Information-Processing Theory Proposed for Implementing Consciousness in AI

A new paper develops an information-processing theory of consciousness and, through it, proposes a way to implement consciousness in AI. This theory aims to explain what adaptive advantage consciousness provided during the evolutionary process of living organisms.

The core theory, the Subject-Object Emergence Theory, argues that consciousness evolved to enable organisms to achieve adaptive coordination between the body and the environment without extensive trial-and-error learning.

The key mechanisms are as follows:

• Subject-Object subsystem: Uses depictive/iconic visual representations that indicate the relative position of the body and the environment.
• Adaptive coordination: These representations 'light up' for the subject as subjective experience, enabling complex coordination tasks to be performed without reinforcement learning.

The research emphasizes that conscious function is a key factor in enhancing human learning efficiency, and explores a path toward AGI (Artificial General Intelligence) by transplanting this into AI systems.