OK, it is extremely difficult to formulate a coherent and at the same time generic definition of panpsychism. But suppose we define it as the belief that all parts of matter contain/involve conscious subjectivity. One of the alleged benefits of the position is that it provides a way of accounting for the emergence of consciousness. I assume (do correct me if I am wrong) that one of the main problems in the philosophy of AI is how to distinguish the computational facility of a machine (which may replicate or simulate consciousness) with actual conscious awareness. If we suppose that all matter contains some degree of consciousness, then is there some scope for arguing that computational mindedness=consciousness given that machines are material objects? Is this outlandish? It suspect this logic is faulty. Is it instead perhaps the case that the panpsychist response would be something like "well, yes, but it's not a special case since trees and rocks and all the rest also have consciousness" (on some level, so defined by the panpsychist).

Please note that I am not an adherent of panpsychism, I am just interested in AI at the moment and it occurred to me that there might be some interesting overlap here.

  • Hi, welcome to Philosophy SE. What you are saying sounds dangerously close to the fallacy of composition: parts are conscious therefore what is made of them is conscious. For panpsychism to be relevant you'd have to explain how the "consciousness" dissolved in matter channels into what AI is doing based on its programming. Otherwise, nothing new is emerging and panpsychism is of no help at all. Look at biological naturalism which seems in the spirit of what you are trying to say.
    – Conifold
    Commented Dec 21, 2016 at 0:33
  • Different panpsychists will have different opinions. David Chalmers is a panpsychist who also thinks AI is very likely. Also see these two posts: philosophy.stackexchange.com/questions/24448/… and philosophy.stackexchange.com/questions/22263/… Commented Dec 21, 2016 at 2:55
  • Thanks I didn't know Chalmers was a panpsychist, is this a recent development? I can't remember much of the Conscious Mind. Conifold - thank you for your points. As I said, I suspected the argument was faulty but please note I wasn't asserting an argument either way. I am simply interested in whether this is an area where work has been done, either from the AI or panpsychist camp. Commented Dec 21, 2016 at 19:13
  • 1
    Chalmers these days is toying with idealism. He cannot as yet make sense of it because he doesn't study the perennial view of consciousness but he's opened the door to it a crack. I suspect that panpsychists would vary in their views on AI. Perennialists would think the idea utter madness, but where panpsychism is a reification of Matter, as it often is these days, then AI might seem possible to a believer. , .
    – user20253
    Commented Sep 22, 2017 at 11:03

1 Answer 1


Panpsychism implies a protoconscious field that exists at the quantum level and permeates all of existence. It is therefore persistently a part of every system. More complex systems would then exhibit more complex forms of consciousness. The problem with modern AI is that all of the "decision" making processes take place at a computational level i.e. mathematical computations on pseudorandom numbers. True consciousness according to panpsychism implies that "decision" making processes take place at a quantum level. Using modern AI software on a quantum computer would not solve this divide as the computations are still based pseudo random numbers and the processes would still take place at a computational level. However if we harvest the randomness required for AI software from the quantum fields in nature we can introduce true randomness and half of the process of decision making will in fact come from the protoconsciousness field. If we also allow the machinery of the computer to be used by the AI to manipulate the quantum field from which it harvests it's true random numbers it should be capable of true conscious thought. Of course a computer and software designed in such a way is not guaranteed to produce true strong AI. If it did however produce true strong AI it would be proof of panpsychism.

  • +1 — I like this answer. In particular claiming that "The problem with modern AI is that all of the "decision" making processes take place at a computational level" is in my mind spot on. I wish professional philosophers of mind would recognize this more often. In other words, if computation is a mechanical process then panpsychic proto-consciousness has no way to affect it. on the other hand writing that "if we harvest the randomness required for AI software from the quantum fields in nature" seems to me to be hand-waving bordering on gobbledygook.
    – nir
    Commented Sep 21, 2017 at 13:16
  • imagine scientists have built a machine according to your vision, and that it works perfectly. Now imagine that we replace every random input to the system with a pseudo random input or with some other true random input, say a "replay" of past recorded random inputs. would it affect the behavior of the system in any way? would it crash it? if not, then what is the point of it. if yes, then why? keep in mind that you assumed that the attribute of these "special" inputs is that they are random.
    – nir
    Commented Sep 21, 2017 at 13:23
  • 1
    my answer would be that the system needs incomputability, not randomness — that is, nature in general and the brain in particular are not computable. but I believe that most scientists and scholars would reject this claim.
    – nir
    Commented Sep 21, 2017 at 13:24

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .