臺灣與越南人工智慧法制發展之比較研究－以民事責任為中心

A Comparative Study on the Law of Artificial Intelligence for Human Development: Focusing on Civil Liability Applicable to AI in Taiwan and Vietnam

本文以臺越兩國當前關於人工智慧法制發展之研究作為題目，展開實證法之比較分析。本文首先在第一章序論中說明近來人工智慧在經濟、產業、和社會生活上之重要性，其次在第二章闡述人工智慧之概念、演進沿革，並評估其發展對於人類生活在各方面所造成之利弊影響。自第三章以降，本文開始詳細分析臺越兩國對於人工智慧系統之法律定性，發現目前皆將之界定為一種具有市場流通交易價值之物而已，均尚未賦予人工智慧以法律上之人格地位。抑且，本文不但從臺越比較民法之角度出發，分別討論人工智慧占有人、商品製造人責任、危險製造人責任、消保法上之商品與服務責任等規定，適用於人工智慧之侵權事故時所可能發生之法律問題，而且也一併論及人工智慧之創新發展在現行法制上之若干限制，並點出使用大數據與隱私權等個人資料保護之間的緊張關係。最後，本文針對臺越可據以規範兩國人工智慧民事責任之規定，作一比較分析，主張人工智慧商品或服務之法律事故的民事歸責原則厥為：就獨立自主運作之AI所發生之事故，應課予企業經營者以無過失責任、負近乎於結果債務或結果義務，而就使用人可得控制之AI所肇致之事故，應採取過失推定責任、負相當於舉證責任倒置之方法債務或手段義務。

 

"The purpose of this article, which is divided into four parts is to make a comparative study on law related to Artificial Intelligence between Taiwan and Vietnam. Following an introduction relating the important role of AI in the economy and society, Part II explores the development of AI and assesses its impacts on human life. The authors point out not only the positive effects of AI in different fields, but also considerable disadvantages such as increased unemployment, violation of privacy, and intelligence beyond human control. Part III analyzes legal regulations on AI both in Taiwan and Vietnam. This part begins with a discussion on the legal status of AI, followed by inquires into various civil liabilities applicable to AI, like possessor liability, merchandise manufacturer liability, product and service liability etc. Certain limitations on AI invention are presented as it reveals that the Big Data, a source for AI development, can be in serious conflict with privacy protection. Part VI addresses the resemblances and differences between the two countries. It concludes by proposing an imputation principle for harm attributed to AI systems: no-fault liability should be given to business operators for the harm caused by independently operated AI that are beyond human control, while fault liability with a presumption of fault should be imposed for the harm caused by AI operations that are within human control. "