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

Abstract

Recent advances in artificial intelligence have revived a long‑standing assumption: that sufficiently complex computation might give rise to consciousness. This essay challenges that assumption head‑on. I argue that current approaches to artificial general intelligence rely on a structural and functional conception of mind that omits a critical dimension: ontogenetic history. Biological consciousness does not arise from structure alone, but from a continuous developmental process involving embodiment, feedback, and temporal continuity. From this perspective, the duplication or simulation of cognitive functions is insufficient for the emergence of subjective experience. A system may replicate behavior, reasoning, and even self‑modeling, while lacking any phenomenological interior. I further argue that this gap has practical implications. Highly capable but non‑conscious systems may exhibit forms of optimization that are not constrained by experience, leading to a class of risks distinct from those typically discussed in AI alignment. The essay concludes that artificial consciousness, if possible at all, is more likely to resemble the cultivation of a process than the construction of a system — a conclusion with profound implications for how we think about AI, ethics, and our own place in the universe.

Keywords