A Critical Evaluation of the The Fitting-Recipient Argument for AI Ensoulment Hypothesis

Will advanced artificial intelligence systems one day become ensouled? An affirmative answer to this question, which is an important issue in the philosophy of artificial intelligence, has various metaphysical, ethical, legal, social, and political implications. One of Brian Cutter’s two arguments in his article “The AI Ensoulment Hypothesis” in defense of the possibility that the AI ensoulment hypothesis might be realized in artificial general intelligence is the Fitting-Recipient Argument. In this article, after presenting some preliminary remarks on AGI, Cutter’s argument is explained together with its underlying assumptions and key concepts, and his reasons for the truth of its two premises are formulated and analyzed. It is then shown that the first premise of the argument, which Cutter treats as expressing a sufficient condition, is in fact defensible only as a necessary condition. Accordingly, the argument commits the fallacy of affirming the consequent, that is, treating a necessary condition as a sufficient one, and therefore fails to establish its conclusion. Consequently, ensoulment cannot be inferred from fittingness for ensoulment, and thus the Fitting-Recipient Argument cannot provide a basis for Cutter’s intended conclusion that we should take seriously the possibility of artificial general intelligence becoming ensouled.

کلیدواژه ها:

Artificial General Intelligence (AGI) ، AI Ensoulment ، Fitting-Recipient Argument ، substance dualism ، Philosophy of Artificial Intelligence ، philosophy of mind ، Affirming the Consequent