Toward a Theory of Intelligent Complex Systems: From Symbolic AI to Embodied and Evolutionary AI

Abstract

In the twentieth century, AI (artificial Intelligence) arose along with Turing’s theory of computability. AI-research was focused on using symbolic representations in computer programs to model human cognitive abilities. The final goal was a complete symbolic representation of human intelligence in the sense of Turing’s AI-test. Actually, human intelligence is only a special example of problem solving abilities which have evolved during biological evolution. In embodied AI and robotics, the emergence of intelligence is explained by bodily behavior and interaction with the environment. But, intelligence is not reserved to single organisms and brains. In a technical coevolution, computational networks grow together with technical and societal infrastructures generating automated and intelligent activities of cyberphysical systems. The article argues for a unified theory of intelligent complex systems.