Primum Non Nocere: The Ethical Beginnings of a Non-Axiomatic Reasoning System.

What is the right action? is a question an agent with artificial general intelligence (AGI) will have to face, especially in applications that deal with the health and well-being of human companions. Previous work has considered psychological aspects for forming ethical decision-making; here we consider a philosophy approach and apply abstract, general principles that drive ethical decision-making. Starting with a maxim that has resonated within the health community: “first, do no harm” , we introduce equivalent beliefs and goals to a non-axiomatic reasoning system and examine the sub-goals that are formed. We provide a simple scenario that shows how an AGI system might reason from contrasting, normative ethical theories and how they might combine to provide an ethical framework for AGI to engage in more complex human affairs.