頁籤選單縮合
| 題 名 | 現階段歐盟與美國對人工智慧規範之比較分析=Comparative Analysis of Current Artificial Intelligence Regulations in the European Union and the United States |
|---|---|
| 作 者 | 陳奕璇; | 書刊名 | 安全與情報研究 |
| 卷 期 | 9:1 2026.01[民115.01] |
| 頁 次 | 頁105-143 |
| 分類號 | 588.36 |
| 關鍵詞 | 歐盟; 美國優先; 人工智慧; 監管性權力; AI韌性; European Union; America first; Artificial intelligence; Regulatory power; AI resilience; |
| 語 文 | 中文(Chinese) |
| 中文摘要 | 歐盟《人工智慧法》2024 年 8 月 1 日正式生效,期間歷經公 開審議、接納各方建議、彙整後由歐盟執委會做出修改,這是公、 私部門相互協商的討論內容。美國總統川普以創新發展為優先, 但美國聯邦政府與州政府之間對人工智慧(Artificial Intelligence, 簡稱 AI)監管立場卻有不同。就管制 AI 系統來說,歐盟以權利 導向 (right-driven) 為基礎,美國則以市場導向 (market-driven) 為 主。目前美國是全球 AI 發展最為快速的國家,歐盟 AI 使用率較 不普及。然而歐盟為何儘早制定《人工智慧法》,目的希望數位 轉型過程預先「管制風險」,防堵 AI 所造成的技術漏洞。歐、 美在 AI 規範設置雖看似立場不同,但實際上仍有部分趨同的現 象。有鑑於此,本文將試圖提出監管性權力 (regulatory power) 與 風險管理及 AI 韌性的概念,用以檢視歐盟、美國的 AI 規範,並 對臺灣 AI 發展提供建議。 |
| 英文摘要 | The European Union (EU) Artificial Intelligence Act will offi- cially take effect on August 1, 2024. This legislation has undergone a period of public consultation, during which inputs from multiple stakeholders were incorporated by the European Commission to revise relevant provisions. The legislative process reflects ongoing negotiations between the public and private sectors regarding arti- ficial intelligence (AI) applications. By contrast, the United States has prioritized AI innovation and development under the policy of the Trump administration, where the federal and state governments have different stances on AI regulations. In terms of regulatory ap- proaches, the EU adopts a rights-driven model, whereas the United States follows a market-driven one. Currently, the United States leads the world in the rapid development of AI, and the EU exhibits lower rates of AI adoption. Nevertheless, the EU’s decision to intro- duce the Artificial Intelligence Act at the early stage of AI develop- ment aims to manage potential risks during digital transformation and to preemptively address vulnerabilities arising from AI technol- ogies. Despite apparent divergences in regulatory philosophy be- tween the EU and the United States, their approaches also converge in several aspects. In light of these dynamics, this study adopts the concepts of regulatory power, risk management, and AI resilience to examine the respective AI governance models of the EU and the United States, in addition to providing policy recommendations for AI development in Taiwan. |
本系統中英文摘要資訊取自各篇刊載內容。