人工智能的法律知識

The Promise and Perils of Legal Artificial Intelligence Thomas Coendet

本文嘗試呈現人工智能在法律框架、學術研究和法律程序中的演變角色，探討將人工智能作為自主智能體與將之作為分析人工智能融入法律實踐的前景和風險之工具的區別。通過這個視角，文章調查了一系列旨在增強法律分析、規則制定和決策過程的人工智能工具。此外，文章還批判性地檢視了人工智能作為代理人在法律程序中發揮積極作用的規範含義。本文認為，解決此問題需要理解法律不僅是一個機械系統，而且是人類生活形式的重要組成部分。參與這種生活形式需要承諾、技能、常識、創造力和判斷力等特質，而人工智能無法或只能在有限程度上模擬這些元素。為了說明這些限制，本文展示了一個利用ChatGPT進行的現場測試，對於任何依賴或開發人工智能工具進行法律研究和應用的人來說，理解它們是至關重要的。

This article reflects on the evolving role of artificial intelligence (AI) within legal frameworks, scholarship, and procedures. It introduces the distinction between AI as autonomous agents and AI as tools for analysing the promises and perils of integrating AI into legal practices. Through this lens, it surveys the array of AI tools that aim to augment legal analysis, rule formulation, and decision-making processes. Furthermore, the article critically examines the normative implications of AI assuming an active role as an agent within legal processes. It argues that addressing this question necessitates an understanding of law not solely as a mechanistic system, but as a vital part of the human form of life. Engaging in this form of life demands traits such as commitment, competence, common sense, creativity, and judgment; traits that AI can simulate only to some degree, if any. To illustrate these limitations, the article presents a field-test utilizing ChatGPT; understanding them is paramount for anyone who relies on or develops AI tools for legal research and application.