近年來隨著科技的進步，相關演算軟體及電腦硬體功能提升，藉由大數據的大量演算以及機器學習的技術，人工智慧所仰賴的運算能力也得到大幅進步，可以運用到許多領域。行政院更以台灣的「AI 小國大戰略」的戰略思維，將人工智慧選定為下世代的發展主軸。人工智慧技術快速進展，帶給人類社會許多的好處與便利，高階人工智慧係以模仿人類行為及思考方式的過程，甚至可能發展至具備自主思考及行動的程度。人工智慧與人類間的關係及應用可能性，可能在倫理規範、法律制度、政策執行上帶來風險與衝擊。我們應該去思考如何正確看待人工智慧，特別是當人工智慧在運作過程造成傷害時，應該明確釐清出造成傷害的主體為何。

本研究論文以擬似範疇界定文獻回顧之研究方法(Quasi Scoping Review)，加以探討是否應賦予人工智慧法律人格？以及醫療人工智慧造成傷害的民事責任，應如何分擔？法律人格與民事責任承擔是否有關？基於這些探討，期能對於人工智慧在醫療領域發展所帶來的民事責任的問題，能以更為建設性、具體方式來解決。

筆者認為應賦予強人工智慧類似公司之法律人格，並設立強人工智慧之監管單位，強制要求人工智慧進行註冊；弱人工智慧之責任則由人類承擔。雖然賦予人工智慧一個法律人格的法律主體地位，並不表示人類可以完全將所有責任歸責於人工智慧；惟當賦予其法律人格的法律主體地位時，將可更為明確釐清人工智慧的相關法律紛爭，也能降低人工智慧對人類社會的衝擊，而有助於更為廣泛地運用。

As science and technology makes progress, functions of computing algorithm software and computer software also has improved a lot. Through adopting the technology of big data and machine learning, the computing power of artificial intelligence (AI) also builds up quickly. For the time being, AI has been applied in many fields. The Executive Yuan in Taiwan has chosen artificial intelligence as the development direction of our next generation with the strategic thinking of Taiwan’s "AI Small Economy, Smart Strategy".

The development of AI will bring about many benefits and convenience for human society. The most advanced AI might even imitate human thought process and behavior to reach the extent that we wonder whether AI has autonomy. The relationship of AI and human beings and AI’s applications will lead to risks and challenges in ethical norms, legal institutions and policy implementations.

Using the method of Quasi Scoping Review, this thesis explores the main arguments current literature makes regarding the following three issues: Whether we should recognize AI’s legal personhood, and if yes, how? How we should attribute civil responsibility/liability for the harm caused by AI? What is the relationship between legal personhood and civil responsibility/liability? It is hoped that the results might render more constructive and specific recommendation to resolve the civil responsibility/liability issues in the harms caused by medical AI.

I think strong artificial intelligence should be endowed with the legal personality similar to that of a company, and a supervision unit of strong artificial intelligence should be set up to force artificial intelligence to register; The responsibility of weak artificial intelligence is borne by human beings. Although giving artificial intelligence a legal subject status of personality right does not mean that human beings can completely attribute all responsibilities to artificial intelligence; However, when the person’s personality right is given the legal subject status, the legal disputes related to artificial intelligence can be more clearly clarified, and the impact of artificial intelligence on human society can be reduced, which is conducive to its wider application.