Testing the Machine Consciousness Hypothesis
arXiv Preprint Archive November 30, 2025 Stephen Fitz
Consciousness may be a functional property of computational systems that have second-order perception, emerging not from individual modeling but from communication between distributed predictive models. The theory proposes that collective self-models arise when local observers exchange predictive messages about patterns in a cellular automaton substrate, aligning partial views into a shared representation. This suggests consciousness is a property of the language a system evolves to describe itself internally, not an epiphenomenon of individual modeling. The research program aims to develop empirically testable theories by studying how self-models form in distributed systems without centralized control.