Speaker: Camilo Sarmiento, PhD student at LIP6, Laboratoire d'Informatique de Sorbonne Université

Abstract:
Utilitarianism is one of the theories that defend the idea that the consequences of actions are the only parameter for assessing the value of actions. Can this theory be adopted as a decision procedure for an algorithm? Indeed, the machine learning techniques used allow more parameters to be taken into account than a human being could take into account and, a priori, the systems we are talking about are neutral, therefore they have no interest. If this intuition proves to be correct, from a utilitarian point of view, the adoption of these systems would be an opportunity to make our society, a more moral society. In this presentation we will show that it is far from being as simple as that.