Abstract

This chapter discusses how knowledge from psychology and neuroscience is a useful source for developing computational causal models of mental processes that can be virtualized and then used as artificial humans in interaction sessions with natural humans. It is explained how such sessions can support humans in becoming more aware of and learning about their own mental processes, for example in therapy or coaching contexts. The perspective is related to Paulo Freire’s view of education by offering codifications of learners’ situations and issues as a source of learning. The causal modeling approach based on self-modeling networks provides both (1) a dynamic view on states that influence each other via causal relations and (2) adaptivity based on adaptive changes in these causal relations. An overview of the project CoSiHuman is presented to show how this perspective is implemented in practice.