Will a computer ever pass the Turing test?

  • Thread starter Thread starter skip0110
  • 25 comments
  • 1,137 views

Will robots take over the world?

  • Total voters
    29
skip0110

skip0110

Messages
5,178
United States
Worcester, MA
Messages
skip0110
Basically, the Turing test is a check to measure a certain aspect of artifical intelligence. A human judge talks to two subjects, one of which is a computer. (Written conversation is assumed.) If the judge can't figure out which one is the computer, then it passes the test.

(And, you can read about it on Wikipedia, too.)

So, do you think we'll ever be able to make a computer that passes the test?
 
Funny, we’re talking about this in a class I’m taking – and no, it’s not cognitive science or psychology. It’s a writing class. It just so happens that our subject for the quarter is “consciousness”.

I definitely think it’s possible. I believe strongly in the power of reductionism, so I believe that once we have a very strong understanding of the exact neurological processes that go on in our brains, and how those pathways and what not lead to the formation of consciousness, we’ll be able to duplicate it all using computer chips. I’m not saying that it’ll happen anytime soon (I somewhat doubt we’ll even get to the first step – a wholesome understanding of neuroscience – within my lifetime), but long-long term it should be feasible.

And now I need to go write my essay for that class…
 
Sage
I definitely think it’s possible. I believe strongly in the power of reductionism, so I believe that once we have a very strong understanding of the exact neurological processes that go on in our brains, and how those pathways and what not lead to the formation of consciousness, we’ll be able to duplicate it all using computer chips. I’m not saying that it’ll happen anytime soon (I somewhat doubt we’ll even get to the first step – a wholesome understanding of neuroscience – within my lifetime), but long-long term it should be feasible.
Click to expand...
Interesting.

I am always surprised by how much we know about neuroscience. For example, the other day I read a paper in the journal Science which described the mathematical models behind several types of neurons in the brain. They were talking about neurons that do integration, differentiation neutrons, cumulative (adding) neurons, etc. The also suggested a model where memories are formed by a group of neurons whose responses amplify with a certain stimulus pattern.

So from what I grasped from this article/paper is that we are making good strides toward making a complete mathematical/physical/chemical model for the brain. On one hand that makes it look like we will be really close to being able to just simulate this model and end up with human-like thought.

But on the other hand, I think of the massive number of parameters that would have to go into such a model. Evolution has taken thousands (millions, maybe?) of years to work through the various combinations of these parameters and end up with something that has the kind of thought processes we are capable of, so would it take us that long (or longer) as well? (I think so.)
 
I saw something on, i believe' national geographic, where they used parts of a rat's brain and hooked it up to a computer.
After a while the mini-brain could control a plane in a flight simulator.
It actually learned how to do that, nobody programmed it or anything.
The computer sends signal to the brain and the brain sends signals back.

Pretty fascinating :)
 
Here's some food for thought. I heard that for a computer to replicate a brain, it would have to be the size of texas...3 storys tall! Of course, that was probably 18+ months ago, so now it probably has to be "only" the size of, say, Kansas.

Based on your question, the answer is yes if your subscribe to the ideology that anything is possible given time. From a practical standpoint, definitly no.
 
I think complete artificial intelligence is entirely possible. Until we find out it's not possible. Likewise, I don't think it'll ever happen, until some smartass makes it work. I'll leave it to the scientists and software people.
 
Will robots take over the world?

0% - Yes
100% - No
Click to expand...

What are you guys doing!? They have us where they want us!
 
No way. They won't. Within a span of 100 year, technology will be less important than now. People are going to return back to nature. Technology will distroy our planet.
 
philly cheese
Here's some food for thought. I heard that for a computer to replicate a brain, it would have to be the size of texas...3 storys tall! Of course, that was probably 18+ months ago, so now it probably has to be "only" the size of, say, Kansas.

Based on your question, the answer is yes if your subscribe to the ideology that anything is possible given time. From a practical standpoint, definitly no.
Click to expand...

I did a calculation a few months back that, based on the difference in stored memory and processing capacity between a human brain and a top-end PC, using Moore's Law, a home computer will exceed a human's brain inside the next 20 years.

In fact, processing capacity is all we currently have - stored memory went WAY past ours in about 1987.
 
Well if we ever invent AI there will always be a fail safe of some kind that will allow a manual override where someone can take over and stop them. This is actually one of my topics, I know this is going to sound weird but I can see the future becoming a lot like Ghost In The Shell. Where you will have cybernetic bodies and more AI and you can DL any information you want directly to your brain. Could you imagine being asked something you don't know (by someone who isn't cybernetic) and all you have to is DL it and tell them......*starts do day dream*......Man that would be sweet :D
 
Diablomonkey
Well if we ever invent AI there will always be a fail safe of some kind that will allow a manual override where someone can take over and stop them.
Click to expand...

Asimov's laws.
 
Diablomonkey
Well if we ever invent AI there will always be a fail safe of some kind that will allow a manual override where someone can take over and stop them.
Click to expand...
Unplug them?
 
Don't invent (make, produce) them!
 
kikie
No way. They won't. Within a span of 100 year, technology will be less important than now. People are going to return back to nature. Technology will distroy our planet.
Click to expand...

I dont think that will happen. It will be too hard for people just to turn around and reject all their modern luxurys because they think it will destroy the planet.


Global warming will get us before the robots:nervous:



:p


I was also thinking, what would be the motivation for intelligent robots to take over? I would assume they would be content serving us? What else could they do when they took over apart from reproduce? Pointless really...
 
Famine
I did a calculation a few months back that, based on the difference in stored memory and processing capacity between a human brain and a top-end PC, using Moore's Law, a home computer will exceed a human's brain inside the next 20 years.

In fact, processing capacity is all we currently have - stored memory went WAY past ours in about 1987.
Click to expand...

So, are you implying thats all it will take? We throw enough processing power and memory at the problem, and eventually we'll have a machine that can communicate like a human?

I'd guess that we've had the processing power/memory of, say, a mouse, for quite some time now, but I don't think we can fool mice into thinking that a computer is another mouse. Or maybe we can? That would be really cool if we could.
 
This is going to get extremely advanced (and thus very confusing for mere fools such as myself) very quickly, but my thoughts are that we can replicate 'consciousness', as it were, in a computer - but the one thing that would be hardest to program, maybe impossible to get right, is morality.

DE
 
I think it's not a matter of "Can it pass the test?", but a matter of "When can it pass the test?"
 
Robots can't take over the world unless we create them :D.
 
I think it would be awesome (and plausible in the future) to be able to answer questions on the computer about your personality, and it would send out information to another computer to be your friend, with the perfect personality. Imagine a computer that isn't quite as smart as you, so you can still show off to it. :D

Artificial friendship FTW!
 
Dark Elite
This is going to get extremely advanced (and thus very confusing for mere fools such as myself) very quickly, but my thoughts are that we can replicate 'consciousness', as it were, in a computer - but the one thing that would be hardest to program, maybe impossible to get right, is morality.

DE
Click to expand...
I can sort of see where you are coming from.

But if we consider morality to be distinguishing between a good or bad decision (or between good and evil), albeit in a very complex and fuzzy situation, I don't see why a computer would struggle with such a decision if it is capable of forming a natural-language response to an open ended question, the construction of which I would image requires many similar decisions in equally complex and fuzzy situations (except, perhaps the effects of each decision are much smaller in the latter case).
MINICOOPER120
I think it would be awesome (and plausible in the future) to be able to answer questions on the computer about your personality, and it would send out information to another computer to be your friend, with the perfect personality. Imagine a computer that isn't quite as smart as you, so you can still show off to it. :D

Artificial friendship FTW!
Click to expand...
Reminds me of a certain Futurama episode, the one with the robotic Lucy Liu's.
 
I remember learning that new memories and thoughts are made when new brain cells are formed. It would be difficult to actually get it to form new curcuit boards without some kind of DNA to tell it what to do. I could be wrong though, maybe it's just empty, but existing brain cells that form thoughts when electricity passes through them in advanced patterns. I really don't how this all works.
 
MINICOOPER120
I remember learning that new memories and thoughts are made when new brain cells are formed. It would be difficult to actually get it to form new curcuit boards without some kind of DNA to tell it what to do. I could be wrong though, maybe it's just empty, but existing brain cells that form thoughts when electricity passes through them in advanced patterns. I really don't how this all works.
Click to expand...
The computer does not need to behave exactly how the brain does in order to be able to emulate it (and thus pass the Turing test). All we need to do is model that behavior and represent it in 0's and 1's.

Just like a PS2 does not need to have rubber, springs, and gasoline inside it in order to (somewhat) mimic the behavior of a car.
 
Still, those 0s and 1s will need to be in a complex arrangement for it to know how to think, and remember important things while not thinking of unimportant things.
 
MINICOOPER120
Still, those 0s and 1s will need to be in a complex arrangement for it to know how to think, and remember important things while not thinking of unimportant things.
Click to expand...

Oh, yes, certainly.

I was just pointing out that there is no a priori reason (I can think of) for a computer to have the same structure as a brain in order for it to have the functionality of a brain.

Which is not to say that building it like a brain would not help in giving it the functionality of a brain.
 
In reply to the poll, by the time computers become self-aware, the true line between machine and humans will finally be known.
 
Back