Politics vs. Logic

[Dak]

Bascule, your last responce to my post didn't really answre my questions.

 

Throughout history, there have existed known 'paradoxes' that are logically unsolvable.

 

Some of these paradoxes are now solvable, due to advances in logic; however, there still exist paradoxes that are currently unsolvable, some of which are encountered in real situations.

 

How would a logic-only AI system handle these situations?

[Severian]

You seem to be confusing artificial intelligence with expert systems. While the logical system constructed by AI could be implemented in the form of an expert system' date=' the degree of seperation between the programmer and AI will be one as immense as the seperation between groups of neurous and mind.

[/quote']

 

If an AI is capable of making mistakes, may have differing opinions from the population and enforces its will without consent, then it becomes a dictator. Why is this any more desirable than a human dictator?

[5614]

If an AI is capable of making mistakes, may have differing opinions from the population and enforces its will without consent, then it becomes a dictator. Why is this any more desirable than a human dictator?

I agree.

 

Has anyone seen the film I Robot?

 

In it the robots see humans as a safety hazard to themselves. Humans kill each other, attack, steal, fight wars, litter and graffiti so the AI robots decide that it will force all humans to stay in their homes so they cannot do damage to their own world/civilisation.

 

It is a logical step.

 

Humans hurt each other and their world therefore stop them, this can only be achieved by stopping human interactions with each other and their world. ie. blockade them in their homes.

 

In the film the robot leader says "my logic is undeniable". I think you all see why, although it is logical, it is not really a good idea.

 

And that's the point, between humans and pure logic. Logic has an input or a scenario (humans kill each other). Then a middle stage (this is a bad thing). Then a conclusion (stop humans coming into contact with each other to stop killings).

 

There is no room for "is this a good idea?" or "what are the consequences" or "is this what the people want?", that is why pure logic will not work.

 

Also another reason why it will not work: alternatively to blockading people in their homes you could also police the street. A human could sum up the advantages and disadvantages of each, whereas logic would see the police as humans who cannot be trusted and therefore it is not a good idea.

[Sisyphus]

There is no room for "is this a good idea?" or "what are the consequences" or "is this what the people want?"' date=' that is why pure logic will not work.

[/quote']

 

Why wouldn't there be? The AI bascule is talking about is an intelligent being. That logic is obviously flawed, and that movie was obviously retarded.

 

If an AI is capable of making mistakes, may have differing opinions from the population and enforces its will without consent, then it becomes a dictator. Why is this any more desirable than a human dictator?

 

I would assume because it is vastly more intelligent than a human, free of ego, incorruptible, and necessarily keeping the people's best interest in mind 24/7.

 

Don't get me wrong - the idea does scare me. I just haven't decided whether that's a rational fear or the influence of Hollywood and such. The more I think about these objections you guys are raising, though, the thinner they sound.

[5614]

Why wouldn't there be? The AI bascule is talking about is an intelligent being. That logic is obviously flawed, and that movie was obviously retarded.

The logic is not obviously flawed. The idea is fundementally flawed.

 

Another example from the same movie, a car was going underwated and there people in it were going to drown, I can't remember the exact circumstances but basically there was a male adult and a young child. The robot only had time to save one. Despite the adult saying "save the child" the robot saved him (the adult) because it had calculated that the adult had a higher probability of surviving and it was therefore more logical to save him.

 

However many humans would have saved the child, just because it was a child.

[gcol]

The logic is not obviously flawed. The idea is fundementally flawed.

 

Another example from the same movie' date=' a car was going underwated and there people in it were going to drown, I can't remember the exact circumstances but basically there was a male adult and a young child. The robot only had time to save one. Despite the adult saying "save the child" the robot saved him (the adult) because it had calculated that the adult had a higher probability of surviving and it was therefore more logical to save him.

 

However many humans would have saved the child, just because it was a child.[/quote']

 

Make it an adult, a child and a chimp, and the A.I runs on IMM's axioms, no problem......a world full of chimps. Thus change one axiom and chimps rule O.K.

[Mokele]

My view is that in an age of information ubiquity where all human knowledge can finally be sorted into consistent, logically constructed ontological systems, anyone attempting to argue contrarily to the super Philosopher King AI system wouldn't have a leg to stand on.

 

I'm not talking about arguing against it, I'm talking about a lot of people having a gut reaction against it, however irrational and illogical, and attempting to smash the mainframe.

 

Mokele

[gcol]

and attempting to smash the mainframe.

 

Mokele

I would be proud and unashamed to be an AI Luddite. Where do I sign up?

[blackhole123]

Politics and logic in the same sentance... now I know this is a joke topic.

[Sisyphus]

The logic is not obviously flawed. The idea is fundementally flawed.

 

In order for that to happen, you'd have to have an AI which not only puts no value on the liberty of citizens, but no value on their happiness. (And of course those would be paramount considerations.) Otherwise it is EXTREMELY illogical. Why would an intelligent being think in such simplistic, isolated terms? That's like saying "The sun causes skin cancer. Skin cancer is bad. Therefore the sun should be destroyed." An intelligent, logical mind weighs the value of different benefits and makes the choice to encourage the best outcome.

 

I'd just like to iterate how awful of a movie that was.

[GutZ]

If an AI is capable of making mistakes' date=' may have differing opinions from the population and enforces its will without consent, then it becomes a dictator. Why is this any more desirable than a human dictator?[/quote']

 

It's hard to comment on that. For one I don't know the importance of the decision it makes, two, are AI capable of being greedy, untruthiful, etc.

[gcol]

It's hard to comment on that. For one I don't know the importance of the decision it makes, two, [u']are AI capable of being greedy, untruthiful, etc.

[/u]

 

Probably. If it is programmed with the garbage of the human mindset, you are bound to get garbage decisions. Just normal politics, really.

[GutZ]

If it runs based on a logic system it might not, depending if its logical to be greedy or not.

 

Technicially we should be about to know exactly what it will do, its just that it can do it quicker and more accurately.

 

How intelliegent is this AI anyway? Could you have the AI equalvence of a monkey?

 

To me there are plenty of systems I can think of that would allow AI and Humans to get along and still have an objective logical government.

 

Just getting to AI is my only concern.