5 ms·
An interesting thing to me is the unending quest may somewhat end during our lifetimes through us being able to build conscious AI and see how it works.
by tim333 7y ago
An interesting thing to me is the unending quest may somewhat end during our lifetimes through us being able to build conscious AI and see how it works.
- rebuilder 7y agoI doubt it. We can't prove other humans are conscious, I don't see how AI will help clarify anything there.
- mantap 7y agoYou could use an MRI machine to show that their brain activity is the same as yours, and if you know you are conscious then that should be evidence enough. Or you could just ask them.
- OrderlyTiamat 7y agoAn MRI scan, while incredible technology, is an indirect measure of an indirect measure of brain activity, has several seconds delay and a spatial resolution of a couple of millimeters. The "same" brain activity in two mri scans to me doesn't approach the kind of evidence I would be looking for here.
- xamuel 7y agoThis. If building a conscious machine was as simple as building a machine that looks conscious on MRI, then it would be as simple as tape-recording a human brain.
- adwn 7y ago> Or you could just ask them. I can easily make a computer program which, when asked "Are you conscious?", will reply with "Yes", but that doesn't mean that it is conscious. Conversely, you can ask someone who doesn't speak English, "Are you conscious?", and they will only stare at you confusedly, but that doesn't mean that they're not conscious. Therefore, "just asking" is not a reliable method at all to find out whether an entity is conscious.
- mantap 7y agoHow could another human being be not-conscious while displaying all the signs of consciousness? Maybe I have a failure of imagination but if ask my mother "are you conscious?" and she says "yes" then that's pretty conclusive evidence to me. This is actually such a dumb thread I think I'm going to stop now.
- Buttons840 7y agoIt's not dumb, it's philosophy. Philosophy is applying logic to thought. We cannot physically interact with "consciousness", so instead we think about it, put forth ideas, and logically scrutinize those ideas. If the answer to the question "are you concious?" is conclusive evidence, then evidently not all people are concious, because some will answer "no" to that question. Would you agree with this logic - this if-then statement?
- rebuilder 7y agoI am not conscious. Prove me wrong.
- dwaltrip 7y agoIf you are human, you are conscious. Consciousness is to humans in the same way breathing is to humans. There are many other arguments that I'm sure you are aware of. If all of these aren't enough for you, well then, I would direct you towards exploring solipsism.
- rebuilder 7y agoI don't see the relevance of solipsism here. I'm taking essentially the opposite position: I am not conscious. You seem to think I am conscious. How can you tell the difference?
- dwaltrip 7y agoIf you genuinely don't think you are conscious, then I have nothing to say. If you are making a roundabout point of how we can't know if other people are conscious, then my argument about solipsism stands.
- tim333 7y agoI'm thinking along the lines of Feynman's "What I cannot create, I do not understand." If we can build machines with dreams, feelings and the like similar to human ones I imagine we'll have a better understanding of how the thing works. Bit like the difference between philosophers of old pondering how birds fly and how to define the word flying vs aircraft designers who build working aircraft and have degrees in aeronautical engineering.
- mellosouls 7y agoUnfortunately I think this greatly underestimates the qualitative difference of the fundamental nature of the phenomenon of consciousness to anything we have modeled or built before. Sure, we can and will build conscious machines (I believe), but we will do that modelling cognition - and yes, we will develop a better understanding of that. But consciousness is very different and will come with the territory unbidden, and unexplained.
- rebuilder 7y agoHow will we know when we've built such machines?
- mannykannot 7y agoAll that shows is that proof is the wrong criterion, and demanding proof is a way of avoiding the issue.
- goatlover 7y agoThis brings up Ned Block's paper on the Harder Problem of Consciousness in which he uses the example of Data from Star Trek to demonstrate the epistemic bind we are in when trying to decide whether anything non-human is conscious. Another example would be a society of people (China for example). Do all their interactions result in consciousness? How would we every determine this?
- erokar 7y agoHighly unlikely, as we are not nearer conscious machines now than we were 100 years ago. There is nothing about computers that make them more conscious than e.g. radiators.
- tim333 7y agoWell if you Google "conscious", the dictionary definition starts: "aware of and responding to one's surroundings" I'd suggest self driving cars do that better than radiators.
- erokar 7y agoThis topic deserves a little more than a dictionary definition. There's no reason to believe self-driving cars has any more of a subjective experience, qualia, than radiators.
- SmellyGeekBoy 7y agoThere's also no reason to believe that our own "consciousness" is any more than a significantly advanced algorithm. It's a very human trait to put ourselves on a pedestal.
- erokar 7y agoPlease substantiate your claim. You'd need to rely on dualism/metaphysics to support the view that consciousness is just an algorithm. Without some medium to implement it, an algorithm is just an abstract idea. The implementation could in principle be arbitrary. I belive consciousness can be explained within a monist, materialistic worldview. I see it as the result of a biological process, much like digestion. To stretch this analogy (perhaps a tad too far): You can simulate digestion on a computer but it will never produce a real shit for you. I think John Searle got this right and can recommend his book The Mystery of Consciousness.
- cr0sh 7y ago
- lawlorino 7y agoSeen as we still can't even define and measure consciousness how do you think this is going to work exactly? AI is not some kind of magic box despite what a lot of "thought leaders" might lead you to think.
- shrimpx 7y agoI think the best we can do is to build AI that is to us externally indistinguishable from consciousness, or "conscious enough." Which I think is totally doable. But that just represents what we know or recognize as conscious, which can be a very superficial understanding. If it looks like a duck...
- taneq 7y agoWe'll never build conscious AI, purely because we'll never admit that "a mere computer" is conscious. I think we're more likely to inadvertently prove that qualia don't really exist in us any more than they do in computers.
- SmellyGeekBoy 7y agoThe invention of conscious AI could spawn a whole civil rights movement. Is a computer capable of consciousness entitled to the same rights as a human? Obviously this has already been pondered in depth by various sci-fi writers over the years, but it will be very interesting to see if it does happen in our lifetimes.
- Buttons840 7y agoWhat if "a mere computer" ends up being superior to a human mind in every way? We may not believe it is concious, but we might end up believing that we ourselves are also not concious in the process.