Assessing the technical feasibility of ethical artificial beings

Digital systems are continuously gaining capabilities, are en­visioned as the solution to some of the societal problems and therefore, associated with bigger decision-making responsibilities (for example as autonomous cars). Humans placed in those situations of responsibility face sometimes tough decisions, which they resolve on a daily basis guided by safeguards acquired through education or imposed by professional codes of conduct. With the latest wave of artificial systems demonstrat­ing higher linguistic capabilities, society is starting to ascribe to them some sort of human-like existence, some sort of "Artificial Being". One concern is that these artificial beings, with increasing decision making power in daily-life events may not have a sense of moral duties towards other beings as good as society expects. In this work we look at past considerations of this societal challenge and examine what may take to transfer some sense of moral understanding to artificial beings, which ethical principles we can consider transferring into them. As a result, we identify some technical concepts which seem relevant to embed a level of moral capability in artificial beings. We see this not as a final solution, rather the beginning of a necessary branch of technology.