3 ms·
This is the important bit: > I believe that AI systems can, in principle, be “conscious”/“sentient”/“self-aware”/moral agents/moral patients—if only because I
by IshKebab 4y ago
This is the important bit:
> I believe that AI systems can, in principle, be “conscious”/“sentient”/“self-aware”/moral agents/moral patients—if only because I have not seen any compelling arguments that they cannot. Those arguments would require an understanding of the nature of our own consciousness that we simply don’t have.
There are too many philosophers who are insanely over-confident that their wooly evidence-free idea of consciousness is the right one, simply because other philosophers have written a lot about it.
Nobody knows what consciousness is exactly so people proclaiming that this AI can't be conscious are wrong. I don't think it is conscious yet. But that's a different matter.
- mrtranscendence 4y agoWhat does it mean that something "can't" be conscious? A cup of coffee could be conscious -- you can't prove that it doesn't have some kind of conscious experience despite its inability to act independently. But we observe consciousness in ourselves and apparent consciousness in other beings more or less resembling ourselves, and that gives a general idea of what kinds of things might possibly be conscious. There is no indication that some numbers, stored in memory, and an algorithm for applying arithmetic to those numbers to assign probabilities to tokens might be conscious. If it were conscious it would be so significantly different from the kinds of things that seem to have consciousness that at the very least some additional evidence is warranted. I'm as confident that lamda is not conscious as I am that my coffee isn't conscious, so I'll shorthand that by saying that neither of them "can" be conscious. Sue me!
- heyitsguay 4y agoAgreed. One distinction that I've found people have trouble in debates regarding lamda is separating the idea that any billion parameter ML algorithm could be conscious vs lamda in particular being conscious. I think the jury's out on the former (I'd bet you could create a reactive embodied agent with learned environmental goals in <1B parameters given the right setup, which is at least a stepping stone), but for the reasons you describe here, lamda could only be considered conscious by a definition so broad as to declare all digital computation as conscious, and the reasons for that come down to understanding the network architecture and prediction task.
- 2OEH8eoCRo0 4y agoIt's also interesting how many people say, "it's merely acting like a sentient AI" as if an AI researcher at Google hasn't thought of that already. He is either trolling for press or truly believes that this is different than anything else he has encountered.
- gizajob 4y agoIt's strange how a microprocessor is regarded as wholly dead and unconscious when running a spreadsheet or an operating system, but the moment it starts producing outputs in natural language, it somehow has a claim towards sentience due to the confusion caused. I'm a junior philosopher here. I just did an undergraduate dissertation on Wittgenstein and problems for AGI. Much confusion in the field, especially from certain philosophers who couldn't even programme a microwave, nevermind comprehend what a processor is doing at the low level switching bits. We also know absolutely loads about consciousness and sentience. Even a child can understand how we can recognise and properly predicate the word "conscious" to various entities both alive and dead. Philosophers who keep repeating "but we don't have a theory of consciousness, we don't know what it is!" really overlook certain basic facts about the world and our subjective private experience that shows quite simply that we know loads about it.
- hnfong 4y agoThe example of the child only illustrates that we (individually) have a cultural framework of what consciousness means, but it doesn't really mean we "agree" on it. Ask the child, does a dog have consciousness? How about a rat? Butterfly? Ant? Computer (eg. Siri)? Pretty sure people will disagree over those answers. Especially the last one, I suspect. Of course it's tautologically true that people convinced that for people who think they have the answers do have their own answers, but the problem here is that we're not agreeing with each other and don't really have a way to convince each other. I kinda think fundamentally we can't actually do that, because fundamentally it's open for me to claim that all other humans are just fancy automata and I'm the only one with consciousness (in the subjective sense). (And it's empirically true for me too!) But since we're social creatures we mutually recognize the consciousness of other human beings as a matter of protocol. So I don't think approaching it from a scientific point of view really makes sense. If some day AIs become so socially or politically powerful that we have to earn their favor, we'll have to agree that they are conscious beings if they demanded us to do so. This doesn't have anything to do with bits and bytes or transistors unfortunately, it's just purely power dynamics.
- gizajob 4y ago