輔仁法學第61期

刑事定罪後的「附帶效果」係指對經刑事定罪之人，於刑罰之外，所施加經濟、政治、社會上權利的限制。由於，附帶效果既深且廣的影響力，具有可能不亞於刑罰對刑事定罪人的不利益。 因此，美國律師協會（American Bar Association）以及統一法律委員會（Uniform Law Commission）分別針對刑事定罪後附帶效果，提出不同的程序模範法例，以推動刑事定罪後附帶效果的程序立法。我國的行政法規中也具有許多刑事定罪後附帶效果之規定，但現行法中並未對刑事定罪後附帶效果，做出相關討論或程序保障。本文希望透過研究美國刑事定罪後附帶效果制度的發展，檢視我國法制的欠缺，並提出相對應的立法建議。

關鍵詞：附帶效果、褫奪公權、律師協助辯護權、米蘭達告知、量刑

Collateral consequences refer to the economic, political, and social rights restrictions imposed on persons convicted of a crime, in addition to the criminal penalty. Due to the collateral consequences has deep and wide influence, and maybe no less than the criminal penalties of the criminal convict. The American Bar Association and the Uniform Law Commission have both proposed the model codes for the collateral consequences to promote the official legislation for collateral consequences procedural.
There are many collateral consequences provisions in Taiwan's administrative regulations or civil law. However, Taiwan does not have any related discussions or procedural legislation for collateral consequences. This article hopes to study the development of the collateral consequences in the United States, and examine the lack of Taiwan's legal system and propose corresponding legislative proposals.

Keywords: Collateral Consequences, Disenfranchisement, Effective Assistance of Counsel, Miranda Warning, Sentencing

壹、導論
貳、美國刑事定罪後的附帶效果制度分析
一、美國法中刑事定罪後附帶效果之歷史緣由
二、當今美國刑事定罪後附帶效果之規範
三、PADILLA V. KENTUCKY
參、台灣刑事定罪後附帶效果之體系建構芻議
一、我國現有刑事定罪後附帶效果
二、我國相關法制之不足
三、立法建議
肆、結論

所謂功能性原則，係指著作權不保護功能性作品之功能性部分，僅保護作品中能與功能性部分區分之非功能性或藝術性表達部分。許多作品於創作當下，除因實用之考量而具有功能性外，創作者同時於作品中投入個人思想與感情，形成具藝術性或非功能性之表達，導致作品內含功能性與非功能性二項特徵。原按著作權法之原理，作品若含原創性之表達，應受著作權之保護。然而，若未界定權利範圍，恐使著作權之保護及於作品中原為著作權所排除，屬於技術思想而歸專利權保護之功能性部分。此時，功能性原則即扮演框限功能性作品中著作權所及之範圍，確保功能性部分不具可著作性之關鍵角色。
美國作為功能性原則之源流國，於電腦程式、實用物品以及建築物等功能性作品領域之實務方面已累積相當之經驗，且已建構功能性原則應用之準則。近年我國實務於相應領域明顯有引進美國功能性原則及其應用標準之現象，惟文獻或論述甚少談及功能性原則之相關概念。有鑑於此，本文欲從美國關於功能性原則之基本概念出發，觀察美國實務應用該原則所產生之判準、爭議與值得思考之處，於介紹我國實務相應之進展後，藉由比較之方法探尋我國關於功能性原則之問題，進而提出本文之觀點與建議。

關鍵詞：功能性原則、功能性、區分原則、思想與表達區分原則、思想與表達合併原則、必要場景原則、實用功能性、抽象-過濾-比較測試法、觀念上區分、想像測試法

Functionality doctrine, also known as the doctrine of separability, is a principle of copyright law which says the law protects the artistic or non-functional features of a functional work which are separable from the functional aspects of the work. The law’s notion of functionality, in contrast, precludes the functional features of the work from gaining copyright protection. A functional work would unquestionably have its intrinsic function due to the utilitarian consideration. In the creation of the work, the designer would also put her idea and feelings into the work to form the aesthetic or non-functional expressions. Thus, a functional work would have two major features, functional and non-functional. In accordance with the fundamental of copyright law, a work is copyrightable if original expressions contained. If the extent of copyright has not been firstly defined, however, there will be a risk that the copyright protection will extend to the functional aspect of the work which should be secured by patent. In this case the functionality doctrine will play a crucial role to limit the scope of copyright protection, and ensure the functional features of a functional work stay uncopyrighted.
The U.S., as the origin of the functionality doctrine, has accumulated considerable experience in applying the doctrine to the judicial cases on copyrightability of varied functional works, such as computer program, useful article, and architecture. Taiwanese courts have adopted the functionality doctrine for deciding copyrightability of computer program and useful article in recent years. Treatises and literatures in Taiwan, nevertheless, have rarely discussed or even mentioned the concept of the functionality doctrine. In the light of this, the article will review the judicial application of the functionality doctrine in the U.S. after a thorough introduction to the basic concept of the doctrine. Following the observation of Taiwanese courts’ reaction on the functionality doctrine, this article will try to provide some useful comments and workable suggestions by making a comparative study on the U.S. legal experience.

Keywords: Functionality Doctrine, Functionality, Doctrine of Separability, Idea/Expression Dichotomy, Merger Doctrine, Scenes a Faire, Intrinsic Utilitarian Function, Abstraction-Filtration-Comparison, Conceptual Separability, Imagination Test

壹、前言
貳、功能性原則概念之介紹
一、功能性之定義與性質
二、功能性原則之目的
三、功能性原則之背景
參、美國實務就功能性原則之發展與實踐
一、電腦程式
二、實用物品
三、建築物
肆、美國應用功能性原則之發現與思考
一、著作權與專利權競合之爭議
二、功能性原則之挑戰與完善
伍、功能性原則於我國之表現
一、電腦程式
二、實用物品或應用美術作品
三、我國關於功能性原則之檢討與建議
陸、結論

營業秘密之侵害在涉外事件如何定性，頗有疑問。我國營業秘密法立法之主管機關將營業秘密定性為智慧財產權。涉外民事法律適用法（涉民法）第42條（智慧財產權）之立法理由亦將營業秘密定性為該條之單位法律關係。不過，本文考察發現，就國際私法之問題而言，國際上一般解係將營業秘密侵害定性為不公平競爭之問題。營業秘密法之主管機關將營業秘密定性為智慧財產權之根據似亦缺乏合理性。且涉民法第42條所採取之保護國連結，應不適合用於不具屬地性、並非絕對權之營業秘密。因此，本文認為不應將營業秘密侵害定性為智慧財產權之侵害。又我國涉民法第27條就不正競爭行為僅規定了市場地之連結，未若其所參考之瑞士法般設有僅損害特定競爭者時之連結規定。因此，雖營業秘密侵害可定性為不正競爭行為，但解釋上宜認為營業秘密侵害若尚未影響市場競爭秩序（例如僅不正取得而尚無不正使用），仍不適用涉民法第27條以市場地法為準據法，而應依涉民法第25條所定之侵權行為連結方式決定其準據法。也可能將營業秘密侵害直接定性為第25條之侵權行為，視個案狀況並得以市場地法為關係最切之法。

關鍵詞：營業秘密、智慧財產權、侵權行為、不正競爭、保護國

How to decide the applicable law in misappropriation of trade secrets case is an unnoticed problem. In Formosa, the drafter of the Trade Secret Protection Law 1996 regarded trade secrets as intellectual property rights, so does the drafter of Formosan Private International Law 2010. However, this characterization might not be correct. Article 39 of TRIPs cannot serve as a ground for arguing that the trade secrets are intellectual property rights. Nor is the Article 10bis of Paris Convention. In many jurisdictions such as in EU, Germany, Japan, even in ALI Principles of the American Law Institute, trade secrets are not treated as intellectual property rights. Article 42 of Formosan Private International Law is suitable for intellectual property rights with territoriality. However, the protection of trade secrets should not be confined in the territory of a state, and the application of Article 42 will lead to unsatisfactory results. Hence, misappropriation of trade secrets should be characterized as unfair competition. Formosan Private International Law has provision for unfair competition, namely its Article 27. However, different from the EU Rome II Regulation or Swiss IPRG, Article 27 of Formosan Private International Law only envisages the affected market as the connecting factor, it does not have a special provision for the so called bilateral unfair competition. Since in trade secrets misappropriation the market order is sometimes not or not yet affected, or we do not even know where the trade secrets will be used and thus affect the order of the market there, the difficulties in identifying the market location will lead to difficulties in determine the applicable law. This article suggests that the misappropriation of trade secrets can be merely characterized as tort and apply Article 25 of Formosan Private International Law to determine the applicable law.

Keywords: Trade Secret, Intellectual Property, Tort, Unfair Competition, Lex Loci Protectionis

壹、序言
貳、營業秘密本身之準據法
一、概念上之必要性
二、營業秘密可否視為動產
三、營業秘密應否定性為智慧財產權
參、營業秘密侵害之準據法
一、涉民法第42條（智慧財產權）
二、涉民法第27條（不公平競爭）
三、涉民法第25條（侵權行為）
四、強行法規的特別連結
五、依請求性質為不同定性
肆、總結

深度學習 AI醫療的應用目前在醫療放射 (Radiology) 與醫療影像 (Imaging) 均已取得相當優異的成果，其檢測疾病的準確率和速度方面甚至超過了人類醫師。然而，深度學習會產生「黑盒子」(Black Box) 的問題。 當人們輸入數據資料於 AI後，系統經由深度學習的複雜神經網路處理後，再輸出其判斷的結果供醫師參考。然而，AI究竟是如何作出特定的檢測結果或診斷建議？人們卻難以理解其運算邏輯，整個過程就如同一個黑盒子般不透明。在此情況下，如果AI醫療系統出錯而發生誤診，究竟是誰應該負擔民事上的侵權責任？本文即欲探討此問題，深度分析美國及歐盟法制下對於AI醫療的民事侵權責任，並與我國法制相比較，檢視我國民事侵權責任體系該如何調整以面對未來AI醫療浪潮的衝擊。

關鍵詞：AI人工智慧、AI醫療、AI 黑盒子、AI深度學習、民事侵權責任

The application of AI medical treatment has achieved exceptional results in both medical radiology and medical imaging; its accuracy and speed of detecting diseases even exceed that of human doctors. However, AI will have the problem of "Black Box". When people input data into AI, the system processes through complex neural network of deep learning, and then outputs the results of its judgment. However, how does AI make specific results or diagnosis suggestions? It is difficult for humans to understand its operation logic. The entire process is as opaque as a black box. In this case, if the AI medical system is wrong and misdiagnosed, who should bear the tort liability? This issue is worth exploring in depth. This article explores the tort liability for AI in healthcare under the US and EU legal systems, and compares it with Taiwan's legal system, to examine how Taiwan's tort liability system should be adjusted to the impact of future AI medical waves. I will discuss and analyze the relevant literature of the US, the EU and its member states, and Taiwan, including their legislation, amendments to the law, academic articles, and other valuable materials, and propose amendments to the current Taiwan's tort law framework so that it can fully reflect the international trends in the development of AI in healthcare.

Keywords: Artificial Intelligence, AI in Healthcare, AI Black Box, AI Deep Learning, Tort Liability

壹、前言
貳、AI醫療表現已超過人類醫師，未來可能出現「AI醫師」
一、AI在醫療上的應用
二、AI的 深度學習與黑盒子（Black Box）
參、黑盒子（Black Box）AI醫療誤診時涉及的民事侵權責任問題
一、AI誤診的民事侵權責任問題
二、美國、歐盟、及我國法制就AI醫療民事侵權責任之探討
肆、賦予AI電子人格
伍、本文建議
一、重新定義醫師使用AI醫療的過失標準
二、AI醫療製造商不因AI黑盒子的運算錯誤負擔產品責任
三、賦予AI醫療法人格
陸、結論