|
|
|
|
Department of Computer Science
The George Washington University 801 22nd Street NW, Suite 704 Washington DC 20052 Voice: (202) 994-7181 Fax: (202) 994-4875 E-mail: cs@gwu.edu Seminar/talk SummaryTitlePain and Some Other Dangers AI Poses to Itself and to Society Date2009-11-30 4:00 PM SpeakerMichael LaChat, Ph.D. The Methodist Theological School PlaceRoom 736 Academic Center, CS Conference Room ContentI will begin by recounting the evolution of my own ethical reflection on the construction of a Personal Artificial Intelligence (PAI), i.e. an entity that could pass the Turing Test, albeit with important caveats and clarifications. In this Colloquy, special attention will be paid to the pain, suffering and moral capacities of this “trans-species” Person and to some other possibly deleterious effects of this accomplishment on human society, e.g. problems of human responsibility and autonomy. In my earliest work, I applied standard norms of Western medical codes (e.g. “Above all, do no harm”) to the PAI as subject of a non-therapeutic and potentially harmful/lethal experiment. I intimated that a deliberate, teleological PAI project was unjustifiable by all but the most extreme utilitarian ethical theories, largely because it would be perceived as unnecessary to human welfare. In later writings, I began to see that the myriad applications of AI would begin to blur the distinction between luxury and necessity and postulated that within the moral framework of a liberal, utilitarian cost/benefits analysis the PAI would emerge not as an intentional design, but as a syncretism of advances in service robotics and in medical prosthetics, particularly from those technologies involved in the repair of human brains. As the intelligence and decision-making capabilities of this entity evolve into its “adult phase” it would be granted the same moral autonomy as competent adult human persons only if it had the capacity of empathy, i.e. it would be required to vividly experience and resonate with the pain and suffering of other persons, even those of different “species” – in short, it would have to possess a superego replete with guilt (a form of suffering) as a constraint upon its (possibly metahuman) actions and thoughts. And as it evolves to a more “God-like” phase, its sufferings and moral burdens would become almost unimaginably immense. I will conclude with some imaginative ethical and theological reflection on God and evolution consequent to the achievement of this PAI before turning to Peter Bock for a response leading to audience participation. DetailAttachment: Michael_LaChat_CS_Colloquium.pdf Back |
|
||||||||||||||||
801 22nd Street NW, Washington DC 20052 :: T (202) 994-7181 :: F (202) 994-4875
contact us :: sitemap :: directions :: Email webmaster :: SEAS Homepage |
||||||||||||||||||