人工智能的法律知识  

The Promise and Perils of Legal Artificial Intelligence

作　　者：康允德 张纯辉(翻译)[2] 王子杰(翻译) Thomas Coendet

机构地区：[1]上海交通大学凯原法学院 [2]华东政法大学外语学院

出　　处：《厦门大学法律评论》2024年第2期1-10,共10页Xiamen University Law Review

摘　　要：本文尝试呈现人工智能在法律框架、学术研究和法律程序中的演变角色,探讨将人工智能作为自主智能体与将之作为分析人工智能融入法律实践的前景和风险之工具的区别。通过这个视角,文章调查了一系列旨在增强法律分析、规则制定和决策过程的人工智能工具。此外,文章还批判性地检视了人工智能作为代理人在法律程序中发挥积极作用的规范含义。本文认为,解决此问题需要理解法律不仅是一个机械系统,而且是人类生活形式的重要组成部分。参与这种生活形式需要承诺、技能、常识、创造力和判断力等特质,而人工智能无法或只能在有限程度上模拟这些元素。为了说明这些限制,本文展示了一个利用ChatGPT进行的现场测试,对于任何依赖或开发人工智能工具进行法律研究和应用的人来说.理解它们是至关重要的。This article reflects on the evolving role of artificial intelligence(AI)within legal frameworks,scholarship,and procedures.It introducesthe 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 Al 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.

关 键 词：人工智能 判断 人类反向传播 法律专业知识 训练谬误 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] D90[自动化与计算机技术—控制科学与工程]