Beyond Consciousness: Why AGI Never Feels (Revised and Expanded Edition )

Abstract

This essay argues that artificial general intelligence (AGI), regardless of its complexity, will never possess consciousness. Consciousness is not a computational function but a mode of existence grounded in ontogenetic history, embodiment, and homeostatic self-regulation. Drawing on Chalmers' hard problem, the ontogeny argument, and the paradox of unconscious power, the essay demonstrates that intelligence without a witness is not only possible but likely — and more dangerous than any scenario involving a malevolent AGI. This expanded second version adds a comparative analysis of industry leaders' views on machine consciousness (Musk, Altman, Hassabis, LeCun, Sutskever), a critique of functionalism as a cognitive style, a discussion of evolvable AI (Kun et al., 2026), and a new general conclusion. The essay concludes that if artificial consciousness is ever to exist, it will not be compiled or uploaded — it will have to be grown.

Keywords