본문 바로가기
  • Home

AI, Moral Machine, Good Person

  • Philosophical Investigation
  • 2020, 59(), pp.213~242
  • DOI : 10.33156/philos.2020.59..007
  • Publisher : Institute of philosophy in Chung-Ang Univ.
  • Research Area : Humanities > Philosophy
  • Received : July 15, 2020
  • Accepted : August 17, 2020
  • Published : August 31, 2020

Maeng, Jooman 1

1중앙대학교

Accredited

ABSTRACT

In this paper I will argue that artificial intelligence cannot be a moral machine because it is not an autonomous agent who can create and establish the motive, intention and purpose of the act on his own. The autonomy of the engineering meaning that an anticipated artificial intelligence agent will have, or the autonomy of artificial intelligence, has a fundamental difference from the moral autonomy of the human agent, and it is merely a pseudo-autonomy. So it's just an incredibly convenient or dangerous machine. Only self-creation and self-establishment and self-modification of motive, intention and purpose are the real moral autonomy, whereas artificial intelligence agents have only formal autonomy at best, and such autonomy is no less than the automation of automatic machines which is limited to a limited purpose, and specialized to implement only that. A machine that always performs certain acts is not really a moral machine or a moral agent. Artificial intelligence can never be a moral machine. Moral machines are theoretically and practically impossible. If there was such an existence, it would certainly be an artificial species with other formidable powers beyond human species. Thus, most of those prospects that take place around the current possible artificial intelligence agents are a product of false imagination. Artificial intelligence robots, which are nothing but pseudo-autonomous agents from an ethical standpoint, along with the impossibility of moral machines, must be thoroughly planned and controlled machines in all respects.

Citation status

* References for papers published after 2023 are currently being built.

This paper was written with support from the National Research Foundation of Korea.