On the evolutionary cognitive pressure for experiential awareness: do machines need it?

The paper tackles a foundational question in AI ethics: do machines need to be conscious to be intelligent? The authors sidestep the philosophical debate over what consciousness is, focusing instead on whether experiential awareness—the subjective feeling of experience—is computationally required for higher-level reasoning. They examine why this awareness evolved in biological organisms from a computational perspective. Their conclusion: it is necessary because of evolutionary constraints like autonomous neurological reactions that interfere with pure reasoning. Biological brains had to integrate awareness to manage those low-level processes.

For artificial systems, which are architected without such evolutionary baggage, there is no computational need for experiential awareness. They can achieve arbitrary levels of intelligence simply through better algorithms and hardware. This has major implications: it suggests that future superintelligent AI may not be conscious at all, which could dramatically simplify ethical considerations (e.g., rights, suffering). It also proposes new empirical tests for distinguishing conscious from non‑conscious systems by looking for signs of that evolutionary legacy.