Rodrigo Vargas
2 min readOct 9, 2021

--

Well, it’s not because of being a creation of humans that I think (current) machines cannot truly be enactive: it’s because their interaction with the environment is mediated by chips, which are artifacts whose behavior is carefully decoupled from the quantum nature of their physical grounding. One could respond that our brain’s behavior, which mediates our interaction with the environment, is based on neurons and all the relevant physics of a neuron activity is classical. That, though, is not quite established—look for the Penrose-Hameroff model of consciousness, in which it is argued that quantum states of the microtubular array are relevant for modelling our brain’s activity. One could perhaps imagine quantum computers roaming around, and those would bypass what I’ve said; but that’s quite into the future, to say the least.

But I don’t want to overemphasize the quantum physics aspect of my argument, although it is certainly very important. Note that enactivism is not just the trivial claim that there’s interaction between us and our environment. Enactivism postulates that “living beings and their environments stand in relation to each other through mutual specification or codetermination” (Varela, Thompson and Rosch). Machines based on chips will never stand in relation to their environment through mutual codetermination. What an AI “perceives” is inevitably mediated by our way of making rational sense of the world: classical mechanics and logic. It interfaces with reality through devices that are structurally built to reflect such worldview, decoupling it from the quantum nature of their physical support. The question arises: is it possible to develop something akin to our intuitive, embodied (as opposed to axiomatic or symbolic) understanding of the world based only on such kind of input? Or is it more plausible that our embodied perception of time, which is so crucial in developing a cognitive worldview and seems to be closely tied with the quantum mechanical nature of our biology, is structurally out-of-bounds for such machines?

--

--

Rodrigo Vargas
Rodrigo Vargas

Written by Rodrigo Vargas

Mathematical physicist turned AI researcher. Knows something about deep learning and the mathematics of quantum fields. Currently works at Ennova Research.

No responses yet