3 ms·

A.I. Is Not Sentient. Why Do People Say It Is?

6 points by zzzeek 4y ago

zzzeek 4y ago
What I find fascinating about the very high level AI researchers who feel that their creations are sentient is not so much that what they see in AI seems to them to be "sentience", but what a dim view they have of human sentience. It seems that people who have rather underdeveloped senses of self and others would be most susceptible to seeing machines as "sentient" since they don't really have much of a view of sentience to start with.
- groffee 4y ago
  I mean when people talk about 'human level intelligence' with machines, we're already there, 99% of people are dumb f*cks. We have stupid people building stupid machines (or not even that, just using 'pre-trained models'), it's not hard to convince themselves and other people that it's 'sentient'.
- IceMetalPunk 4y ago
  On the other hand, there's also a lot of confusion about what sentience is. There's a common conflation between sentience and sapience, there's an overly mystical view of what emotions and human cognition are, there's a habit of thinking of sentience and sapience both as binary rather than a continuum, etc. So while I wouldn't say current AI is sentient yet, I also don't think there will be an AI that's just "suddenly" sentient, and I don't think it's an undebatable fact that current AI isn't sentient or sapient. I think there's an argument to be made that it is, to an extent, just nowhere near to the level of humans. I think it'll just continue to get more and more sentient, more and more sapient, until it becomes full P-Zombie levels of indistinguishable from human-levels of those traits.
blacksqr 4y ago
tl;dr The concept of artificial sentience makes me feel icky, and the people who believe in it are weird, therefore it cannot exist.
- blacksqr 4y ago
  > A neural network is really a mathematical system that learns skills by pinpointing patterns in vast amounts of digital data. And that's different from a brain, because... ?
  - stop50 4y ago
    The "AI"s i saw are oneway streets and don't change unless they are training. If we assume the brain is empty at birth, it is constantly training and working. additionally the brain is from my perspective more complex because it is more meshed and has more input parameters. Some things are even processed before the brain even got them.
ksaj 4y ago
When all of its responses are modeled from input derived from prior work of sentient beings (humans), how could one expect that the output would not appear sentient? It only "knows" how to respond like a sentient being would since 100% of its inputs actually were from sentient beings. Whatever it modeled should look like what it modeled. It's pretty much the definition, or at least the purpose. What blurs the apparent distinction with modern AI is the sheer number of knowledge inputs. But the computer will never be sentient until we figure out what sentience is and how it emerges. Only then can we build something specifically able to do whatever that is. Aside: I just finished reading Blindsight by Peter Watts. It envisions an alien race that lacks sentience, and what that would be like. Its conclusion questions whether sentience is actually a flaw that holds intelligence back. Pretty interesting thought experiment.

6 points by zzzeek 4y ago

What I find fascinating about the very high level AI researchers who feel that their creations are sentient is not so much that what they see in AI seems to them to be "sentience", but what a dim view they have of human sentience. It seems that people who have rather underdeveloped senses of self and others would be most susceptible to seeing machines as "sentient" since they don't really have much of a view of sentience to start with.

groffee 4y ago
I mean when people talk about 'human level intelligence' with machines, we're already there, 99% of people are dumb f*cks. We have stupid people building stupid machines (or not even that, just using 'pre-trained models'), it's not hard to convince themselves and other people that it's 'sentient'.
IceMetalPunk 4y ago
On the other hand, there's also a lot of confusion about what sentience is. There's a common conflation between sentience and sapience, there's an overly mystical view of what emotions and human cognition are, there's a habit of thinking of sentience and sapience both as binary rather than a continuum, etc. So while I wouldn't say current AI is sentient yet, I also don't think there will be an AI that's just "suddenly" sentient, and I don't think it's an undebatable fact that current AI isn't sentient or sapient. I think there's an argument to be made that it is, to an extent, just nowhere near to the level of humans. I think it'll just continue to get more and more sentient, more and more sapient, until it becomes full P-Zombie levels of indistinguishable from human-levels of those traits.

blacksqr 4y ago

tl;dr The concept of artificial sentience makes me feel icky, and the people who believe in it are weird, therefore it cannot exist.

blacksqr 4y ago
> A neural network is really a mathematical system that learns skills by pinpointing patterns in vast amounts of digital data. And that's different from a brain, because... ?
- stop50 4y ago
  The "AI"s i saw are oneway streets and don't change unless they are training. If we assume the brain is empty at birth, it is constantly training and working. additionally the brain is from my perspective more complex because it is more meshed and has more input parameters. Some things are even processed before the brain even got them.

ksaj 4y ago

When all of its responses are modeled from input derived from prior work of sentient beings (humans), how could one expect that the output would not appear sentient? It only "knows" how to respond like a sentient being would since 100% of its inputs actually were from sentient beings. Whatever it modeled should look like what it modeled. It's pretty much the definition, or at least the purpose. What blurs the apparent distinction with modern AI is the sheer number of knowledge inputs. But the computer will never be sentient until we figure out what sentience is and how it emerges. Only then can we build something specifically able to do whatever that is. Aside: I just finished reading Blindsight by Peter Watts. It envisions an alien race that lacks sentience, and what that would be like. Its conclusion questions whether sentience is actually a flaw that holds intelligence back. Pretty interesting thought experiment.