An Approach to the Concept of Consciousness from the Positions of Dynamic Theory of Information

The concept of “consciousness”, central in cognitive science, still lacks a clear, established definition. This issue arises particularly acutely in discussion regarding the possibility of consciousness for artificial systems (e.g., GPT). From the perspective of the General Theory of Complex Developing Systems and its specialized branch, the Dynamic Theory of Information (DTI), consciousness may emerge in complex self-organizing systems as a tool for self-diagnosis and for processing internal and external information. There are two key principles for the functioning of any developing system (including cognitive one): the presence of noise (for adaptation to sudden environmental changes) and a homeostasis mechanism (to ensure its integrity). In the human brain, both these factors are provided by subcortical structures, which are also responsible for emotions. In GPT (in its current form), there presents the noise, but not the homeostasis. We propose to define consciousness as an ability to generate, protect, and propagate one’s own subjective information.