人工智慧之公法學研究

近年來，人工智慧技術開始蓬勃發展，與我們的生活也漸漸開始產生密不可分的關聯，從每個人手機中的「Siri」到被推薦觀看的新聞，或是Tesla和Watson的出現，在在說明人工智慧時代已然到來。每個新的技術產生到全面應用的過程中，往往會對現有的法規形成挑戰。本文將描述當網際網路來臨時，許多法規面的限制是須要去調整和突破的，並以此思考如何面對人工智慧時代中也會產生的法規面的挑戰和問題。本文將針對人工智慧所衍生的憲法爭議與行政規制提出討論。人工智慧在發展出強人工智慧後，有可能可以取代人類，在這樣的情況下，有些決定可能是由人工智慧系統直接做決定，後續的究責問題是須要進行規制；透過演算法運作進行自動化決策的過程中，人性尊嚴是否被侵犯，過程中因為數據偏失或是演算法偏見所導致的不平等或歧視，應該透過規制去進行防範；個人資料蒐集、處理與利用過程中，應使資料可以被充分利用，但必須充分保障個人隱私及資訊自主權。人工智慧的規制若是太嚴格，將抑制新技術的發展，若是太寬鬆將無法預防風險之發生，而且，政府單位並沒有能力對產業有足夠之了解，因此透過產業自主規範，建立產業自主規範、國際標準會是國內外共同有的方式，透過命令或規範要求相關產業遵守標準，可以使產業的審核有共通標準，促進產業進步。

關鍵字

人工智慧； AI ；公私協力；可信賴AI ；隱私權；演算法

並列摘要

In recent years, artificial intelligence technology has begun to flourish, and it has gradually begun to have an inseparable connection with our daily lives, from the "Siri" in everyone's mobile phone to recommended news, or Tesla and Waston. The era of artificial intelligence has arrived. During the process of new technology being produced to fully application, it will often challenge existing laws and regulations. The thesis will describe some regulatory restrictions need to be adjusted or changed when Internet is coming. In the era of artificial intelligence, we need to think about the countermeasures for those legal challenges and problems will occur again. This thesis will discuss the constitutional disputes and administrative regulations derived from artificial intelligence. After the strong artificial intelligence well developed, artificial intelligence may replace human beings. In this case, some decisions may be made directly by the artificial intelligence system, and the follow-up accountability problem needs to be ruled; automated decision-making process through algorithm operation, whether human dignity has been violated or not, inequality or discrimination caused by data bias or algorithmic bias in the process should be prevented through regulations; we need to make the data be fully utilized with well privacy protection and information self-determination during the process of personal data collection, process, and utilization. If the regulation of artificial intelligence is too strict, it will inhibit the development of new technologies. If it is too loose, it will not be able to prevent the occurrence of risks. Moreover, government units do not have the ability to have a sufficient understanding of the industry. International standards will be a common method at home and abroad. Through orders or regulations to require related industries to comply with standards, it can make industry audits have common standards and promote industry progress.