Abstract
Social robots are increasingly used in different areas of society such as public health, elderly care, education, and commerce. They have also been successfully employed in autism spectrum disorders therapy with children. Humans strive to find in them not only assistants but also friends. Although forms and functionalities of such robots vary, there is a strong tendency to anthropomorphize artificial agents, making them look and behave as human as possible and imputing human attributes to them. The more human a robot looks, the more appealing it will be considered by humans. However, this linear link between likeness and liking only holds to the point where a feeling of strangeness and eeriness emerges. We discuss possible explanations of this so-called uncanny valley phenomenon that emerges in human–robot interaction. We also touch upon important ethical questions surrounding human–robot interaction in different social settings, such as elderly care or autism spectrum disorders therapy.