Bridging the gap of neuroscience, philosophy, and evolutionary biology to propose an approach to machine learning of human-like ethics

Abstract

The growing explosion of ideas such as Artificial Intelligence (AI), smart environments and ubiquitous computing has led to the creation of the Ambient Intelligence (AmI) paradigm. As AmI begins to take place, moving from a futuristic idea to a reality, we are gradually witnessing the creation of an omnipresent, responsive, and intelligent atmosphere in which thousands of tiny sensors and natural user interfaces will be embedded in our natural movements and in our social and physical interactions. Hence, a key challenge in this multi-disciplinary approach is to get machines to act according to ethical priorities that make sense to human beings. In this study, we improve the capacity for machine ethics to approach human ethics by assessing the computation of transaction values and we argue that it is possible to perform such a computation using recent work that describes the effects of human decision-making using an axiomatic framework. This paper clarifies the relationship between the brain’s 3-axes of neuroscience, the 3 Plato’s Transcendentals of philosophy and the biological evolution’s 3-components, as well as the top-down vs. bottom-up approaches to machine ethics.