Comment by slibhb
9 hours ago
There's currently no way to prove that something is or isn't conscious. I would bet that rocks and LLMs aren't conscious. I would bet that humans and dogs are conscious. But I can't prove it. Whenever people try to prove it (in either direction), they end up resorting to sophistry.
This state of affairs could change. If neuroscientists develop a deeper understanding of consciousness, we could talk about it with more precision. Perhaps we could prove that LLMs do or don't possess consciousness. Some neuroscientists claim we already have the requisite knowledge, but I'm skeptical (and there's no consensus).
I actually don't think neuroscience can solve it either. Science could decide whether, e.g. equivalent physical processes are occurring in dogs and humans, and you could argue that solves the question for things with physical brains, but even then it can't measure whether another brain has the same subjective experience as you do. It can't really do that for other humans. It certainly can't answer that question for human askers regarding a different substrate where humans have no ground truth to compare with. The Hard Problem is going to remain Hard.
> I actually don't think neuroscience can solve it either. Science could decide whether, e.g. equivalent physical processes are occurring in dogs and humans, and you could argue that solves the question for things with physical brains, but even then it can't measure whether another brain has the same subjective experience as you do. It can't really do that for other humans
Imagine a device that you put on your head and press "record". After 5 minutes, you press "stop" and then I put it on my head and press "play". I then experience those 5 minutes as you experienced them. You could also replay your own 5 minutes and confirm the recording is accurate.
That device doesn't exist today -- it's sci-fi -- but there's no known law of physics that forbids it. If it is built it would be a lot of progress towards, perhaps even a solution for, the hard problem.
The only realistic way we either build that device or prove that it can't be built is science (probably neuroscience). So, my opinion is that, to a large degree, when we opine about "subjective nature of experience" we're bloviating a bit. Our experience is subjective now but there's no law of physics that says that must always be true.
It wouldn't solve the hard problem. I still couldn't verify that the recording gives you the exact subjective experience it gives me. Yes, I'm leaning into the unfalsifiability but that's kind of the point of why it's Hard. We have this nugget of unfalsifiability at the core of our experience.
Anyway, if someone claimed to create such a machine I would, in fact, very much doubt that it actually creates the same experience simply because no human brain is quite physically the same, so it will interact differently with the machine. That's true even if experience is entirely physical.
> The only realistic way we either build that device or prove that it can't be built is science.
There's another possibility: the device remains in the realm of hazy infeasibility forever, where no one succeeds convincingly but we also never articulate why it's impossible. I think this is more likely. Certainly the engineering would be extremely difficult.
Any candidate device will face the usual objections about the relationship between experience and its physical correlates, plus the one I mentioned above about physical differences between brains, and probably a dozen more depending on the details. You'll be able to choose to believe it proves consciousness is physical, or not, but you can already choose that today with equally strong evidence. It's like a binary Rorschach test for your assumptions about metaphysics.
2 replies →
Strange Days
I've heard an interesting argument about this while going down the rabbit hole of the hard problem of conciousness. Unfortunately can't rember the source - if you can recognize it, do let me know!
Back in the XVII/XIX century, a similar problem existed regarding life - the problem of "what makes living things tick". The assumption at that time was that while we can understand the biological processes around life, we will never understand the so-called "vital force", which causes things to live - life itself. I know it sounds weird now, but back in the day the mental models were different. Phenomenas like "water boils" and "organisms self-replicate" were treated as completely different domains of reality, without an overarching uniform scientific model.
It turned out that after around 100 years, we can figure out the chemical/physical processes and the need for the term "vital force" became redundant.
While this is certainly not an argument proving that the Hard Problem is not in fact hard, it is an interesting idea to think about. Perhaps its all a matter of developing better, higher-resolution neurological models which will at some point give us the tools to decompose qualia.
The problem of subjective is different, because while we might be able to get to a point of being able to say it can't affect anything, pretty much by definition we can't experience another entity's subjective experience or lack thereof.
Even if we were to e.g. identify some field that seemed to coincide with entities reporting a subjective experience, we wouldn't have a way of determining if they truly do, or just act as if they do, nor is it clear such entities would be able to report the difference.
As it is, we struggle to quantify even much more basic differences in experience that we can introspect. E.g. I have aphantasia - I don't see things in my minds eye - and I regularly come across people who insists both that can't be true, and that it can't be true that others see things. And some of the people I've spoken to who insist aphantasia isn't real clearly has it based on digging into their thinking about it.
Even at that level we rely on trusting people's claims about their introspection - we don't know, we assume based on testimony.
Worth keeping in mind, yes, but biological processes are readily observable whereas subjective mental phenomena are... not. Everything about them is inaccessible and frankly unfalsifiable except for the fact that they're the bedrock of all the rest of our observations. Not directly comparable.
Anyway, further reading: https://en.wikipedia.org/wiki/Vitalism I was in fact looking this up recently for fiction-writing purposes.
I agree re: subjective experience. It seems like something inherently unknowable.
What we could do is break down other facets of what we talk about under the umbrella of consciousness, and find measurable subsets.
E.g. a lot of people insists LLMs can't reason either. Coming up with a testable definition of what that means might be doable.
Overall also just separating the subjective experience from the rest leaves us in a position where proving the possibility of AGI "just" rests on whether or not human brains exceed the Turing computable.
If they don't, then subjective experience or not is irrelevant for the question of reasoning and intelligence, as in that case a subjective experience either can't affect the computation or must itself be at least possible to fully simulate by any Turing complete system.
The problem of subjective experience then would largely be down to faith and feelings but would also be entirely orthogonal to the rest.
Yes, our whole concept of "consciousness" is very anthropocentric. There's no particular reason to think subjective experience, moral standing, and general problem-solving ability are inherently connected. (self-nitpick: you could argue moral standing requires subjectivity, but it's still easy to imagine a being with subjectivity that's incapable of suffering and indifferent to its fate)
Simply put, even solipsism is unfalsifiable.
I mean, outside of our own internal experience of qualia, there's no way for us to prove anyone or anything is conscious. But my biggest issue with the debates about consciousness and LLMs is that the "LLMs are conscious" crowd usually has to presuppose a completely materialistic/mechanistic understanding of the universe and of our own consciousness, and that's something that we really don't know for certain. Also, if we were running LLMs on a mechanical or pneumatic or water based computer (albeit very slowly) then the non-electronic computer would be just as "conscious" as the normal electronic computer, and I think it would give a lot of the "LLMs are conscious" crowd pause to think that a bunch of pipes and valves can be conscious. I think there's a lot of magical thinking that arises from the fact that computers are electronic.
> a bunch of pipes and valves can be conscious
Billions * billions of pipes and valves can result in emergent behavior that appears conscious while at the same time the sound of a single independent water pipe can moan and sound like human speech or otherwise lifelike and evoke human emotions.
I think LLMs are doing both of these things and often people are more impressed by the independent fixtures (the moan) rather than the emergent behavior. Both the sound and the emergent behavior can be built on purpose or on accident.
I think it helps to look at this through an Information Theory lens. What information is coming into the system (the human or the machine)? What information goes out of the system which is novel? How much of this can be attributed to attempting to parse random noise aka. `Random_Imagination_Engine` vs something else? The number of inventors who come up with a breakthrough idea after mis-hearing someone is surprisingly high.
If we make the distinction between phenomenal consciousness from access consciousness we can see that LLMs clearly can make decisions based on input (A-Consciousness) but they probably don't have raw feelings and sensations (P-Consciousness).
> What information goes out of the system which is novel?
I think there is very little truly novel information. Most information including the information of "breakthrough idea after mis-hearing someone" is just a mix of previous information.
I guess you are already aware of that... just for completness sake.
You will need immense evidence to go against the entire arc of history where all bets about supernatural or aphysical elements have failed. Until you do so, the brain is another not fully solved physical science problem.
And yes, turing equivalence is turing equivalence, I don't see why a system of pipes can't make an AI.
We only need to assume the hypothesis that brains don't exceed the Turing computable to say that LLMs have the same computational power as a brain.
We can't prove it, but given the total absence of evidence of anything exceeding the Turing computable, as a hypothesis it is a reasonable one that would require truly extraordinary evidence to rise above "magical thinking".
Now, that is also far from proving they are "conscious".
Lets institute a voting system where entities will vote on the other's consciousness. Voting will be done on 8.5 x 11 paper and mailed to the consciousness adjudicator
Seems to me that the burden of proof lies with anyone claiming consciousness. Saying it might be conscious like a human but we don’t know is no more useful than any other conspiracy theory.
I dare you to objectively prove to me that you're conscious.
I am not making that argument.