論儒家的「道德主體」與AI作為「人工道德主體」之差異──以陽明的「心」觀念為例的反思

近年由於人工智慧技術的進步，故「人工道德主體」（artificial moral agent, AMA）概念亦受關注，特別是對「主體」一詞的定義學界有諸多爭議；本文將提供學界中贊成、反對與對AMA採取折衷者三種主張的分析，進一步再將其與儒家的「道德主體」進行比較。儒家的主體性主要以陽明的心觀念為例，展示儒家語境中道德存有的屬性；同時，作者亦透過「心之本體」與「人造意識」、「動靜感應之心」與「行為主義主體」、「自省主體」與「自我檢查程序」等向度，檢視前述兩種主體的差異。本文比較的目的，在於避免對AI的探討，不斷走向數位人文等功能性的方向；以哲學而言，人文對AI領域的參與仍應先回歸基礎層次的探討，理解人性主體與機器產品的差異，進而方能承繼人文學對人之所以為人的獨特性的重視，此亦儒家立足於人禽之辨對當代科技的反思。

關鍵字

道德主體；人工道德主體（AMA）；儒家； AI倫理

並列摘要

Due to the progress of artificial intelligence in recent years, the concept of the "artificial moral agent" (AMA) is paid much attention, as the definition of "agent" is much debated in the academic circle in particular. This article would analyze the claims that are respectively agreeable, opposite and eclectic about the AMA, and then compare it with the Confucian moral agent. The Confucian agency is mainly exemplified by Wang Yang-ming's idea of mind, showing the attributes of moral being in the context of Confucianism. In the meantime, the writer would examine the differences between the two types of agency with regards to the "noumenon of mind," "artificial consciousness," the "responsive mind," the "behaviorist agent," the "self-reflective agent," and the "procedure of self-examination." The purpose of such comparison is to prevent the inquires of AI from dwelling in such functions as digital humanism. From the perspective of philosophy, the involvement of humanism in the field of AI should first go back to the fundamental levels, figuring out the distinctions between the human agent and mechanical products; only then we can carry on the humanist attention to the uniqueness of man as man, which is the reflection on contemporary technology by Confucianism, based on the distinctions between human beings and creatures.