This paper discusses the feasibility and potential risks of artificial emotions. It is an era of worrying about the overtaking of artificial intelligence from a cognitive point of view, which is considered to be a human domain, and people are now trying to find out human nature with emotion, not reason. Recently, the task of infusing emotions into artificial intelligence robots is emerging as a new topic. First, let's take a look at the development of emotional robots and their main motivations, and why robots' emotions matter. In order to examine whether a robot possessing true emotions is possible, we do not define emotions a priori, but introduce some important roles that emotions play and propose criteria for giving emotions to certain objects. Against these criteria, true emotional robots are unlikely to become a reality in the near future. However, even before robots possessing emotions appear, unilateral emotional communication with robots with a certain degree of autonomy is potentially dangerous, and it is thought that it is good to prepare for it.

Keywords

artificial intelligence, AI, deep learning, artificial emotion and emotion