Cognitive Robots and the Conscious Mind: A Review of the Global Workspace Theory

Abstract

Purpose of Review The theory of consciousness is a subject that has kept scholars and researchers challenged for centuries. Even today it is not possible to define what consciousness is. This has led to the theorization of different models of consciousness. Starting from Baars’ Global Workspace Theory, this paper examines the models of cognitive architectures that are inspired by it and that can represent a reference point in the field of robot consciousness. Recent Findings Global Workspace Theory has recently been ranked as the most promising theory in its field. However, this is not reflected in the mathematical models of cognitive architectures inspired by it: they are few, and most of them are a decade old, which is too long compared to the speed at which artificial intelligence techniques are improving. Indeed, recent publications propose simple mathematical models that are well designed for computer implementation. Summary In this paper, we introduce an overview of consciousness and robot consciousness, with some interesting insights from the literature. Then we focus on Baars’ Global Workspace Theory, presenting it briefly. Finally, we report on the most interesting and promising models of cognitive architectures that implement it, describing their peculiarities.

Artificial consciousness: the missing ingredient for ethical AI?

“…Their proposed system is based on the Global Workspace Theory (GWT) originally proposed by Baars (1997) , which is, to date, one of the most widely followed theories in the field of consciousness studies. In addition, there are several implementations of it ( Signa et al, 2021 ).…”