Thinking Machines

Abstract

This chapter investigates ethical questions surrounding the possible future emergence of self-aware artificial intelligence (AI). Current research into ethical AI and how this might be applied or extended to future AI is discussed. It is argued that the development of self-aware machines, or their functional equivalents, is possible in principle, and so questions of their ethical status are important. The importance of an objective, reality-based ethics in maintaining human-friendly AI is identified. It is proposed that the conditional nature of life and the value of reason provide the basis of an objective ethics, whose implications include rights to life and liberty, and which apply equally to humans and self-aware machines. Crucial to the development of human-friendly AI will be research on encoding correct rules of reasoning into AI and, using that, validating objective ethics and determining to what extent they will apply to and be followed voluntarily by self-aware machines.