2

To those posting opinionated comments or answers (i.e. not factual research), please first describe your perspective of sentience and artificially intelligent machines - this is not the same as their definitions. Because this question is broad, abstract, and actively debated, readers have their own beliefs of what these terms should mean.


I was reading the Wikipedia article on a philosophical zombie:

[A philosophical zombie] is a hypothetical being that is indistinguishable from a normal human being except in that it lacks conscious experience, qualia, or sentience. When a zombie is poked with a sharp object, for example, it does not feel any pain though it behaves exactly as if it does feel pain (it may say "ouch" and recoil from the stimulus, or tell us that it is in intense pain).

[Philosophical zombies] think they are conscious, think they have qualia, think they suffer pains – they are just 'wrong' (according to this lamentable tradition), in ways that neither they nor we could ever discover.

The excerpt above describes my perspective of what I believe it means for something to be psuedo-sentient (not a real term). It seems to me that an artificial machine (constructed as modern computer programs are constructed, processing instructions sequentially with the ability to conditionally branch its execution) would not be capable of surpassing this quality of psuedo-sentience. What research has been done in this area and what are its answers?

4 Answers 4

3

The naturalistic answer--that is, the answer that you implicitly accept when performing scientific research--is yes.

The reasoning is pretty simple: all nontrivial computational devices are (when abstracted to devices with the ability to access infinite memory) equivalent, and everything in the physical universe is in principle describable mathematically (i.e. as a computation). Even quantum mechanics admits computations which are arbitrarily close.

We've no evidence that this picture (of reality being comprehensible/calculable) is wrong, and a vast and ever-growing pile of instances where it seems to work. As parts of the universe, we therefore are presumably inherently calculable. Since there's no difference in principle (just in implementation detail) between artificially intelligent machines and naturally intelligent humans, AIs can in principle do anything we can, including be sentient. Whether there will ever be AIs that meet the preconditions to be sentient (whatever those are) we don't know. But we do have a pretty good idea now that there's no more magic behind cognition (including sentience, presumably) than there is magical vital force behind life.

1
  • Methodological naturalism does NOT assume computability. Instead it assumes empiricism. It is a rival to your actual POV here, which is rationalism. Naturalists would force you to spell out your assumptions then test them. One is the logical coherence of rationalism. That is refuted by Godel, and logical pluralism.With a falsified assumption, this claim doesn’t get off the ground. Unsurprisingly computability also fails tests. Indeterminate systems are not computable and physics is indeterminate. And even for computers the famous halting problem is not computable.
    – Dcleve
    Jul 16, 2022 at 4:30
3

What research has been done in this area and what are its answers?

There are a lot of people doing research in this area, and as of yet, there are no answers.

Nobody (yet) knows how to make a machine that has qualia, or if such a thing is possible.

If this topic interests you, reading Chalmers (and his critics) is a good place to start.

2
  • Significant progress has been made with projects like this one. Though we have no way (yet) to compare individual artificial neuron behaviors to organic ones.
    – user2411
    Dec 23, 2012 at 11:41
  • Qualia are extremely basic. Even very simple animas writhe in pain. Life therefore has qualia even in living things that do not clearly have sentience.
    – Dcleve
    Jul 16, 2022 at 4:13
0

What research has been done in this area and what are its answers?

See "A computational architecture of cognition, motivation, and emotion. Dörner, Dietrich; Güss, C. Dominik. Review of General Psychology, Vol 17(3), Sep 2013, 297-317."

Doerner is one of the first German psychologists who consider emotion in a setting, which allows to implement emotion in animate beings from artificial life. His computer program is named "PSI". A popular version of his theory is described in his book "Bauplan fuer eine Seele". Unfortunately the book has not yet been translated to English.

If you like, I can send you the paper above in electronic form.

-4

can artificially intelligent machines ever become sentient?

I would actually remove artificially intelligent from that question, because a sentient being need not be intelligent. So the question is really, can machines ever become sentient?

The answer is actually, we don't know. The processes from which sentience arises are not fully understood (okay, more like barely understood). There is the possibility that the processes from which sentience arises are irreducibly complex. If that is the case, then it is not possible for us to create machines that are sentient. The answer would be no.

In theory, however, since we are essentially biological machines it should be possible to create an artificial machine which mimics our state.

1
  • 1
    There is at least one controversial statement here, but no references. Please rewrite your answer from "I think X" to "There are philosophers [ref] who say X". Ideally, provide an overview of the debate, with arguments from both sides.
    – user2953
    Dec 10, 2015 at 7:19

You must log in to answer this question.

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