Why AI will never Be

We’ve come to strongly suspect that there’s a structural, circular causation between cognition and existence. In other words, one can only speak of cognitive beings, and that would actually be somehow redundant: only beings can have cognition; and, among all the worldly self-sustaining structures, only those that have cognition are really beings. Thus, the enactive paradigm in artificial intelligence. But we argue here that letting AI’s roam the world by giving them control over a robotic body will never be enough to create that magic, circular causation.

Computers rely on a meticulous, extremely delicate decoupling of material substrate and logical behavior. We’ve managed to build some quite exceptional physical structures that reflect very precisely our most inherent, structural prejudice about the world we inhabit: logic. We know, though, by our struggling understanding of quantum physics, that if there is such a thing as an objective reality, then that reality does not obey our rules of logic. There have been attempts at rigorously studying quantum logic and henceforth understanding the world in terms of its seemingly inherent, native logic. But those attempts have yet to bear some fruit and, after all, it does looks like they are not really fit to cut through that cake. The essential difficulty with quantum physics is the measurement problem. And, much as some physicists would like to definitively dismiss the strongly compelling suspicion that there should be a link between the measurement problem and the fundamental nature of consciousness, no air-tight argument has been produced to that effect.

So, computers exist because we’ve managed to severe a link in the circular chain that connects cognition and existence. All the interesting stuff that comes out from a computer is structurally detached from the physical substrate that made it possible. Chips work in the digital regime, in which their analog parts are irrelevant. As soon as analog parts of chips become relevant to a computer’s output, that’s because the computer has crushed.

The conclusion is that AI’s will never be. The deep mysteries of an existence that evolves through time, namely sentience, consciousness and agency, are out of the reach of a chip-powered artifact. Those artifacts are structurally bound to reflect our eternal, time-deprived way of making rational sense of our cognitive world. They cannot go beyond a mixture of connectivist and representationalist approaches to artificial intelligence. True enactivism is out of their reach.