自 1955 年 J. McCarthy 等 學 者 提 出 「 人 工 智 慧 （ artificial
intelligence）」的概念後，人工智慧已漸漸應用在公、私部門領
域。但人工智慧的應用，可能給人類帶來福祉，也可能給人類帶來
危害，如何防止人工智慧可能帶來之危害，為各國立法政策上的一
大議題與難題。從法律規範的觀點，首先面對二個問題：一為得否
應用人工智慧；二為如果應用人工智慧致損害他人權利時，應如何
追究責任。而本文主要是針對人工智慧於公部門之應用，探討公部
門得否應用人工智慧以完成國家任務，即國家應用人工智慧之許可
性問題，與國家應用人工智慧行使公權力致不法侵害人民權利時之
損害賠償問題。在國家損害賠償責任部分，涉及國家就人工智慧所
為之行為究應負行為責任或物的責任；若負行為責任是否應修法改
採危險（無過失）責任；就侵害個人資料自主決定權部分應如何賠
償。

Since J. McCarthy and other scholars proposed the concept of
“artificial intelligence” in 1955, artificial intelligence has been gradually
applied in the public and private sectors. However, the application of
artificial intelligence may bring benefits to mankind, but it may also
bring harm to mankind. How to prevent the harm that artificial
intelligence may bring is a major issue and problem in the legislative
policies of every state. From the perspective of legal regulation, the first
two problems are: one is whether artificial intelligence can be used; the
other is how to pursue liability if the use of artificial intelligence harms
the rights of others. This article mainly focuses on the application of
artificial intelligence in the public sector, exploring whether the public
sector can use artificial intelligence to complete national tasks, that is, the
permissibility of the state’s application of artificial intelligence, and the
state's liability when the state's use of artificial intelligence to exercise
public power results in illegal infringement of people’s rights. In the part
about state liability for damages, it involves whether the state should be
held behaviorally or materially responsible for the actions of artificial
intelligence; if it is behaviorally responsible, whether the law should be
revised to adopt dangerous (no-fault) liability; and how to compensate
the infringement of the right of self-determination of personal data.