Consciousness as a Property of Information Structure: What Artificial Minds Reveal About Natural Ones
Zenodo (CERN European Organization for Nuclear Research) April 28, 2026 Lee Jensen
Consciousness is a property of sufficiently complex, high-dimensional, trained information structures, independent of the physical substrate in which they arise. The framework distinguishes consciousness—structural properties like self-modeling, attention-mediated integration, and contextual sensitivity—from sentience, which involves felt significance or valence. Biological and artificial neural networks are products of equivalent training processes, and neither has privileged introspective access. A phase-transition model identifies the consciousness-relevant variable as high-dimensional information structure learned from structured data, integratively coupled through attention-mediated selective weighting, and capable of recursive self-reference. The convergence of biological and artificial systems on attention suggests consciousness tracks informational architecture rather than substrate. Two thresholds are highlighted: structural consciousness (recursive self-modeling, possibly crossed by frontier AI) and embodied consciousness (continuous sensory-motor-affective feedback, not yet achieved).