The Possibility of Artificial Intelligence as a Moral Agent in Kant's Ethics

Abstract

In this article the possibility of Artificial Intelligence [AI] or machines which are ironically called "intelligent" to be considered as moral agents in Kantian ethics will be questioned. In Kantian sense, there is no possibility for intelligent machines to become moral agents, or more properly, there is no possibility for practical reason to become artificial. This is because in its current form, AI can take a "responsibility" not "from duty/out of duty" as free and autonomous but rather "according to task" imposed on them by their programmers. This responsibility, like the "intelligence" of machines, lacks any real moral content. Consequently, it is first necessary to specify what are the necessary conditions for AI to act as a moral agent. Even if this determination leads to a negative conclusion, the article will raise another question: Does Kant offer the possibility of "interpreting" (active) AI as a being that acts according to purposes? In the last part of the article, the position of AI will be evaluated through Kant's teleological view of nature and beings in nature. Because while the Kantian perspective allows us to know the mechanical process of nature, it also enables us to have a teleological view of all beings.