If AI is to be used in practical ways such as medicine, job interviews, or surveillance, it must be able to understand human morals. In order to build real robots in the real world with a neural model of how humans do certain things, an AI system must follow something like Isaac Asimov's laws.