查詢結果分析
來源資料
頁籤選單縮合
| 題 名 | 智慧政府資料驅動決策的法治挑戰--以「模仿型」及「相關性發掘型」人工智慧系統為例=Algorithmic Governmentality and Its Challenges to Rule of Law: A Study of "Imitative" and "Correlation-Discovery" AI Systems |
|---|---|
| 作 者 | 邱文聰; | 書刊名 | 中研院法學期刊 |
| 卷 期 | 37 2025.09[民114.09] |
| 頁 次 | 頁27-97 |
| 專 輯 | 「資料治理與以資料進行治理」專號 |
| 分類號 | 312.83 |
| 關鍵詞 | 全自動化決策; 人工智慧; 法治; 目的解釋; 可解釋性; 相關性; 因果關係; 剖繪; 差異製造事實; 行動理由; Fully automated decision-making; Artificial intelligence; Rule of law; Teleological interpretation; Explainability; Correlation; Causation; Profiling; Difference-making facts; Reasons for action; |
| 語 文 | 中文(Chinese) |
| 中文摘要 | 本文探討全自動化決策在公部門的應用,特別關注機器學習型 人工智慧所帶來的資料驅動決策革命,及其對法治理念的衝擊。本 文認為,理解人工智慧的潛力與威脅,需從區分「模仿型」與「相 關性發掘型」人工智慧系統著手,方能在理解不同人工智慧系統的 限制後,找到與法治相容的應用模式,遠離破壞民主基本價值的真 正威脅。本文主張,模仿型人工智慧系統經過足夠的學習後,能在 法律適用上突破目的解釋屏障,而具有模仿各種法律推論的能力; 運用模仿型人工智慧作出全自動化的法律決策,也與現有法治理念 有條件地相容。然而相關性發掘型人工智慧透過發現隱藏的相關模 式,不斷創造出採取新控制介入措施的理由,這種能力在公部門的 應用將導致形式上合理的實質權力擴張,超越現有法治原則所能控 制的範圍。 |
| 英文摘要 | This article explores the issues of the application of fully automated decision-making in the public sector, with a particular focus on the datadriven decision-making revolution brought about by machine learningbased artificial intelligence (AI) and its impact on the principles of the rule of law. The article argues that understanding the potential and threats of AI requires distinguishing between “imitation-based” and “correlation discovery-based” AI systems. This distinction helps to recognize the limitations of different AI systems and to identify applications compatible with the rule of law, thereby avoiding genuine threats to fundamental democratic values. The article asserts that imitation-based machine learning systems, after sufficient training, can overcome the barriers of teleological interpretation in legal application and possess the ability to mimic various forms of legal reasoning. The use of such imitation-based AI in fully automated legal decision-making may be conditionally compatible with rule of law principles. On the other hand, correlation discovery-based AI, by uncovering hidden patterns and continuously generating new reasons for intervention, could lead to an expansion of power in the public sector that exceeds the control afforded by current rule of law principles. |
本系統中英文摘要資訊取自各篇刊載內容。