Towards a Principle-Based Healthcare Agent

Abstract

To feel comfortable allowing healthcare robots to interact with human beings, we must ensure that they act in an ethically responsible manner, following an acceptable ethical principle(s). Giving robots ethical principles to guide their behavior results in their being ethical agents; yet we argue that it is the human designers, not the robots, who should be held responsible for their actions. Towards the end of designing ethical autonomous robots that function in the domain of healthcare, we have developed a method, through an automated dialogue with an ethicist, for discovering the ethically relevant features of possible actions that could be taken by a robot, with an appropriate range of intensities, prima facie duties to either maximize or minimize those features, as well as decision principles that should be used to guide its behavior. Our vision of how an ethical robot assistant would behave demonstrates that an ethical principle is used to select the best action at each moment, rather than just determine whether a particular action is acceptable or not. Further, we maintain that machine ethics research gives us a fresh perspective on ethics. We believe that there is a good chance that this research may lead to surprising new insights, and therefore breakthroughs, in ethical theory.