Artificial Intelligence

[jasmine]

Just wondering what all you techies (there are a few here) think about AI. A new Turing Test is about to take place:

http://www.guardian.co.uk/technology/2008/oct/05/artificialintelligenceai

I think that as long as a computer is programmed it's not intelligent or conscious - it's the programmer that's conscious and has provided the options and the intelligence that fools the questioner.

However, were a chip to be invented that allowed a none-programmed machine to think would that be consciousness and would the machine have rights?

[Trystan]

I'm not really a transhumanist fanboy . . . meh . . .

[jasmine]

I'm not really a transhumanist fanboy . . . meh . . .


Thanks for the info. Really thought-provoking.

[Trystan]

Thanks for the info. Really thought-provoking.

I've avoided looking like a paranoid sci-fi nutcase, so it's really problem at all. :)

[jasmine]

I've avoided looking like a paranoid sci-fi nutcase, so it's really problem at all. :)

No you haven't. What are you trying to say?

[Trystan]

No you haven't. What are you trying to say?

I think that all this excitement over technology is somewhat dangerous.

[Pirate turtle the 11th]

he has transformers pajamas.


But seriously i think advances in AI are positive.

[jasmine]

I think that all this excitement over technology is somewhat dangerous.

Why?

I really can't think of anything else to ask as you could have explained two or three posts previously or don't bother as the case may be.

[jasmine]

Okay, well I give up. Enjoy your games with the racist TomK. But do ask yourselves why you find a reactionary so appealing and so entertaining. Perhaps you are more like him than you care to think.

[Trystan]

Okay, well I give up. Enjoy your games with the racist TomK. But do ask yourselves why you find a reactionary so appealing and so entertaining. Perhaps you are more like him than you care to think.

It's called a sense of humor. Don't be so dour. :)

[Lord Testicles]

Okay, well I give up. Enjoy your games with the racist TomK. But do ask yourselves why you find a reactionary so appealing and so entertaining. Perhaps you are more like him than you care to think.

Were did this come from? :huh:

[Dean]

I think that as long as a computer is programmed it's not intelligent or conscious - it's the programmer that's conscious and has provided the options and the intelligence that fools the questioner.
Agreed.


However, were a chip to be invented that allowed a none-programmed machine to think would that be consciousness and would the machine have rights?
I don't see how that would generate consciousness. It is still a programmed set of responses, the only difference is that the second case allows the machinery to adapt.

[Qwerty Dvorak]

Just wondering what all you techies (there are a few here) think about AI. A new Turing Test is about to take place:

http://www.guardian.co.uk/technology/2008/oct/05/artificialintelligenceai

I think that as long as a computer is programmed it's not intelligent or conscious - it's the programmer that's conscious and has provided the options and the intelligence that fools the questioner.

However, were a chip to be invented that allowed a none-programmed machine to think would that be consciousness and would the machine have rights?
Personally I think that how a being is treated by the law is a function not of any internal factors like biology or conciousness but rather of the being's relation to and role in society. What this says for the artificially intelligent robot I'm not sure but it's food for thought.

[Bud Struggle]

Were did this come from? :huh:

Girlfren has a "love Jones" for me--but I told her, no. :(

[ÑóẊîöʼn]

I think the Turing test is far too limited. A clever programmer can easily create a series of canned responses that the computer spits out when it see the right series of words typed by the tester.

A definitive test of consciousness would able to test a subject's creative ability and decision making capability.

I think ultimately real AI will be achieved as a synthesis of the "top-down" and "bottom-up" approaches. Pre-programmed responses will be the equivalent of "instincts" while responses made up on the fly will be the equivalent of "interactions" - like when a human drives a car, they are constantly relying on sensory feedback in order to keep the vehicle on the road.


I think that all this excitement over technology is somewhat dangerous.

Please elaborate.

[apathy maybe]

Just wondering what all you techies (there are a few here) think about AI. A new Turing Test is about to take place:

http://www.guardian.co.uk/technology/2008/oct/05/artificialintelligenceai
Firstly, it isn't a "new Turing Test", it is computers being subject to the Turing Test. (I thought that someone had come up with a "better" test or something.

I think that as long as a computer is programmed it's not intelligent or conscious - it's the programmer that's conscious and has provided the options and the intelligence that fools the questioner.
If you are chatting online with someone you have never meet, and you do so multiple times over a long period of time, and you enjoy the conversations, have fun, discuss philosophy, men, fashion, politics (whatever), and it turns out that this person is actually a computer, would you feel that your conversations suddenly became worth less because it wasn't with a "real" person? If you can't tell the difference, what does it matter if the "intelligence" is "merely" simulated instead of being real?


However, were a chip to be invented that allowed a none-programmed machine to think would that be consciousness and would the machine have rights?
Just because a computer has been programmed, doesn't mean that it can't think. A computer program that has been created to "teach" itself, can theoretically become more knowledgeable on various subjects, and have complex discussions about these subjects, yet they were still programmed originally.


I think the Turing test is far too limited. A clever programmer can easily create a series of canned responses that the computer spits out when it see the right series of words typed by the tester.
You under estimate the complexity of the task. Considering spelling differences (and mistakes), the complexity of grammar and so on, the database of "cues", to which "canned responses" would be given, would have to be quite large...

A definitive test of consciousness would able to test a subject's creative ability and decision making capability.
And how would you test that then?


I think ultimately real AI will be achieved as a synthesis of the "top-down" and "bottom-up" approaches. Pre-programmed responses will be the equivalent of "instincts" while responses made up on the fly will be the equivalent of "interactions" - like when a human drives a car, they are constantly relying on sensory feedback in order to keep the vehicle on the road.
Computers are perfectly capable of "driving" cars, without any real intelligence. Using the same idea of "cues" and "canned responses" that can be used in a conversation. A person is near that zebra crossing, therefore slowdown. They look like they want to cross, therefore stop.

The traffic lights are orange/yellow, therefore stop if safe, the traffic lights are red, stop (if safe). The traffic lights are green, go go go! (But watch out for pedestrians and other cars.)



The Turing Test was created partly as a tongue in cheek response to the question "can machines ever think?" The question is basically meaningless, but the test provides for a way around this meaninglessness. Can a machine think can never be answered (any more then "do other people think?").
To quote myself:
A possible problem [with the Turing Test] is knowing whether the machine was really thinking or was just cleverly programmed. The point is however, if you can't tell the difference, what does it matter?

[pusher robot]

A possible problem [with the Turing Test] is knowing whether the machine was really thinking or was just cleverly programmed. The point is however, if you can't tell the difference, what does it matter?

This always reminds me of Charles C. Clarke's aphorism that "any sufficiently advanced technology is indistinguishable from magic." The corollary, I guess, is that if you can't distinguish something from magic, it must be sufficiently advanced.

[Lynx]

From model to modeled.

[apathy maybe]

This always reminds me of Charles C. Clarke's aphorism that "any sufficiently advanced technology is indistinguishable from magic." The corollary, I guess, is that if you can't distinguish something from magic, it must be sufficiently advanced.

Surely that would be Arthur...

And yes, I have seen on Slashdot someone with a similar corollary in their signature. If you can distinguish something from magic, it is insufficiently advanced.