicles. In addition, the changes in the supply chain of the automotive industry after the development of autonomous driving technology are explained, pointing out that the era of 'Software-Defined Vehicles' is coming. This different traffic mode will have a certain degree of impact on the existing laws and regulations. Regarding products liability, in Article 191-1 of the Civil Code, the presumption-of-negligence-liability system is adopted, while in the Consumer Protection Act, the liability-without-fault system is adopted. However, the decision-making instructions that drive the self-driving car are highly automated and may involve whether the driver (user) regularly updates the program. In this article, the author argues that it may not be convincing enough for the self-driving supply chain-related businesses to assume liability without fault or presumption of negligence liability based on risk allocation alone. Furthermore, the author of this article also suggests that regarding Taiwan's self-driving car legislation, reference can be made to German and China's promotion of self-driving car acts or related safety standards. Finally, in this article, two suggestions are provided. One is to establish benchmarks for product safety assessments to ensure the overall safety of vehicles and/or corresponding systems, so that manufacturers can claim that the self-driving cars they produce comply with industry standards and they have performed their duty of care at the time of production to prevent damage. The other is to introduce other compensation systems, such as the relief fund set up by the government-led industry alliance, as a compensation in the event of a special accident, which can eliminate the concerns between related manufacturers and consumers and make self-driving cars have more opportunities to popularize.