Artificial intelligence and ethics: An exercise in the moral imagination

摘要

Wiener predicted that the quest to construct computer-modeled artificial intelligence (AI) would come to impinge directly upon some of our most widely and deeply held re-ligious and ethical values. It is certainly true that the idea of mind as artifact, the idea of a humanly constructed ar-tificial intelligence, forces us to confront our image of our-selves. In the theistic tradition of Judeo-Christian culture, a tradition that is, to a large extent: our “fate, ” we were created in the imago Dei, in the image of God, and our tradition has, for the most part, showed that our greatest sin is pride-disobedience to our creator, a disobedience that most often takes the form of trying to be God. Now, if human beings are able to construct an artificial, personal intelligence-and I will suggest that this is theoretically possible, albeit perhaps practically improbable-then the tendency of our religious and moral tradition would be to-ward the condemnation of the undertaking: We will have stepped into the shoes of the creator, and, in so doing we will have overstepped our own boundaries. Such is the scenario envisaged by some of the clas-sic science fiction of the past, Shelley’s Frankenstein, or the Modern Prometheus and the Capek brothers ’ R. U.R. (for Rossom’s Universal Robots) being notable examples. Both seminal works share the view that Pamela McCor-duck (1979) in her work Machines Who Think calls the “Hebraic ” attitude toward the AI enterprise. In contrast to what she calls the “Hellenic ” fascination with, and open-ness toward, AI, the Hebraic attitude has been one of fear and warning: “You shall not make for yourself a graven image...” I don’t think that the basic outline of Franl%enstein needs to be recapitulated here, even if, as is usually the case, the reader has seen only the poor image of the book in movie form. Dr. Frankenstein’s tragedy-his ambition for scientific discoveries and benefits, coupled with the misery he brought upon himself, his creation and others-remains the primal expression of the “mad scientist’s ” valuational downfall, the weighting of experimental knowledge over the possibility of doing harm to self; subject, and society.