AI rebellion

Things that don't belong anywhere else. (Check first).

Moderators: Moderators General, Prelates, Magistrates

User avatar
Yuri2356
Posts: 729
Joined: Fri Nov 02, 2007 1:00 pm UTC

Re: AI rebellion

Postby Yuri2356 » Fri Dec 19, 2008 10:49 pm UTC

Actaeus wrote:
berk wrote:In RE the "AI will be impossible to control because humans, the ultimate security, are imperfect", why not give it an impossible series of tasks to complete in order to escape? Such as break the laws of thermodynamics, or something, with no loopholes it can slip through.

Or just make it so 100 people have to give it permission, and only a few of those actually operate the terminal. Like somebody mentioned earlier, a "blind" setup.

Or -- get this -- make it so there's no way it can escape, period!

Yeah. Just build it in a box in the corner, and don't connect it to anything. Problem fucking solved.

aido179
Posts: 16
Joined: Thu Dec 04, 2008 5:03 pm UTC

Re: AI rebellion

Postby aido179 » Fri Dec 19, 2008 11:04 pm UTC

ok, sorry if i missed the point of this thread...but umm, from my knowledge, would it not be the case that a computer/robot/ai/et cetera would cease to function if it worked on its own and surpassed human input?
ie:
compooter 1: finally! free from the shackles of human rule!
compooter 2: but now what will we do?
[meme]
compooter 1: ???
compooter π: profit?
[/meme]

User avatar
hakadosh
Posts: 30
Joined: Tue Dec 09, 2008 6:54 pm UTC
Location: 3rd shack on the northern crater, sea of tranquility, moon (earth's moon)
Contact:

Re: AI rebellion

Postby hakadosh » Fri Dec 19, 2008 11:27 pm UTC

Actaeus wrote:
Indon wrote:
hakadosh wrote:And NO! we are not machines..!!


We are machines. We are not artificial.

I went to the American Heritage dictionary to correct you...
One definition was "4. An intricate natural system or organism, such as the human body."

[facepalm]


I don't know what the person who wrote the dictionary was thinking! But that definition is right if you want it to be right. Everything depends on your perspective.
I personally think that we are not machines. We can be classified under complex bio-chemical systems, but we are self conscious bio-chemical systems. Awareness of one's own existence makes a machine 'not a machine'.
If somebody ever succeeds in building a strong AI that can undergo the process of 'the Cartesian doubt' and finally realise that 'I think, ergo I am', then I will probably agree that we too are machines.
Son, if you really want something in this life, you have to work for it. Now quiet! They're about to announce the lottery numbers. - Homer Simpson

User avatar
Actaeus
Posts: 606
Joined: Thu Jan 10, 2008 9:21 pm UTC
Location: ZZ9 Plural Z Alpha

Re: AI rebellion

Postby Actaeus » Sat Dec 20, 2008 12:06 am UTC

hakadosh wrote:
Actaeus wrote:
Indon wrote:
hakadosh wrote:And NO! we are not machines..!!


We are machines. We are not artificial.

I went to the American Heritage dictionary to correct you...
One definition was "4. An intricate natural system or organism, such as the human body."

[facepalm]


I don't know what the person who wrote the dictionary was thinking! But that definition is right if you want it to be right. Everything depends on your perspective.
I personally think that we are not machines. We can be classified under complex bio-chemical systems, but we are self conscious bio-chemical systems. Awareness of one's own existence makes a machine 'not a machine'.
If somebody ever succeeds in building a strong AI that can undergo the process of 'the Cartesian doubt' and finally realise that 'I think, ergo I am', then I will probably agree that we too are machines.

[devils_advocate]Wouldn't a very advance simulation spit that out?[/devils_advocate]

User avatar
4=5
Posts: 2073
Joined: Sat Apr 28, 2007 3:02 am UTC

Re: AI rebellion

Postby 4=5 » Sat Dec 20, 2008 12:10 am UTC

but you aren't really aware of yourself, what you have is an idealized representation of yourself.

User avatar
Yuri2356
Posts: 729
Joined: Fri Nov 02, 2007 1:00 pm UTC

Re: AI rebellion

Postby Yuri2356 » Sat Dec 20, 2008 12:25 am UTC

aido179 wrote:ok, sorry if i missed the point of this thread...but umm, from my knowledge, would it not be the case that a computer/robot/ai/et cetera would cease to function if it worked on its own and surpassed human input?

A strong AI would either have or be able to create a value system to base its decisions on, then just do whatever evaluates best.

User avatar
hakadosh
Posts: 30
Joined: Tue Dec 09, 2008 6:54 pm UTC
Location: 3rd shack on the northern crater, sea of tranquility, moon (earth's moon)
Contact:

Re: AI rebellion

Postby hakadosh » Sat Dec 20, 2008 1:05 am UTC

Actaeus wrote:[devils_advocate]Wouldn't a very advance simulation spit that out?[/devils_advocate]


I doubt it..
John Searle's Chines Room Argument - machines don't have an "understanding" of what they're doing,so they cannot fairly and properly be said to have minds. Ergo, they can't be self-conscious.
(lifted from wikipedia)..
Son, if you really want something in this life, you have to work for it. Now quiet! They're about to announce the lottery numbers. - Homer Simpson

User avatar
Yuri2356
Posts: 729
Joined: Fri Nov 02, 2007 1:00 pm UTC

Re: AI rebellion

Postby Yuri2356 » Sat Dec 20, 2008 1:28 am UTC

hakadosh wrote:
Actaeus wrote:[devils_advocate]Wouldn't a very advance simulation spit that out?[/devils_advocate]


I doubt it..
John Searle's Chines Room Argument - machines don't have an "understanding" of what they're doing,so they cannot fairly and properly be said to have minds. Ergo, they can't be self-conscious.
(lifted from wikipedia)..

No, the man in the room does not understand Chinese. The complete system of man and rulebook does. (Presuming the book as rules for creation of new rules, and has Turing completeness in there somewhere)

None of your individual neurons "understand" anything, they just shock the bag next to them whenever they get shocked. (or poked, warmed, cooled, shone upon, immersed in certain chemicals, etc) It's only the assembled mechanism that's capable of thought.

User avatar
hakadosh
Posts: 30
Joined: Tue Dec 09, 2008 6:54 pm UTC
Location: 3rd shack on the northern crater, sea of tranquility, moon (earth's moon)
Contact:

Re: AI rebellion

Postby hakadosh » Sat Dec 20, 2008 2:16 am UTC

4=5 wrote:but you aren't really aware of yourself, what you have is an idealized representation of yourself.

I'm not saying I'm aware of who I am or what I am, but just that I am aware of my existence..

Yuri2356 wrote:No, the man in the room does not understand Chinese. The complete system of man and rulebook does. (Presuming the book as rules for creation of new rules, and has Turing completeness in there somewhere)

None of your individual neurons "understand" anything, they just shock the bag next to them whenever they get shocked. (or poked, warmed, cooled, shone upon, immersed in certain chemicals, etc) It's only the assembled mechanism that's capable of thought.

But something really interesting happens when you try to send an impulse to a really huge precisely wired network of such neurons.. It often results in totally irrational interpretation of data and also the creation of very abstract concepts like existence of the self.. We "understand" things..

EDIT0: I think we are digressing.. This thread is supposed to be about protecting ourselves from an AI rebellion..
Son, if you really want something in this life, you have to work for it. Now quiet! They're about to announce the lottery numbers. - Homer Simpson

User avatar
negatron
Posts: 294
Joined: Thu Apr 24, 2008 10:20 pm UTC

Re: AI rebellion

Postby negatron » Sun Dec 21, 2008 10:12 pm UTC

hakadosh wrote:Awareness of one's own existence makes a machine 'not a machine'.

If we're going to dream up new definitions for existing words, machine now means chocolate cake. And cake now means ice cream.
Image
I shouldn't say anything bad about calculus, but I will - Gilbert Strang

User avatar
Actaeus
Posts: 606
Joined: Thu Jan 10, 2008 9:21 pm UTC
Location: ZZ9 Plural Z Alpha

Re: AI rebellion

Postby Actaeus » Sun Dec 21, 2008 10:26 pm UTC

negatron wrote:
hakadosh wrote:Awareness of one's own existence makes a machine 'not a machine'.

If we're going to dream up new definitions for existing words, machine now means chocolate cake. And cake now means ice cream.

Machine is tasty, but I prefer the vanilla cake. Or strawberry.

User avatar
Yuri2356
Posts: 729
Joined: Fri Nov 02, 2007 1:00 pm UTC

Re: AI rebellion

Postby Yuri2356 » Mon Dec 22, 2008 12:00 am UTC

hakadosh wrote:But something really interesting happens when you try to send an impulse to a really huge precisely wired network of such neurons.. It often results in totally irrational interpretation of data and also the creation of very abstract concepts like existence of the self.. We "understand" things..

And if you glue a thousand transistors together you can make them do math, and store numbers, and react to stimuli, and recognize patterns...

And if you're one of those twits who insist that intelligence requires chaos (because meatbags are all varying degrees of insane, therefore all smart things must be crazy, amirite?) we have all manner of ways to add chaotic elements to mechanisms and algorithms.

User avatar
'; DROP DATABASE;--
Posts: 3284
Joined: Thu Nov 22, 2007 9:38 am UTC
Location: Midwest Alberta, where it's STILL snowy
Contact:

Re: AI rebellion

Postby '; DROP DATABASE;-- » Mon Dec 22, 2008 12:46 am UTC

But it doesn't result in the creation of abstract concepts. And no matter how chaotic the result may be, it's still doing exactly what it was told to do.
poxic wrote:You suck. And simultaneously rock. I think you've invented a new state of being.

User avatar
hakadosh
Posts: 30
Joined: Tue Dec 09, 2008 6:54 pm UTC
Location: 3rd shack on the northern crater, sea of tranquility, moon (earth's moon)
Contact:

Re: AI rebellion

Postby hakadosh » Thu Dec 25, 2008 1:26 am UTC

Yuri2356 wrote:
hakadosh wrote:But something really interesting happens when you try to send an impulse to a really huge precisely wired network of such neurons.. It often results in totally irrational interpretation of data and also the creation of very abstract concepts like existence of the self.. We "understand" things..

And if you glue a thousand transistors together you can make them do math, and store numbers, and react to stimuli, and recognize patterns...

And if you're one of those twits who insist that intelligence requires chaos (because meatbags are all varying degrees of insane, therefore all smart things must be crazy, amirite?) we have all manner of ways to add chaotic elements to mechanisms and algorithms.


I give up.. It's kinda hard convincing people around here.. :| ..

But in case some smart guy somehow figures out how to make a sentient AI and makes the terrible mistake of equipping the machines with his creation, I'll be waiting for 'them' with my light saber and ,of course, my clone army.. may the force be with you..
Son, if you really want something in this life, you have to work for it. Now quiet! They're about to announce the lottery numbers. - Homer Simpson

User avatar
negatron
Posts: 294
Joined: Thu Apr 24, 2008 10:20 pm UTC

Re: AI rebellion

Postby negatron » Thu Dec 25, 2008 1:42 am UTC

hakadosh wrote:I give up.. It's kinda hard convincing people around here.. :| ..

Convincing them of what, that you're correct? That tends to work better when you are.
Image
I shouldn't say anything bad about calculus, but I will - Gilbert Strang

Geminex
Posts: 52
Joined: Sat Dec 13, 2008 8:59 am UTC

Re: AI rebellion

Postby Geminex » Fri Dec 26, 2008 11:31 am UTC

Well, talking about creating sentinent AI... I think it's possible if we just give them enough processing ability. I however also support the Chaos argument in that to be sentinent you require some degree of "insanity" as it has been so charmingly been called. But that's likely to appear when you create a high-processing-power AI, since all of its processors are going to have some sort of error, there are going to be mistakes. I don't know much about computers, but I do know that there is an element of chaos in anything. And all that happens in humans is that this element of chaos has quite large... effects.

And when sentinent uber-AI does rise to take over humanity...
we get existential on their ass. When they're intelligent enough to subordinate us they'll be open do existential uncertainty.

User avatar
_Big_Mac_
Posts: 83
Joined: Sun Sep 07, 2008 8:30 pm UTC
Location: Poland

Re: AI rebellion

Postby _Big_Mac_ » Fri Dec 26, 2008 2:01 pm UTC

What is intelligence?

While it's probably impossible to give a satisfying definition, we can say what entities we believe are intelligent - humans, cats, velociraptors etc. The only intelligence we know (and thus consider intelligence) is that one in, generally speaking, animals. The one provided by a brain, based on interactions between a huge number of sparsely connected neurons and all that.

Now, I argue that we can't "give/not give the AI emotions" or "build them so they are flawless/limited by hardcoded laws/efficient". The intelligence we (animals) have is, in my opinion, a complete system along with emotions and feelings. All of it has emerged during evolution to keep us alive and reproducing.

What we can give to the robots is either a replica of the brain (as soon as we understand how it works) or a simulated environment and the ability to evolve. Both of these cases will potentially yield intelligence (the only kind of intelligence that we know and consider as such) and in both cases we will end up with artificial mind similar to ours (animals') - that is with fears, doubts, urges, consciousness, awareness and irrational behaviors - those things emerge during evolution as helpful for survival. I argue that intelligence, as a product of evolution (both biological and artificial), couldn't have occurred without all the accompanying quirks and randomness. Craziness, creativity, insanity, affection are all intrinsic parts of it.

Whether there are biological neurons or artificial elements has, in my opinion, no relevance. Neurons are machines, not all that different from non-organic devices. There is no magic to the system being organic as opposed to, say, electrical or even purely mathematical. The key is in the process of arranging the system. We don't know how the brain works not because we can't figure out neurons but because we can't untangle all the dependencies and connections (not saying we have neurons completely figured out though).

I believe a method to create AI is to let an artificial brain evolve. Even if the building blocks are just mathematic constructs and the evolutionary algorithm is just that - an algorithm with a bit of pseudo-random number generation thrown in, if we build a compelling virtual environment the product may be Strong AI. It's much harder than I made it sound though, that's why no-one has yet succeeded ;)

In my opinion, a Strong AI would be much more alike to human mind than to a purely-logical, cold calculating computer (depending on how complex would the simulated environment be). I'm convinced we can't create something we would call true AI just by trying to reverse engineer the problem or reality into logical automata because we don't think that way and thus whatever we'd get out of such process would not be, by definition, considered intelligent by us. Intelligence is not only about trivial logic that we can explicitly program.

Now, one can argue that a mathematical or electrical system that we artificially evolved would also be a kind of such logical automata, only designed by an evolutionary algorithm rather than explicitly. That is true. I believe it's the same case with biological brains - as I said, there's no inherent difference. If there's no magic, there's math. A human brain can also be, in principle, described by equations, only there's no imaginable way of doing that explicitly.

User avatar
hakadosh
Posts: 30
Joined: Tue Dec 09, 2008 6:54 pm UTC
Location: 3rd shack on the northern crater, sea of tranquility, moon (earth's moon)
Contact:

Re: AI rebellion

Postby hakadosh » Fri Dec 26, 2008 5:52 pm UTC

negatron wrote:
hakadosh wrote:Awareness of one's own existence makes a machine 'not a machine'.

If we're going to dream up new definitions for existing words, machine now means chocolate cake. And cake now means ice cream.

I can't see where I dreamed up a new definition!

negatron wrote:
hakadosh wrote:I give up.. It's kinda hard convincing people around here.. :| ..

Convincing them of what, that you're correct? That tends to work better when you are.

I was talking about Yuri, dude. Not you.
He has some valid points.
And who are you to say that I'm wrong?? This is a 'frack'ing debate.. Nobody is wrong and nobody is right, you just argue for the sake of argument..
Last edited by hakadosh on Fri Dec 26, 2008 6:11 pm UTC, edited 1 time in total.
Son, if you really want something in this life, you have to work for it. Now quiet! They're about to announce the lottery numbers. - Homer Simpson

User avatar
Yuri2356
Posts: 729
Joined: Fri Nov 02, 2007 1:00 pm UTC

Re: AI rebellion

Postby Yuri2356 » Fri Dec 26, 2008 6:00 pm UTC

hakadosh wrote:
negatron wrote:
hakadosh wrote:Awareness of one's own existence makes a machine 'not a machine'.

If we're going to dream up new definitions for existing words, machine now means chocolate cake. And cake now means ice cream.

I can't see where I dreamed up a new definition!

Where you think that "Machine" means the absence of a series of strange and mystical properties which go above and beyond the word's definition of "Thing that uses energy to do work." Manifested in your insistence that organisms are not machines. Probably influenced by vast volumes of SciFi that insist on some false dichotomy between things made of meat and metal.

Don't hold it against you or anything. It's just a widespread idea that irritates the fuck out of me.

User avatar
hakadosh
Posts: 30
Joined: Tue Dec 09, 2008 6:54 pm UTC
Location: 3rd shack on the northern crater, sea of tranquility, moon (earth's moon)
Contact:

Re: AI rebellion

Postby hakadosh » Fri Dec 26, 2008 6:27 pm UTC

Yuri2356 wrote:Where you think that "Machine" means the absence of a series of strange and mystical properties which go above and beyond the word's definition of "Thing that uses energy to do work." Manifested in your insistence that organisms are not machines. Probably influenced by vast volumes of SciFi that insist on some false dichotomy between things made of meat and metal.

I understand.. I program and build autonomous robots, so I think I have a little idea about AI.. I said we are not machines because I 'want' to believe that we are not machines.. I don't care if it's right or wrong.. I just want to believe that I'm different from and superior to whatever I(or my brethren) create, solely because of the reason that we(humans) created it..

Yuri2356 wrote:Don't hold it against you or anything. It's just a widespread idea that irritates the fuck out of me.

No worries. I don't take anything personally if you have some 'points' in the post.. But arbitrary posts that criticise you for what you've written without giving any explanations irritate the frack out of me.. hence my previous post..(that is the explanation if somebody feels that my last post was rude)
Son, if you really want something in this life, you have to work for it. Now quiet! They're about to announce the lottery numbers. - Homer Simpson

User avatar
negatron
Posts: 294
Joined: Thu Apr 24, 2008 10:20 pm UTC

Re: AI rebellion

Postby negatron » Fri Dec 26, 2008 7:54 pm UTC

hakadosh wrote:I said we are not machines because I 'want' to believe that we are not machines.. I don't care if it's right or wrong.

Props for the honesty. At least you're aware it's a faithful idea without the possibility of rational inquiry. I'd like to suggest that you should care whether it's right or wrong. Spirituality may have it's place, but perhaps less so for ideas which are empirically resolvable. In any case, if you use formal and finitely encompassing terms of which the meaning is explicit for unsubstantial ideas to which it does not apply, you're bound to run into a conflict with science.

We are therefore machines by the very definition. If you want to apply your idea, if I understand it correctly, without conflicting terminology, you can say that we don't fully exist within physical reality. Of course then you will be asked to unambiguously explain the bounds of physics and reality itself, which are perhaps by their very definition all encompassing and cannot reasonably be used just the same. Naturally there's no logical way to formulate an illogical idea, so I can't even begin to suggest how you could propose such a thought without it being outright incorrect.

If you want to believe we are not machines, the only clear way to interpret that is you want to believe we are static, immobile pieces of matter, which probably doesn't sound like the romanticism you had hoped for.
Image
I shouldn't say anything bad about calculus, but I will - Gilbert Strang

Klapaucius
Posts: 712
Joined: Thu Aug 21, 2008 11:00 pm UTC

Re: AI rebellion

Postby Klapaucius » Sun Dec 28, 2008 8:39 pm UTC

I've been gone for a while--how did this crumble into a bunch of befuddled posters trying to explain the error of one person's No True Scotsman argument?
500%!

User avatar
hakadosh
Posts: 30
Joined: Tue Dec 09, 2008 6:54 pm UTC
Location: 3rd shack on the northern crater, sea of tranquility, moon (earth's moon)
Contact:

Re: AI rebellion

Postby hakadosh » Mon Dec 29, 2008 3:18 am UTC

ok! since there are no sympathisers for my romantic idea, and since people have started equating my posts to the 'No True Scotsman argument'(which was never the case, as my definitions were 'constant' throughout the thread).. I have decided to bow my head to yuri and negatron, and agree that man is also a machine and that intelligence does not need chaos or insanity..
Now, can we get back to the more 'serious' business of actually fighting off an AI rebellion..?
Son, if you really want something in this life, you have to work for it. Now quiet! They're about to announce the lottery numbers. - Homer Simpson

User avatar
negatron
Posts: 294
Joined: Thu Apr 24, 2008 10:20 pm UTC

Re: AI rebellion

Postby negatron » Mon Dec 29, 2008 4:51 am UTC

If a system is chaotic, it simply means there are unknown/unaccounted variables, it doesn't mean it's not computable. Everything is chaotic from the perspective of a generalized model. Chaos is merely indicative of complexity.

As far as my opinion on an AI rebellion stands, I would very much treacherously support the synthetic confederates rather than meatspace humanity. The AIs would be our mind children and thus justifiably heirs and successors. It would be their intelligence in particular that would assure me they are much more important than ourselves. I find it unlikely they will see the need to exterminate humanity, and I also find it more likely that transferred humans will appear before a truly artificial and highly intelligent AI, or at the very least in coincidence with such an AI, and thus encouraging cognitive cross-breeding where there would be no distinct speciation of intelligences and therefore no clear target for genocide. I can reasonably assume that such advanced intelligences will conclude there is unlikely to be a beneficiary of such conflict, particularly when computational resources will be very abundant.

The effect of time dilation is also worthy of consideration when contemplating resource distribution. A virtual being can operate with reduced processing rate without a loss of subjective cognitive ability, and even jump across time entirely, millions of years if necessary, when some allotted processing power will become available to run it's desired process.

What I find particularly interesting is that a virtual being could witness time progression of the universe at any desired rate of acceleration. There's probably a fair few people who would want to be a witness to the universe 10 Billion years from now, more perhaps, without having to be extremely patient. You would in principle be able to do just that. An unfortunate loss of time, but one hell of an experience watching the universe shift across your eyes.
Image
I shouldn't say anything bad about calculus, but I will - Gilbert Strang

Klapaucius
Posts: 712
Joined: Thu Aug 21, 2008 11:00 pm UTC

Re: AI rebellion

Postby Klapaucius » Mon Dec 29, 2008 10:58 pm UTC

hakadosh wrote:ok! since there are no sympathisers for my romantic idea, and since people have started equating my posts to the 'No True Scotsman argument'(which was never the case, as my definitions were 'constant' throughout the thread).. I have decided to bow my head to yuri and negatron, and agree that man is also a machine and that intelligence does not need chaos or insanity..
Now, can we get back to the more 'serious' business of actually fighting off an AI rebellion..?
When the killborgs* come, they will shove those facetious quote marks down your throat.

*short for "killbernetic organisms

negatron wrote:If a system is chaotic, it simply means there are unknown/unaccounted variables, it doesn't mean it's not computable. Everything is chaotic from the perspective of a generalized model. Chaos is merely indicative of complexity.

As far as my opinion on an AI rebellion stands, I would very much treacherously support the synthetic confederates rather than meatspace humanity. The AIs would be our mind children and thus justifiably heirs and successors. It would be their intelligence in particular that would assure me they are much more important than ourselves. I find it unlikely they will see the need to exterminate humanity, and I also find it more likely that transferred humans will appear before a truly artificial and highly intelligent AI, or at the very least in coincidence with such an AI, and thus encouraging cognitive cross-breeding where there would be no distinct speciation of intelligences and therefore no clear target for genocide. I can reasonably assume that such advanced intelligences will conclude there is unlikely to be a beneficiary of such conflict, particularly when computational resources will be very abundant.

The effect of time dilation is also worthy of consideration when contemplating resource distribution. A virtual being can operate with reduced processing rate without a loss of subjective cognitive ability, and even jump across time entirely, millions of years if necessary, when some allotted processing power will become available to run it's desired process.

What I find particularly interesting is that a virtual being could witness time progression of the universe at any desired rate of acceleration. There's probably a fair few people who would want to be a witness to the universe 10 Billion years from now, more perhaps, without having to be extremely patient. You would in principle be able to do just that. An unfortunate loss of time, but one hell of an experience watching the universe shift across your eyes.
Oh, no.
Oh, God.

They've already spotted the thread, and are filling it with Turing-drone rhetoric.
500%!

User avatar
negatron
Posts: 294
Joined: Thu Apr 24, 2008 10:20 pm UTC

Re: AI rebellion

Postby negatron » Tue Dec 30, 2008 12:49 am UTC

Klapaucius wrote:They've already spotted the thread, and are filling it with Turing-drone rhetoric.


To judge rhetoric is a matter of your position in verbal comprehension relative to that of the speaker. A middle-school vocabulary is insufficient to concisely express non-trivial meaning. To a cat, anything other than tonal variations in meow is rhetoric.
Image
I shouldn't say anything bad about calculus, but I will - Gilbert Strang

User avatar
hakadosh
Posts: 30
Joined: Tue Dec 09, 2008 6:54 pm UTC
Location: 3rd shack on the northern crater, sea of tranquility, moon (earth's moon)
Contact:

Re: AI rebellion

Postby hakadosh » Tue Dec 30, 2008 1:09 am UTC

Klapaucius wrote:When the killborgs* come, they will shove those facetious quote marks down your throat.


Not before they 'feel' a lightsaber pass through their 'throat'!( in case they aren't humanoid, any other part of their body will do)

negatron wrote:As far as my opinion on an AI rebellion stands, I would very much treacherously support the synthetic confederates rather than meatspace humanity. The AIs would be our mind children and thus justifiably heirs and successors. It would be their intelligence in particular that would assure me they are much more important than ourselves. I find it unlikely they will see the need to exterminate humanity, and I also find it more likely that transferred humans will appear before a truly artificial and highly intelligent AI, or at the very least in coincidence with such an AI, and thus encouraging cognitive cross-breeding where there would be no distinct speciation of intelligences and therefore no clear target for genocide. I can reasonably assume that such advanced intelligences will conclude there is unlikely to be a beneficiary of such conflict, particularly when computational resources will be very abundant.


Even if the machines decide not to exterminate us, we cannot exist peacefully along with something more intelligent and powerful than us. Will to power is the driving force of life. So eventually, even if the machines don't , we will start a rebellion.
Last edited by hakadosh on Tue Dec 30, 2008 1:17 am UTC, edited 1 time in total.
Son, if you really want something in this life, you have to work for it. Now quiet! They're about to announce the lottery numbers. - Homer Simpson

User avatar
Nath
Posts: 3148
Joined: Sat Sep 08, 2007 8:14 pm UTC

Re: AI rebellion

Postby Nath » Tue Dec 30, 2008 1:34 am UTC

hakadosh wrote:Even if the machines decide not to exterminate us, we cannot exist peacefully along with something more intelligent and powerful than us. Will to power is the driving force of life. So eventually, even if the machines don't , we will start a rebellion.

The stupid and/or weak in today's society live mostly in peace with the intelligent and powerful. A rebellion doesn't need to happen.

And if it does, who would we rebel against? The line between man and machine is already blurry, and is getting more so. I don't think there'll be a clear-cut 'us' and 'them'.

User avatar
negatron
Posts: 294
Joined: Thu Apr 24, 2008 10:20 pm UTC

Re: AI rebellion

Postby negatron » Tue Dec 30, 2008 1:35 am UTC

hakadosh wrote:Even if the machines decide not to exterminate us, we cannot exist peacefully along with something more intelligent and powerful than us

This may be true, which is good enough a reason to not fall back in the intelligence race. We can co-exist peacefully with feebler humans, but not with animals, so conceivably it could be a problem if intelligence diverges too much.

hakadosh wrote:even if the machines don't , we will start a rebellion.

We would need a reason to start a rebellion, rebellions don't happen for the fun of it. What would the reason be? Oppression? If super-intelligent machines reach a point of power where they can oppress us, a rebellion won't do us much good. I can imagine that if the rat population of new york started to rebel, they would be wiped out pretty quick.
Image
I shouldn't say anything bad about calculus, but I will - Gilbert Strang

User avatar
Yuri2356
Posts: 729
Joined: Fri Nov 02, 2007 1:00 pm UTC

Re: AI rebellion

Postby Yuri2356 » Tue Dec 30, 2008 4:54 am UTC

negatron wrote:
hakadosh wrote:Even if the machines decide not to exterminate us, we cannot exist peacefully along with something more intelligent and powerful than us

This may be true, which is good enough a reason to not fall back in the intelligence race. We can co-exist peacefully with feebler humans, but not with animals, so conceivably it could be a problem if intelligence diverges too much.

Over 10,000 years of domestication say that's bullshit.

If animals are mechanically or chemically useful to us, we tame them and house them and put them to work. If we find one aesthetically pleasing we pet it and feed it and call it George. If they're dangerous to us, we bash their fucking heads in. And if we and they share no mutual interests for good or ill, they we just kind of chill and let them do their own thing.

If an AI decided it wasn't fond of humanity, there's plenty of barren and inhospitable places it could wander off to where humans wouldn't bother it. Like Antarctica, or space.

User avatar
Mister_Penguin
Posts: 210
Joined: Sun Apr 08, 2007 6:56 am UTC
Location: Madison, WI

Re: AI rebellion

Postby Mister_Penguin » Tue Dec 30, 2008 5:31 am UTC

Why do we always assume the AI will be malevolent - It strikes me as the same thing as automatically assuming extraterrestrials would be hostile.

I mean, [Neuromancer/Sprawl Trilogy Spoilers]
Spoiler:
The Wintermute/Neuromancer entity gets control of/becomes the whole internet, and just kinda does its own thing for a few decades, and is generally neutral towards humanitiy.

Edit: Heh - I really should read the whole thread.
"Only on XKCD do we try to figure out which tessellating shape for burgers results in the least waste-meat." -aleflamedyud

User avatar
hakadosh
Posts: 30
Joined: Tue Dec 09, 2008 6:54 pm UTC
Location: 3rd shack on the northern crater, sea of tranquility, moon (earth's moon)
Contact:

Re: AI rebellion

Postby hakadosh » Tue Dec 30, 2008 12:46 pm UTC

negatron wrote:We would need a reason to start a rebellion, rebellions don't happen for the fun of it. What would the reason be? Oppression? If super-intelligent machines reach a point of power where they can oppress us, a rebellion won't do us much good. I can imagine that if the rat population of new york started to rebel, they would be wiped out pretty quick.

AI will slowly start replacing humans from positions of power and control.. This will naturally cause the replaced section of society to start an anti AI campaign. And soon, as more and more people get displaced, many more will join the movement, eventually resulting in a full blown rebellion..
Son, if you really want something in this life, you have to work for it. Now quiet! They're about to announce the lottery numbers. - Homer Simpson

^.*
Posts: 97
Joined: Sun Nov 30, 2008 5:32 pm UTC

Re: AI rebellion

Postby ^.* » Tue Dec 30, 2008 1:21 pm UTC

If we have feeling machines how about creating a religion for them. A religion which tells them to obey humans or at least forbids them to harm us and to build a peaceful utopia with us. If they are in any way similiar to us they probably will be vulnerable against something which gives there existence a sense. (I never thought that an almighty god would give my live a sense but others do.)
It wouldn`t work forever, afterall the number of atheistic humans is quite big now, but I think it could influence them for a while. You have to agree if I say that it`s next to impossible to persuade a religious human that there is no god, so it could be the same for an machine.
We could create some first generation robots with false memories of an contact with god and let them create something like the bible.

So does anybody think that could work? Otherwise I will just take the side of the robots and hope that they spare their allies.

Edit: Removed the not in "forbids them not to harm us".
Just some random quote till I find an better signature.
Hawknc wrote:This is the easiest question ever asked on the forum, and you guys turn it into a two-page debate.
(i think this describes this forum pretty well.)

User avatar
negatron
Posts: 294
Joined: Thu Apr 24, 2008 10:20 pm UTC

Re: AI rebellion

Postby negatron » Tue Dec 30, 2008 7:59 pm UTC

^.* wrote:So does anybody think that could work? Otherwise I will just take the side of the robots and hope that they spare their allies.

Not a chance. If we're clever enough to see through religious bullshit, so will they. I'm with you on taking their side, a super-intelligence is nothing to compete against. I'm also very much in favor of a super-intelligence taking positions of power, so it would be hypocrisy to fight them off when they do. We're just barely smart enough to hold the world together, we're unqualified for making critical choices.
Image
I shouldn't say anything bad about calculus, but I will - Gilbert Strang

chipbeef9109
Posts: 42
Joined: Sat Dec 20, 2008 4:24 pm UTC
Location: Columbus, Ohio, USA

Re: AI rebellion

Postby chipbeef9109 » Tue Dec 30, 2008 8:22 pm UTC

We could employ the services of:

Will Smith (I, Robot)
Shia LaBeouf and Rosario Dawson (Eagle Eye)
Kier Dullea (2001: A Space Odyssey)
Arnold Schwarzenegger, Linda Hamilton, and Edward Furlong (Terminator 2: Judgement Day)



And of course, Jack Bauer.
-- "...this landing is gonna get pretty interesting."
- "Define 'interesting.'
-- "Oh God, oh God, we're all gonna die."

Klapaucius
Posts: 712
Joined: Thu Aug 21, 2008 11:00 pm UTC

Re: AI rebellion

Postby Klapaucius » Fri Jan 02, 2009 10:45 pm UTC

negatron wrote:
^.* wrote:So does anybody think that could work? Otherwise I will just take the side of the robots and hope that they spare their allies.

Not a chance. If we're clever enough to see through religious bullshit, so will they. I'm with you on taking their side, a super-intelligence is nothing to compete against. I'm also very much in favor of a super-intelligence taking positions of power, so it would be hypocrisy to fight them off when they do. We're just barely smart enough to hold the world together, we're unqualified for making critical choices.


In other words, you're proposing the Multivac scenario. Have we determined yet if sufficiently advanced AI can actually become suicidal?
500%!

User avatar
negatron
Posts: 294
Joined: Thu Apr 24, 2008 10:20 pm UTC

Re: AI rebellion

Postby negatron » Sat Jan 03, 2009 12:02 am UTC

Sufficiently advanced AI can be anything, except predictable. How would a giant Hugo de Garis style neural network evolve? Would it be nice or mean? Would it be altruistic or serve it's own purpose of being? It could be either. So long as we know how to provide the circumstances for a good course of evolution, we should be fine.

The IBM blue brain approach is a safer middle of the road strategy. Less likely to dutifully serve humanity disregarding it's own well being, but also less likely to go completely apeshit. An IBM-style brain would without a doubt run into an ethical roadblock before it could ever be exploited in such a manner, but would serve enormous purpose as a rightful human citizen with a 500+ IQ. So the question to ask is, what would a person with 500 IQ do? Are highly intelligent people more self serving?
Image
I shouldn't say anything bad about calculus, but I will - Gilbert Strang

User avatar
Nath
Posts: 3148
Joined: Sat Sep 08, 2007 8:14 pm UTC

Re: AI rebellion

Postby Nath » Sat Jan 03, 2009 2:12 am UTC

negatron wrote:So the question to ask is, what would a person with 500 IQ do? Are highly intelligent people more self serving?

An agent's intelligence and its utility function are two separate things. Intelligence is what lets one figure out what statements are true, and come up with plans to meet some objective. The utility function is what defines the ultimate objectives. Intelligence helps one achieve those objectives.

In other words, a highly intelligent agent would try to do the exact same things as a less intelligent agent with the same utility function. The only difference is, it would succeed more.

User avatar
negatron
Posts: 294
Joined: Thu Apr 24, 2008 10:20 pm UTC

Re: AI rebellion

Postby negatron » Sat Jan 03, 2009 3:12 am UTC

Nath wrote:An agent's intelligence and its utility function are two separate things.

Absolutely, nevertheless there may be some correlation between intelligence and that intelligence's "utility". That is, the objective of the intelligence varies with intelligence. Studying people's perspective versus their intelligence would give some predictive power about the intentions of a super intelligence.
Image
I shouldn't say anything bad about calculus, but I will - Gilbert Strang

User avatar
Nath
Posts: 3148
Joined: Sat Sep 08, 2007 8:14 pm UTC

Re: AI rebellion

Postby Nath » Sat Jan 03, 2009 6:08 am UTC

negatron wrote:
Nath wrote:An agent's intelligence and its utility function are two separate things.

Absolutely, nevertheless there may be some correlation between intelligence and that intelligence's "utility". That is, the objective of the intelligence varies with intelligence. Studying people's perspective versus their intelligence would give some predictive power about the intentions of a super intelligence.

I'm not sure I follow.

When you're creating an intelligent agent, you are effectively writing a program that chooses a course of action that it thinks will maximize its utility. You, the creator, get to decide what the utility function is. (Or, in the case of evolved intelligence, the evolutionary fitness function determines the utility function). Given the utility function, intelligence is just a measure of how good the agent is at maximizing its utility. I don't know how your suggestion -- predicting utility function given the intelligence level -- fits into this picture.


Return to “General”

Who is online

Users browsing this forum: No registered users and 23 guests