Norm Internalisation in Human and Artificial Intelligence

In this article, principles of architectures relating to normative agents are evaluated with regard to the question whether and to what extend results of empirical research are incorporated in the architecture. In the human sciences, internalisation is a crucial element within the concept of norms. Internalisation distinguishes normative behaviour regulation from mere coercion. The aim of this article is to begin answering the question of to what extent normative agent architectures represent the theoretical construct of norm internalisation. The relevant research in this area may be found in socialisation research in psychology and sociology. Evaluation of conclusions from the empirical sciences allows to identify drawbacks and opportunities in existing architectures, as well as to develop suggestions for future development.