人工智能是道德代理人吗——人工智能伦理身份的形而上学分析

Is Artificial Intelligence a Moral Agent?——a metaphysical analysis of the ethical identity of artificial intelligence

人工智能的伦理身份是人工智能伦理面临的基本哲学问题。人工智能伦理身份研究的道德行动者路径通常以人格化的道德行动能力作为人工智能配享伦理身份的标准。这种自上而下的研究路径把人工智能获得伦理身份视为技术问题,有预设身份之嫌。只有以现象学方法直面人工智能的伦理身份本身,才能澄明其伦理身份为何。通过现象学还原的方法自下而上的分析人工智能的行动结构,可以发现人工智能实际上代理了人类道德意向,人工智能的行动以“守护”人类价值为宗旨。人工智能的真实伦理身份是人类的道德代理人。只有以道德代理人为基本单位,人类价值为形而上学基础,一种以人为本的人工智能伦理体系才得以建立。

The ethical identity of Artificial Intelligence(AI)is a basic philosophical problem the AI Ethics facing.The path of moral actors in the study of the ethical identity of AI usually takes the personified moral action ability as the criterion for AI to possess the ethical identity.This top-down approach regards the acquisition of an ethical identity by artificial intelligence as a technical problem,with the suspicion of a preset identity.Only by confronting the ethical identity of AI with phenomenology method can the ethical identity of AI be clarified.Through the bottom-up analysis of the action structure of AI through phenomenological reduction,it can be found that AI actually is the agent of human moral intention,and the action of AI takes“guarding”human value as their purpose.The true ethical identity of AI is the moral agent of human beings.Only by taking the moral agent as the basic unit and human value as the metaphysical basis can a human-oriented AI ethical system be established.