Ethics of autonomous information systems towards an artificial thinking

Projects, situated in the crossroads of the cognitive sciences, the neurosciences, the computing and robotics currently concern the creation of autonomous artificial beings capable of thinking. This article presents a model of comparison of the human thinking with a numerical thinking that relies on four hierarchies: the classification of information systems, the cognitive pyramid, the linguistic pyramid and the numerical information hierarchy. After a state of the art on the nature of the human thinking, the feasibility of autonomous multi-agents systems endowed with an artificial consciousness and capable of thinking is discussed. The ethical aspects and the consequences for the humanity of such systems are evaluated. This subject causes at present the reaction of the scientific community.