Teaching Ethics to Autonomous Systems – Do androids dream of being good
people?
Intelligent systems are going to be increasingly present in our lives, from autonomous vehicles to systems that advise judges in their decisions, to banks to grant a loan, etc. In short, they are systems that will have to make decisions with a strong ethical burden. In this situation, one of the questions that arise is: Can we teach ethics to machines? When asked this question, we hear the two opposite answers with equal ease: of course Not, of course Yes. In this talk I intend to clarify in which sense each of these two answers is partially true. I will explain to the uninitiated what artificial intelligence is, and in particular how a computational system can modify its behavior by learning from the environment, or from the behavior of human beings. I will then explain the possibilities and difficulties of «programming» the ethical behavior in a machine. In the end, the most important reflection is: What we can learn about the ethics of human behavior when we look at ourselves in the mirror of machines.
Gonzalo Génova – Carlos III University of Madrid