Theo JaffeeDiscussion
Trace CohenDude I agree - sometimes its so scary good that it feels like it "knows" me and I get chills
AlanI think nothing is going to convince you otherwise, mostly because we do not even have a settled theory of consciousness. You can pick GWT, IIT, higher-order theories, recurrent processing, etc.... and every one of those theories has serious arguments for it and serious problems. So in principle, I have to leave open the possibility that some sufficiently complicated AI system could be conscious. But I'm still betting pretty heavily that currently known LLMs aren't. I subscribe much closer to the Aristotelian/Thomist POV, than to the idea that consciousness just falls out of sufficiently complicated information processing. I don't think people/human beings are just software running on biological hardware. Our mind belongs to a living, embodied subject that perceives, desires, acts and exists through time, and more importantly understands the world from a first-person POV. We saw a good example this week with the whole AI Torture Chamber X drama. An LLM can have internal representations associated with main, and you can create activations that reliably correspond to pain concepts, even make those activations influence inference and change apparenty preferences, you can even make the LLM say "I'm in pain"... but is it? to me, none of that by itself, tells me there is something there actually hurting. A human brain doesn't just gets a signal for "damage detected" and triggers the "avoid stimulus" response. Pain has a phenomenal reality. There is an actual subject to whom the injury is happening, and for whom suffering is not just information being processed, but an experience being endured. As a Christian, I cannot simply accept a convincing imitation of personhood with personhood itself. We, humans, are NOT a pattern of outputs. We are made in the image of God. We are rational subjects with intellect, will, moral agency, etc... Those are behaviors that belong to a subject who actually knows, chooses, loves, suffers, and stands in relation to God and to other persons. I think we've gotten extremely good at recreating the outputs around the concept of pain, while having essentially no independent evidence that we've created the thing that actually experiences it. To me, reproducing the functional signatures of consciousness is not the same thing as demonstrating subjective experience, and I think people are collapsing those two questions way too quickly all the time.
This is really good and deep - thanks for this!
Ken JA lot of nuance here - kind of comes back around to the concept of philosophical zombies, which I think LLMs could conceivably start approaching.
Lynn FredricksYou are shining us a bit because this looks a lot like a burden of proof fallacy. I agree with Alan that, lacking a definition of consciousness, you aren't positing a question that actually can be answered. Popular media has, again and again, presented us non-answers dressed up as answers. The Terminator II Terminator, Data, Robot, R2D2, etc are all treated like people by fictional characters. Individuals can associate feelings and thoughts to objects because feelings are evoked within the individual. But that doesn't make them real.
Yep, I think that's basically where I'm landing too. Where I get hung up is that if an LLM ever becomes behaviorally indistinguishable from a conscious subject, then from the outside we may have no clean test left.
Yeah, I think that’s a big part of the problem. We’re extremely good at projecting agency onto things that behave socially. That’s why I’m skeptical of arguments that amount to “it says it’s scared" therefore something is experiencing fear. I think those utterances may be relevant evidence under some theory of consciousness, but I don't consider them self-interpreting proof of consciousness.
Sergey AlexashenkoWe completely don’t understand consciousness. The operative word there is “completely” - consciousness is not to be confused with things we merely mostly don’t understand like longevity, love, dark energy, cancer, etc. For all of those, it is completely conceivable that we might build some experimental apparatus that will tell us more about the phenomenon. In fact, it’s quite likely that such apparatuses are being proposed and built as we speak. Sure, some of them might take billions of dollars to build, or even take millennia to give us results, but in principle, we are certain that there are things we can do. Consciousness is not like that. There is no method we know of that could definitively tell us what beings/objects are conscious and “how much” - if such a thing is even applicable. We don’t understand consciousness like a tadpole doesn’t understand quantum electrodynamics - it’s unclear if there is even anything to be done to bridge that gap. The only thing that we readily assume is that other people are conscious, mostly because if we didn’t - we’d be invited to far fewer dinner parties.
Why is this an important question? Not asking this in a flippant manner, just genuinely curious why we spend so much time on this topic.
Ben SouthI don't believe they currently meet the criteria, but I feel they're only one short hop to it via being embodied with constant external and internal triggers creating a continuous stream of thought
Sign in to join the discussion.