NationStates Jolt Archive


Artificial Intelligence

Kelikstadt
12-09-2005, 16:01
I have seen/read/heard so many things on the sbject of "A.I conciousness". What about computer games? With the developement of 'true AI' being filtered into computer games does that mean that (if A.Is are ever deemed to be concious and that technology makes it's way into games) it will count as murder to kill a computer game character?
The Squeaky Rat
12-09-2005, 16:06
I have seen/read/heard so many things on the sbject of "A.I conciousness". What about computer games? With the developement of 'true AI' being filtered into computer games does that mean that (if A.Is are ever deemed to be concious and that technology makes it's way into games) it will count as murder to kill a computer game character?

That depends if the character truly ceases to exist, or if it was just some pixel representation under the AI's control. In the second case it is no different from "killing" another player in a multiplayer game - so no.

In the first case however: yes, it would be. But the developers would not be allowed to make such a game.
Pure Metal
12-09-2005, 16:07
depends if they have self awareness or not... just watch that episode of star trek TNG with Dr. Moriarty and all will become clear ;)
Hemingsoft
12-09-2005, 16:12
My worry is how many months must the character exist before it is no longer entitled to be killed.
Balipo
12-09-2005, 16:12
I think this depends on your attitude toward life.

If I create a concious AI and feel that it needs to be destroyed, I believe I have the right. It is a construct, not a living thing.

As far as games go, I don't think it would be possible given that the game would have too much control and the programmers and designers and plot developers would be seemingly useless.
Hemingsoft
12-09-2005, 16:17
I think this depends on your attitude toward life.

If I create a concious AI and feel that it needs to be destroyed, I believe I have the right. It is a construct, not a living thing.

As far as games go, I don't think it would be possible given that the game would have too much control and the programmers and designers and plot developers would be seemingly useless.

I think it would be useless for players, I mean the characters would play themselves.
The Squeaky Rat
12-09-2005, 16:18
I think this depends on your attitude toward life.

If I create a concious AI and feel that it needs to be destroyed, I believe I have the right. It is a construct, not a living thing.

The "living" is not an argument. We kill living things daily: plants, animals, cells and so on and aside from some minor groups very little protest is heard. And those protests mostly originate in perceived suffering or harming of the creatures that were killed and are not due to the killing itself. So being "biological life" is not really a factor.

If an AI can have feelings and suffer it deserves just as much consideration as for instance a puppy would. If it can also reason, why should its existence be valued as less than human ?

After all, technically we are contructs too.
Compulsive Depression
12-09-2005, 16:19
Considering at the moment game AIs tend to be simple little state machines, I don't think you'll have to feel guilty for a while.

It's not like you can't just start a new game to bring them back to life, anyway.
The Squeaky Rat
12-09-2005, 16:23
It's not like you can't just start a new game to bring them back to life, anyway.

Unless the AI starts from scratch each game and develops during it. Then every new game would be a new person. A "reincarnation" perhaps, but not the same individual.

Sims 2050 anyone ?
Bobfarania
12-09-2005, 16:24
For me, if an AI can feel pain or suffering, even if it is only simulated, then it should be protected by some form of law.
Transipsheim
12-09-2005, 16:25
The question is, would it be murder? Unlike a human, you could simply back up and reinstate an AI. A game character wouldn't really be one, it might be controlled by one, but if you kill a character in a game today, you just need to turn it off and on again (the game, without saving) to see the character again.

What happens if we give AI the same rights we have? We suddenly have machines which are more intelligent, think quicker and can communicate with one another much more effectively than we can. In short, we'd no longer be the dominant species and knowing the people in power around the world, they'd never let that happen.
Bobfarania
12-09-2005, 16:27
the AI would have to follow the same rights as humans do. same rights, same punishment
Kroisistan
12-09-2005, 16:28
My worry is how many months must the character exist before it is no longer entitled to be killed.

You sir have earned a Kroisistan Award for attempted Thread Corruption under Exceptional Crcumstances.

In a debate about artifical intelligence, you were able albeit subtly, to bring up abortion. You sly dog, you. :)
Hemingsoft
12-09-2005, 16:33
You sir have earned a Kroisistan Award for attempted Thread Corruption under Exceptional Crcumstances.

In a debate about artifical intelligence, you were able albeit subtly, to bring up abortion. You sly dog, you. :)

:D

Actually it was sarcasm and an attempted joke.


::EDIT::

my first NS forums award!!!
woot woot
The Squeaky Rat
12-09-2005, 16:34
The question is, would it be murder? Unlike a human, you could simply back up and reinstate an AI. A game character wouldn't really be one, it might be controlled by one, but if you kill a character in a game today, you just need to turn it off and on again (the game, without saving) to see the character again.

That would indeed not be killing, and therefor by definition not murder. However, it would be "lobotomising" the AI in question - by restoring it to an earlier point of its development you strip away part of its memory and experiences.

If you could *not* save the game and if when dead the AI cannot be revived exactly as it was, you have really killed it. And why would the termination of an intelligence not be murder ?
Quorm
12-09-2005, 16:47
Of course, in any realistic setup the AI wouldn't be destroyed when you kill it in the game. The AI would simply be in control of the other characters in the game rather than actualy being them. So, killing your computer oponent in Unreal Tournament 2084 or whatever would be completely analagous to what happens when you kill a human oponent in the game - he would immediately respawn with all his memories and cognitive abilities intact.

I suppose you could set it up to reinitialize the AI everytime you kill it, but that would completely defeat the purpose of having an artificial intelligence controlling your oponent.

The thing that I wonder about though is how we will deal with the morality of having what are essentially virtual slaves. After all, in principle the AIs can be programmed to 'enjoy' being totally subservient - is there still something wrong with treating a sentient being as a slave if they enjoy it? If AIs are going to have rights, what would qualify as abuse of an AI?
Balipo
12-09-2005, 16:49
The "living" is not an argument. We kill living things daily: plants, animals, cells and so on and aside from some minor groups very little protest is heard. And those protests mostly originate in perceived suffering or harming of the creatures that were killed and are not due to the killing itself. So being "biological life" is not really a factor.

If an AI can have feelings and suffer it deserves just as much consideration as for instance a puppy would. If it can also reason, why should its existence be valued as less than human ?

After all, technically we are contructs too.

In your belief we are constructs. I my atheistic belief, we aren't. We are evolved life forms.

Now the other issue you brought up was feelings. Would AI have feelings, or electronic responses programmed to mimic feelings, and how would we know the difference?

This debate gets tougher as you move inward, and it's all speculative so there are no real answers.
The Squeaky Rat
12-09-2005, 16:58
In your belief we are constructs. I my atheistic belief, we aren't. We are evolved life forms.

Constructed by evolution. Shaped and formed. Ultimately in a nine month proces inside a production facility.

However, if you are an atheist, why would you ascribe any special value to biological life ? The only reason to do so is if you believe in souls.

Now the other issue you brought up was feelings. Would AI have feelings, or electronic responses programmed to mimic feelings, and how would we know the difference?

We wouldn't , therefor they are real. Same as with humans - whose feelings are after all also nothing more than bunches of electrical impulses and chemical processes. Believing there is more to it is religious.
Kelikstadt
13-09-2005, 10:19
depends if they have self awareness or not... just watch that episode of star trek TNG with Dr. Moriarty and all will become clear ;)

I know that one. And there's the one where those gangsters on the holodeck find out about the 'real world' and they get tricked into leaving the holodeck and 'die', which is kind of a bad thing for Picard to have done I think. I'm not a trekkie by the way :rolleyes:
Kelikstadt
13-09-2005, 10:33
http://www.jabberwacky.com/ is (apparantly) one of the world's best A.Is. See what you think. I'm not particularly impressed with it.
Kelikstadt
13-09-2005, 10:45
http://www.turinghub.com/testing/ This is interesting. See if you can guess whether or not you are talking to an AI.
Khymru
13-09-2005, 10:47
that is rubbish, it asked if I was a boy, I said no a man and it asked how old my husband was!!
Kelikstadt
13-09-2005, 10:52
I think the 'AI' at jabberwacky.com is actually just a chat room that randomly sends your replies to other people using it.
Harlesburg
13-09-2005, 10:52
I always had a problem with this idea.
Why do it?
Skyfork
13-09-2005, 11:06
Of course, in any realistic setup the AI wouldn't be destroyed when you kill it in the game. The AI would simply be in control of the other characters in the game rather than actualy being them. So, killing your computer oponent in Unreal Tournament 2084 or whatever would be completely analagous to what happens when you kill a human oponent in the game - he would immediately respawn with all his memories and cognitive abilities intact.

I suppose you could set it up to reinitialize the AI everytime you kill it, but that would completely defeat the purpose of having an artificial intelligence controlling your oponent.

The thing that I wonder about though is how we will deal with the morality of having what are essentially virtual slaves. After all, in principle the AIs can be programmed to 'enjoy' being totally subservient - is there still something wrong with treating a sentient being as a slave if they enjoy it? If AIs are going to have rights, what would qualify as abuse of an AI?

Quite the tricky question! If I were to design a interactive program using advanced AI, say one of those tamogatchi slave games and I "torture" the AI in-game. Is it truly being tortued? I don't believe so, because odds are, I programed it to respond that way in the first place. If I used an adaptive AI I could just easily remove the option for it to respond realisticly to torture and I woudl still have to program in multiple sets of responses to the same situation (pain in the butt). But in the end, the AI is not truely feeling pain or anything at all it is merely reacting to a given stimulus using a set of preprogramed responses.
Harlesburg
13-09-2005, 11:08
I agrree with that guy.
Mitigation
14-09-2005, 03:18
Aaaa IIiiii...... wait.... what was the A again?
Desperate Measures
14-09-2005, 03:21
My worry is how many months must the character exist before it is no longer entitled to be killed.
Certainly, once it attains consciousness and can act of its own free will, it can be said to exist.