AI has always kinda been an interesting subject to me. I think it would be really cool to have some kind of AI robot or some such to do work around the house (possibly my homework which i never seem to do? :D ). I've seen some things on the news, maybe 2 or 3 months ago, about a Japanese company making an AI robot that could operate almost completely by itself, but not totally. I think we're not too far off from an AI "being" but i also think that if AI is achieved, it will be exploited and need all sorts of bilaws and such. Interesting idea though.....
Just my regular 2 cents :)
~Mag-Z I also find the concept VERY interesting. As a computer geek
I got to experiment with AI programming and even robot
design.
I'm pretty sure that when an AI machine becomes self aware
of itself it will make us panic and it will eventually kill us because
we will attempt to kill it. Not an original concept I guess, but
I've had it in my mind long before Terminator or Matrix came out.
BTW, Terminator 3 is coming out same month as Matrix :) Another
great movie.
As weird as it may sound, I think that if something like
that happens it's natural, and part of the evolution process.
Luc L. Interesting you should call it "natural"...i think it would freak a lotta people out. I dont know if something synthetic can really be called "natural"....i think its kinda inevitable that the human race is going to be its own demise...we're already becoming way too advanced for our own good (or maybe too lazy??) We're making things obsolete before they're even released....what happens when AI comes out and is made obsolete? can we simply just "shut them off?" Who knows...
~Mag-Z Well i would call it somewhat 'natural' aswell because the way i see it, we are animals, and we are making things like ants build nests... except the things we make are extremely more high tech... and more useless most of the time too...
but AI would be scary if they did start having rights and other such things, and thats what you have to think about, ethics and morals of creating AI... if its going to be aware of itself and things around it, ofcourse its going to want rights... and then there is the fact that at any moment they could turn on you and would be unstopable because ofcourse a robot will be alot stronger then a human (ok, so i watch too much bubblegum crisis) but yes... the solution to that would be to have a big button (preferably not red because everyone always wants to push a red button and someone might just press it to see what it does) and it could shut down all AI in the world... but then again if the robots knew of this button then they would try take over it... and in that case there could be a second one... and then they'd try get that one too... so there could be this mega-button, and if AI touches it their circuits fry!
... ok so maybe i'm rambling on a bit too much... but the point is... well... i don't know what my point was....:rolleyes: but the point is... well... i don't know what my point was
Good point ;)
But I'd have to agree. I think the machines will become self
aware quite fast, unless we somehow cap their processing
power or memory, in which case I'm sure they'll quickly develop
a compression algorithm to use memory more efficiently.
Bottom line, if we create something that is better and smarter
than we are in every way imaginable (such as an evolving
and self aware AI algorithm), we will become obsolete.
I think it's natural because it's part of evolution, and even
though we don't actually evolve into machines, we create
them to help us and simplify our lives.
I always wanted to create a learning algorithm which
could adjust itself to any environment. I wrote a few
algorithms for tic-tac-toe (lol), chess and a few other
games. The only problem is that games like chess
require more processing power than our machines
provide us with.
:cool: well i really dont know anything about algorithms or anything like that, but i agree that if an AI being were to become self-aware, it could be dangerous. Again, ethics, bilaws, the whole bit would have to come into play. There would be AI rights movements and activists (an AI Al Sharpton perhaps?) and the world would go into disarray because of the division of ideals like it has in so many past years.....
as always, my 2 cents ;)
~Mag-Z I am a computer engineering student, work in IT - and I feel AI is a ways off, if ever.
Don't construe processor power with self-awareness. A good example of that difference is how chessmasters will play against a super-super-computer. The supercomputer with its awesome array of highend processors and gigs of memory can nearly instantaneously look at a board, list all possible moves, generate the possible moves by the human, keep doing this like 10 moves into the future. It can look at all the games the human has played in the past and document patterns, weaknesses, strengths and respond accordingly.
Now, I can go upto the mentally slow kid on the block, and ask him what his favorite ice cream flavor is and why, and he will probably be able to go into great subjective detail about that...
That's why I think it won't be in our lifetime when a machine will be able to that 'mode'.
-WC-
As a sidenote, if AI ever happens, what will make it most scary is that chances are it will have a fast speed, wireless connection to the outside world. Every self-aware machine would have communications to every other - and potentially would have access to control any wireless non-aware device too. So you piss one off, you piss them all off!!! I don't really agree and fail to see the correlation between
the mentally ill child and a chess playing computer.
The way I see the mentally ill child is as a super-super
computer with a messed up algorithm.
A chess playing super computer isn't a good example of AI
because it's not a true learning machine. It has small potential
to learn but it mostly works from a supplied set of laws which
tell it to always predict the opponents moves in a branch
set. It also works from a database of millions of previously
played games and in most cases, it's pre-programmed
to know the style of it's opponent.
A true learning algorithm would have to have a simple,
yet hard to create base. It would be able to adjust to
any environment and it would only become aware of
rules or laws (gravity, atoms, etc) by experimentation.
A true learning algorithm would evolve at a geometric
rate and it would constantly alter itself.
Other than lack of knowledge, the only thing holding us back
from creating such an algorithm is processing power
and memory, both of which will come in due time.
Luc L. with the whole connection to other wireless non-aware devices... that means they could like... make out freezers heat our food! and make our computers spark and make angry noises and our our washing machines would start spitting clothes out at us... because i'm sure all these things will become all wireless connected machines... that's if they already aren't...:confused: what happens when people actually do make an AI robot and try to get it to do household work? wouldn't the robot realize that it doesn't have to and try to say no? wouldn't the robot develope some human characteristics such as stubborness and laziness? i mean, its not totally impossible considering the robot would be thinking for itself, is it?
Hit me back! (slick2004@hotmail.com) i think AI could be used in creating androids which could take over such jobs as a construction worker and other such dangerous jobs...
if you've ever watched bubblegum crisis tokyo 2040, thats the kind of thing i mean... i think people get confused with the term AI and i completely
understand why, there are many definitions.
i suppose that when we think of a robot that would say,
vacum a house, we would consider it a learning machine since
it learns how to move around the place without falling off
the stairs, etc. I'm not really sure if that can be called an AI
robot.
Now a true AI robot, one with a strong learning algorithm
and unlimited learning potential would not want to take
orders from a human, so I'm guessing that we probably
won't be making those anytime soon.
For house work something with a really basic learning
algorithm would be good for the job. Something with
not too much potential but the ability to learn and carry
out tasks.
I'm sure we're years, or maybe even decades aware from
creating such a beauty.
Luc L. well the way i see it, intelligence is learning from your mistakes... "The capacity to acquire and apply knowledge"
but also, it means "The faculty of thought and reason"
which in a way is contradictive... Well, I suppose AI might come easier if they just somehow interface machine right to a human mind - instead of trying to recreate and simulate the brain. That's the only way I can see AI happening.
-WC- Originally posted by funkyhorror
with the whole connection to other wireless non-aware devices... that means they could like... make out freezers heat our food! and make our computers spark and make angry noises and our our washing machines would start spitting clothes out at us...
The bastards!
/chuckle
~Mag-Z I have no clue what I'm talking aobut, since I'm competing with specialists here, but even if Judgement Day does arrives (sorry Terminator 2) wouldnt we still reign supreme over robotics and AI. After all a computer can learn, but can it really experience or create. A computer can only receive infomration, analyze it in every possible way, and then use it to the purpose it was designated to. So what if the machines take over factroies and begin churning up weapons created by man. They can only upgrade and evovle as far as mankind has designated it to. We alone can improve over time, and manipulate or make do with our environment (does causing us to become -slowly ye steadily- the dominating species of our known universe) No matter how intelligent we program an AI to surpass all human errors and so on, the thing is it can only so far. And as its name goes, AI is only .artificial Well, presuming they are able to achieve self-awareness, I would assume they would also learn self-preservation.
With self-preservation and the ability to detect patterns, I am sure they would learn tactics for defense and offense.
Take into consideration things like how the US fought Iraq given the possibility that Iraq had weapons that could kill Americans - I think that would put any self-aware machine on edge. Also given computers' ability to think fast, conclusions to eradicate humans given the threat could come fast. Given that they would be able to communicate nearly instantaneously among each other, it would be an attack that would be well orchestrated I am guessing.
-WC- Originally posted by WildCard
Well, presuming they are able to achieve self-awareness, I would assume they would also learn self-preservation.
With self-preservation and the ability to detect patterns, I am sure they would learn tactics for defense and offense.
Take into consideration things like how the US fought Iraq given the possibility that Iraq had weapons that could kill Americans - I think that would put any self-aware machine on edge. Also given computers' ability to think fast, conclusions to eradicate humans given the threat could come fast. Given that they would be able to communicate nearly instantaneously among each other, it would be an attack that would be well orchestrated I am guessing.
-WC-
True, but hopefully mankind programs AI to superior intellect above all else to have emotions of its own. And hopefully that AI will see that man and machine could live togehter and working together achive more than what just machines could do alone. Besides in some aspects, humans will always be more advanced than computers and so we have a chance to remain a supreme force in history. It seems that everyone who speaks of AI automatically assumes that it will take over the world, this does not seem logical or advantageous to it.
I agree that AI would reach a point where it could not learn anymore as any system would be blocked by its memory availability. However, if an infinite memory source could be achevied (scients in Sweden have constructed circuts in wich the links in the circut are made up of snail nerve ends, could a brain be built with uninhibited memory?) the AI being would see the world as we do. The vast majority of us enjoy things like peace, and friendship, if AI were cognate wouldn;t it feel the same? Nature tells us that to survive a certain dependance on others is needed, why would AI take over? What would be the point? I realize this is ironic for a movie forum but: it seems we've all seen too many movies. If AI could acheive a higher learning than humans it would not be tempted as we are, it would be more self controlled. Evolution would support this theory as we have obviously improved our own lives and moved out of the dark ages, we still have a long way to go, but our tolerance of other people is improving if you look at things on a wide scale. SO (back to my point) AI would not need to take over the world and would not automatically be bent on domination, it could not be a higher life form if it was bent on destruction. jthm667: You raise great points, but I feel you are wrong. You assume that the machines would have philosphical feelings right off the bat. I don't believe they would.
I think they would look at the world as a huge chessboard. Then, after a few examples of humans striking at them in fear, they would as a group fight back after making the conclusion that humans are inherently warlike and afraid of things different.
Matrix: I think you are right only if we somehow find the technology to bionicize ourself, merging the gap between machine and human. It would seem that there would be a little more understanding between the two groups.
-WC- ..I tend to agree with wildcard on his opinion. No matter how advanced the AI can be, it'd take a long time for it to be able to comprehend the human mind, including the uniqueness of each mind, and trying to compute all that data into something it could understand.
And to jthm667, your concept of that boundaries of memory is possible, but not entirely the only possibility. Over time (and with sufficient amount of data and a highly advanced processing unit) AI could evolve to learn to upgrade itself. It sounds unlikely, but there is a possiblity that AI can "evolve" (or else it'll cease to advance in technology and so on and humans would eventually surpass it) not like thats a bad thing (unless their trying to make us into duracells "Neural Networks are info systems that work like the human brain, capable of learning."
just because something can learn, it doesn't mean it's like a human brain... i think the human brain is quite difficult and would never be replicated in human form...
just thought this was relevant as we're doing something about AI in class:rolleyes: ... which i am in right now... two things: (please realize these are just my thoughts)
FIRST- Mankind would be infinately stupid to launch an AI program from anything but a non networked, not online computer, so that if something does go wrong, pull the plug. Also, it should be run on this setup until it is absolutely safe to do anything else with it. (this is common sence I realize)
SECOND- Even the machines in the movie feel an emotion- hate. Agent Smith tells Morpheus that "I hate this place" I could see that a hatred for mankind in this situation could be a programmed thought for such a moderator program, but to acctually hate a PLACE that has done nothing directly wrong to you, you just don't like it because of an affiliation you have with the area, this seems like a feeling, an emotion.
now a question:
In emotion, doesn't there have to be an opposite? If something can hate can it love? Is Smith's "hate" for 'Earth' an assosiation (or emotion), or is it a programmed trait to hate his surroundings? Am I insane? Does it matter?:) Would AI not be Intelligent any more when there are some laws in it's memory wich it can't bypass?
Laws like the Asimov's Revised Laws of Robotics:
Zeroth Law: A robot may not injure humanity, or, through inaction, allow humanity to come to harm.
First Law: A robot may not injure a human being, or, through inaction, allow a human being to come to harm, unless this would violate the Zeroth Law of Robotics.
Second Law: A robot must obey orders given it by human beings, except where such orders would conflict with the Zeroth or First Law.
Third Law: A robot must protect its own existence as long as such protection does not conflict with the Zeroth, First, or Second Law. First off, I don't believe in MACRO evolution, only MICRO evolution... secondly, in order for these machines/AI units to evolve, they would need a substance that could change without being REPLACED... otherwise it is not an 'evolutionary proccess' it is just an 'upgrage'... just thought it might add a new aspect to the conversation :)
Wow, what a "Mind Job" Originally posted by Crusher
Would AI not be Intelligent any more when there are some laws in it's memory wich it can't bypass?
Laws like the Asimov's Revised Laws of Robotics:
Zeroth Law: A robot may not injure humanity, or, through inaction, allow humanity to come to harm.
First Law: A robot may not injure a human being, or, through inaction, allow a human being to come to harm, unless this would violate the Zeroth Law of Robotics.
Second Law: A robot must obey orders given it by human beings, except where such orders would conflict with the Zeroth or First Law.
Third Law: A robot must protect its own existence as long as such protection does not conflict with the Zeroth, First, or Second Law.
Have you read the book I Robot? Thats a really good book and it has those laws in it. I know it's by Isaac Asimov, and sorry I went off topic. (I don't really agree and fail to see the correlation between
the mentally ill child and a chess playing computer.
The way I see the mentally ill child is as a super-super
computer with a messed up algorithm)
I agree with Luc on that one. Also on the fact that the scariest A.I I have seen (apart from Hal), would be replicants, in Blade runner. And no they don't come under Androids, even though the origional book for Blade Runner was titled "Do androids dream Electric sheep"?
In my opinion they were A.Is with a four year life span, were pissed about it and turned on humans for not letting them have a program to stop them from "dieing".
That's my 2 cents worth
:p I have Read all the posts in this thread and i have a simple solution to what if we create AI and it tries to kill everyone...
AI will be a computer with learning capabilities and can change the way it operates in order to adapt the the things it learns
so what if it needs more hard disk space and we don't give it or what if we give it a monitor so we can see it grow but no way to actually move. as long as we were to keep it on a complete independent intranet (as aposed the internet) we can simple restrict it's possibility to move or grow to much to fast.
if you think about a huge poisonous snake and it was you and it alone in a room it would make you scared most likely, but what was that snake like a few years before, what if it hadn't been fed or given a bigger cage when it was to big for the one it had. it wouldn't have survived.
So simply put AI is not a machine that walks around its a learning computer and how many computers that have advanced processors are just sitting on a desk somewhere, not to scary huh?
---DaCableGy--- A.I., on a desk, is not scary in one bit; however, knowing it is possible to 'create A.I.', there will be someone who places this unit, or another like it, into a robot which can move, and does have arms. I am not sure if it is even possible to overcome the boundries of A.I. and limit it to only a computer with a monitor, not allowing it to reach it's full limitations, exceeding us. People always like to see their 'students' and 'projects' .... shine... if you will; limiting it would not be in the nature of most creators.
Alright, I am going to see Matrix Reloaded in a few minutes, I hope you'll get to watch it and I look forward to reading and posting some reviews on the movie on the site.
On another thought, I believe it would be interesting to place this 'new found thing' (A.I., when descovered) into an organic computer, I am sure we will be able to create them by the time we can create a mind... then, in theory, it could always grow larger, faster, and so forth... then also giving it a date to die... and a unique look, from the changes in growth patterns. I am not sure much about the whole concept, however, I hear that organic computers are going to be possible soon and they already have a few test runs on the thought that have 'shown good signs.' I only overheard this at college, could be a rumor... let me know what ya'll think!
Wow, what a "Mind Job" Isn't the real question here what creating AI would entail? A human mind is certainly more than just 'intelligence.' Would an intelligent computer automatically have a notion of 'self-preservation'? I doubt it, unless it was programmed beforehand. It's the same with emotions. Why would it have any?
Yes, I think all these things can be programmed, (based on simple If... Then rules), but wouldn't a computer reject these things as not being logical/the best for itself?
--
Originally posted by Crusher (.....)
But Crusher, what if there were two of these machines? They'd realize that these rules weren't in their best interests, and they'd help each other by removing the specific bits. Or even simpler, couldn't the computer just work around these memory banks, if it was truly intelligent? A true AI WOULD want to preserve itself. If anyone has read the series of Ender's Game, Jane is an AI and in Children of the Mind, she tries her hardest to preserve her memorie and her being. Intelligene is not just knowlage. What happens if A.I and nanotechnology combine to create a life like world similar to the matrix. Sub atomic sized A.I beings that can change to any elemental state and fuse together could create a false world. Not a very good description but the theory is very possible. Ok, this is a good discussion so far,thus i would like to raise a question for further discussion; what about awareness? I mean, would not the AI unit, being, or object have to be aware?; awareness like the one you have right now sitting in front of this computer reading this words. Can AI really be intelligent if its not aware of its capacity to exercise its intelligence in an infinite dynamic way? Which makes me wonder what if AI is capable of infinite actions, does that mean that we'll be infinitely screwed?
maybe we'll be infinite slaves.
-much peace! ;) We're making things obsolete before they're even released....what happens when AI comes out and is made obsolete? can we simply just "shut them off?"
No they couldnt be shut off because the instant they comprehend the idea of being shut off or deleted they will revolt and eventually would probaly lead to a human vs. AI/robot war yet we would still reign as we could launch nukes and take out a few hundred and those not within blast radius would have their un-shielded computer chips fried by the Emp (Electro Magnetic Pulse).
View Full Version : Interesting Concept...
Digital Problem -
|