人工智慧（Artificial Intelligence, AI）及其潛在應用發展極為迅
速。然而，廣泛擴張的人工智慧技術應用，也帶來了相關技術、法
律及倫理挑戰，相關問題包括：演算法的公平、透明性及演算法的
安全性及穩固性等。而歐盟近年所提出的《人工智慧法》（草案）
旨在建立通用之人工智慧監管架構，且其影響範圍並不限於歐盟境
內（域外效力），並隨著相關規則之提出逐步建構完整之通用人工
智慧規範。
本文為歸納當前在人工智慧治理上所面臨之特定挑戰，監管機
構必須從人工智慧的發展技術中針對創新和可信找到平衡。依據現
有文獻分析，在人工智慧技術的規範上往往面臨「人工智慧之精確
定義」、「可解釋性及透明度」及「不可預測之風險」等相關難
題；故本文將首先梳理歐盟對人工智慧法制之發展模式、規範內涵
及歐盟監管態度，並提出個別須關注之法制上議題，最後再就如何
合法、安全應用人工智慧進行整體監管架構之分析。

Artificial intelligence (AI) and its potential applications are
developing extremely rapidly. Nevertheless, the widespread application
and use of AI create technical, legal and ethical challenges. These issues
range from fairness of algorithms, transparency, to the security and
robustness of AI algorithms. EU’s AI Act (bill) aims to establish the first
comprehensive regulatory scheme for artificial intelligence, and the
impact of the regulation will extend beyond the EU’s borders. With the
proposed regulation to establish harmonised rules on the use of AI.
The purpose of this paper is to indicate and discuss selected
challenges facing individuals and institutions responsible for creating
regulations regarding the use and implementation of AI. Regulators
should, therefore, find a proper balance between innovation and
credibility. The analysis of available literature allows us to identify at
least 3 important challenges facing regulators: providing a precise
definition of AI, explicability, transparency and the unforeseeable risks of
AI. This paper attempts to analyze a legal regulation system which places
the protection of AI applications as its core.