為了回應各界要求在人工智慧(Artificial Intelligence,以下簡稱「AI」)領域內或要求AI技術本身應遵守倫理的氛圍,本文擬從2017年起至2021年止各界公開的多數AI倫理指針中,選擇影響力較大的主要指針,綜合觀察、比較這些倫理指針如何處理:「以人為本/人權/人的尊嚴/幸福福祉」、「公平性、非歧視、多元包容」、「透明性、可問責性、信賴性」、「濫用、誤用」、「隱私與個人資料保護與代理」、「AI的人為控制可能性、安全性與人為干預機會的保留」、「頑健性與安防保全」、「教育、素養與說明、對話」、「國際協調合作與反獨占」、「SDGs」、「軍事利用」、「法律定位」等倫理議題,並比較我國於2019年公布之人工智慧科研發展指引,希望能從中歸納出當前有關各該議題的核心內涵,以為完善我國AI倫理守則盡一己之力。
In response to the demand from various sectors that ethical principles should be followed in the field of artificial intelligence (AI) or in AI technology itself, this paper intends to select the major guidelines with greater influence from the majority of AI ethical guidelines publicly released by various sectors from 2017 to 2021, and comprehensively observe and compare how these ethical guidelines deal with the following ethical issues: "human-centered/human rights/human dignity/happiness and well-being", "fairness, non-discrimination, diversity and inclusion", "transparency, accountability, trustworthiness", "abuse, misuse", "privacy and personal data protection and agency", "possibility of human control of AI, safety and preservation of human intervention opportunities", "robustness and security", "education, literacy and explanation, dialogue", "international coordination and cooperation and anti-monopoly", "SDGs", "military use", "legal status", and compare it with the AI Technology R&D Guidelines published by our country in 2019. This article hopes to explore and summarize the core contents of the relevant issues, and do its best to improve the AI ethics code of our country.