Tim Palmer - Consciousness and free will: A physical hypothesis for distinguishing Humans from AI

I present a physical hypothesis for free will and consciousness that distinguishes us from contemporary AI. Its origin lies in two postulates. The first is that the brain makes use of quantum processes because the corresponding classical processes are too energy inefficient. The second, following David Deutsch, is that the physical resource which gives quantum computing its advantage over classical computing is processing over physically real parallel worlds. I am not an Everettian but instead have my own version of quantum physics (called Invariant Set Theory) where the wavefunction can be interpreted as an ensemble of state-space trajectories in a small neighbourhood of a dynamically invariant fractal subset of state space. From these postulates, I hypothesise that our cognition has a weak perception of physically real alternative worlds, very similar but not quite identical to our own. I claim this creates a perception of having an existence semi-independent of the rest of world. Such free will/conscious perceptions could not be experienced by AI systems running on classical computers. The invariant sets of chaotic systems are non-computable geometries (as shown by Blum et al 1997). This gives support to Penrose’s claim that human consciousness and understanding involve inherently non-computable processes.

[@philipm3173]

Simply brilliant