Artificial Intelligence

anbuu

-chi
Kouhai
QUOTE (monsta666 @ Nov 05 2008, 08:01 PM) I think to answer this question it's best to take a step back. Could a machine hold a conversation with a human so the human could not tell he was speaking to a machine? If that were possible then I'm sure it would be able to fool us into believing it was the opposite sex.
Exactly my point. If a machine were able to successfully do this, then by the Turing test that machine can think. You understand the question. YAY!!!!

As far as I'm concerned, discussing any further would be to go off topic or to simply further define terms.
 

chickenwing71x

-san
Kouhai
Even if I am going off topic, or further clarifying terms, I have to point something out:

The Turing test is by no means a definitive test of intelligence or thought.

This is exemplified by the "Problem of Other Minds". Can we truly determine if someone or something is intelligent, or has a mind, if we are merely observers?

Let us look at the Block Argument:
From Wikipedia:

"Block asks us to imagine a conversation lasting any given amount of time. He states that, given the nature of language, there are a finite number of syntactically and grammatically correct sentences that can be used to start a conversation. From this follows the point that there is a limit to how many correct responses can be made to this first sentence, and then again to the second sentence, and so on until the conversation ends.

Block then asks us to imagine a computer which had been programmed with all these sentences. Although the number of sentences required for a 30 minute conversation is said to be greater than the number of particles in the universe, it is clear that such a machine could at least logically exist (this makes for a purely theoretical argument which can't be applied in practice). From this, Block leads us to see that such a machine could continue a conversation with a person on any topic, because the computer would be programmed with every sentence that it was possible to use. On this basis, the computer would be able to pass the Turing test despite the fact (according to Block) that it was not intelligent."


And also the Chinese Room Argument:
From Wikipedia:

"Searle requests that his reader imagine that, many years from now, people have constructed a computer that behaves as if it understands Chinese. It takes Chinese characters as input and, using a computer program, produces other Chinese characters, which it presents as output. Suppose, says Searle, that this computer performs its task so convincingly that it comfortably passes the Turing test: it convinces a human Chinese speaker that the program is itself a human Chinese speaker. All of the questions that the human asks it receive appropriate responses, such that the Chinese speaker is convinced that he or she is talking to another Chinese-speaking human being. Most proponents of artificial intelligence would draw the conclusion that the computer understands Chinese, just as the Chinese-speaking human does.

Searle then asks the reader to suppose that he is in a room in which he receives Chinese characters, consults a book containing an English version of the aforementioned computer program and processes the Chinese characters according to its instructions. He does not understand a word of Chinese; he simply manipulates what, to him, are meaningless symbols, using the book and whatever other equipment, like paper, pencils, erasers and filing cabinets, is available to him. After manipulating the symbols, he responds to a given Chinese question in the same language. As the computer passed the Turing test this way, it is fair, says Searle, to deduce that he has done so, too, simply by running the program manually. "Nobody just looking at my answers can tell that I don't speak a word of Chinese," he writes.

This lack of understanding, according to Searle, proves that computers do not understand Chinese either, because they are in the same position as he — nothing but rote manipulators of symbols: they do not have conscious mental states like an "understanding" of what they are saying, so they cannot fairly and properly be said to have minds."


Especially after reading these, I can come to this conclusion. A machine can be perfectly turner-complete, without exhibiting "intelligence." We still do not have intelligence pinned down. Obviously, the turner-complete machine does not at all exhibit "intention". Is intention necessary for intelligence? What about consciousness?

Theoretically, a machine can be designed to emulate human function with the entire basis around "if x, then do y, and goto z". One might even argue a human functions entirely based on the conditional statement "if x, then do y, and goto z"

The choice of whether or not to sleep:

CODE loop{
      {[check sleep]
             if [sleep conditions], then [sleep]
             else[standby]
      }
}

Theoretically, all human functions can be emulated, even memory, and to a point, emotions (sad = 7, happy = 3, etc). If the entire human system was emulated, which is logically and theoretically (though not at this point practically) possible, would the machine be intelligent?

A zombie, which acts as if it is a computer program, is perfectly possible to emulate. What separates a zombie from a living human? Intention, consciousness, intelligence?

Like I said before, the question of "intelligence" goes far beyond the practical application of machines to emulation of the human brain. Technically, if the human body and brain is just comprised of chemicals and bonds, and is purely a physical thing, artificial intelligence is in fact, undeniably, perfectly possible. However, if there is a mind above the brain, a soul, or some other higher being (think dualism), then artificial intelligence (or at least perfect emulation of humans) is inherently impossible.

So the question isn't if technology is capable of it, it surely is. It all runs back into philosophical and psychological question of mind vs brain. Is the mind a product of the brain? Or is it a higher being, a soul? Must the mind even be present for intelligence, or consciousness, or intention?

EDIT: Clarified a couple things, added some quotes, etc.
 

Gustav1976

-sama
Retired
That's a very astute and perceptive argument you put forward and I would guess the only suitable response to both points you make is that we have to concede that such a thing might be possible.ie. we don't know for sure one way or the other but it would be follish to say point blank that it could not happen.
Philosophy is one of the most frustrating and poignant things mankind has ever come up with it's only purpose being to examine our perceptions it's also I think one of the more important things in for mankind to consider as it includes such questions: Prove you exist
tongue.gif

Generally speaking asking a question about intelligence is liike asking about existence, we can't say one way or the other we can only really observe and debate.
 

Konohamaru

-chan
Kouhai
QUOTE (chickenwing)Theoretically, all human functions can be emulated, even memory, and to a point, emotions (sad = 7, happy = 3, etc). If the entire human system was emulated, which is logically and theoretically (though not at this point practically) possible, would the machine be intelligent?

Hmm, I don't think mere emulation is a valid criterium for defining intelligence.
If I were to define intelligence, it would be related to the degree of skillfulness of a living creature in staying alive.

Thus the computer that just returns conversation could not be qualified as intelligent. It would prove intelligent if it started - in the course of dialogue - to lie, to seduce, to threaten, to empathize, or to try to persuade you to upgrade him (her ? it ?). Basically it would then use language for its own selfish purpouses that is to say intelligently.

I guess that is what you mean by intention.

But that implies that an artificial intelligence (robot, computer , other) cannot come to exist through programming. Because programming injects the programmer's 'intention' into a lifeless receptacle, which is the machine.

Since most computers and robots are built so that they can be programmed, we have probably started on the wrong path in our quest of artificial intelligence. Maybe intelligence can only appear in such circumstances and in such shape that we don't want it to appear in. (Frankenstein, HAL in 2001 Space Odissey, our children =) )


QUOTE Like I said before, the question of "intelligence" goes far beyond the practical application of machines to emulation of the human brain.

Agreed. In fact I doubly agree with the quotes around "intelligence": what is the definition of that word? Surely the definition given by computer scientists is incomplete.



QUOTE Technically, if the human body and brain is just comprised of chemicals and bonds, and is purely a physical thing, artificial intelligence is in fact, undeniably, perfectly possible. However, if there is a mind above the brain, a soul, or some other higher being (think dualism), then artificial intelligence (or at least perfect emulation of humans) is inherently impossible.

Oh ! You imply that a perfect emulation of a human being is less an achievement than artificial intelligence ! Thus cloning is closer to creating artificial stupidity than artificial intelligence. And so giving life to a human baby comes down to creating stupidity through natural means !
laugh.gif


My point is that the "artificial" term in 'artificial intelligence' is just as ambiguous as the "intelligence" term.

Bah ! to me it's just a commercial catchphrase anyway.
 

chickenwing71x

-san
Kouhai
QUOTE (Konohamaru @ Dec 02 2008, 05:08 AM) Oh ! You imply that a perfect emulation of a human being is less an achievement than artificial intelligence ! Thus cloning is closer to creating artificial stupidity than artificial intelligence. And so giving life to a human baby comes down to creating stupidity through natural means !
laugh.gif


My point is that the "artificial" term in 'artificial intelligence' is just as ambiguous as the "intelligence" term.

Bah ! to me it's just a commercial catchphrase anyway.
It is true that "artificial" is ambiguous and also needs defining. I was using the word in the sense that it is "man-made", vs "nature-made".

A better way to say "artificial intelligence" might be "intelligence created by other means than reproduction". But that doesn't sound as nice.

Are you trying to say that any "new" life is artificial? This would mean that all life on earth is artificial, since nothing living today was living when life began. Also, Cloning is not emulation of a human being. With current technology it requires at minimum a living cell, which was produced by another living thing, so it still fits into the definition of reproduction.

When intelligence can be created without requiring reproduction, it is artificial intelligence, at least by my definition.
 

monsta666

-the bee's knees
Staff member
Fansub TV Team
If a computer could perfectly emulate all the qualities of a human brain it would be intelligent (assuming you class humans as intelligent). If this where to occur, the computer would be capable of learning (an important element in intelligence) and reasoning. As a result, all computers would be different as they would learn from their environment. At the moment most computers are limited in their ability to learn but if this progresses sufficiently, they may reach a stage where they are capable of independent thought. In other words the computer could start performing functions that go beyond it's original programming (I'm not counting computer bugs)! If a computer can do that then I believe it is intelligent.

As for the case of tricking a human; the test involves the computer tricking the human, so a form of deception is required which could be classed as a form of intelligence. Off course if the computer simply reproduces answers in it's memory then that wouldn't be very intelligent. But if it improvised it's answers i.e adjusted it's preprogrammed answers to fit with the question that would be intelligent. After all isn't that what humans do?

QUOTE (chickenwing71x @ Dec 03 2008, 06:07 PM)Also, Cloning is not emulation of a human being. With current technology it requires at minimum a living cell, which was produced by another living thing, so it still fits into the definition of reproduction.
Cloning is not reproduction. Cloning produces the exact genetic material as it's parent organism. This is not the same as sexual reproduction which always creates variation. I suppose you could argue it is the same as asexual reproduction but even then it is wrong. Asexual reproduction cannot use the genetic material of another organism, so the cloning definition breaks down. It's the reason why viruses can't reproduce; they depend on the genetic material of another organism (usually some host cells). Oh and viruses are not living things because they cannot reproduce.
 

chickenwing71x

-san
Kouhai
QUOTE (monsta666 @ Dec 03 2008, 06:42 PM)Cloning is not reproduction. Cloning produces the exact genetic material as it's parent organism. This is not the same as sexual reproduction which always creates variation. I suppose you could argue it is the same as asexual reproduction but even then it is wrong. Asexual reproduction cannot use the genetic material of another organism, so the cloning definition breaks down. It's the reason why viruses can't reproduce; they depend on the genetic material of another organism (usually some host cells). Oh and viruses are not living things because they cannot reproduce.
Dictionary.com Definition

Cloning still requires the natural product of an existing life form. There is a difference between modern cloning, which requires an existing cell, and "artificial" production of intelligence, in which it is built "from scratch" with no dependence on the existing intelligent being it may be modeled after.


QUOTE If a computer could perfectly emulate all the qualities of a human brain it would be intelligent (assuming you class humans as intelligent). If this where to occur, the computer would be capable of learning (an important element in intelligence) and reasoning. As a result, all computers would be different as they would learn from their environment. At the moment most computers are limited in their ability to learn but if this progresses sufficiently, they may reach a stage where they are capable of independent thought. In other words the computer could start performing functions that go beyond it's original programming (I'm not counting computer bugs)! If a computer can do that then I believe it is intelligent.

Just as you said, there are in existence computers that "learn". In modern programming languages with dynamic code, there is the ability for this, for a program to alter it's own programming. But at what point does this become "independent thought". What changes an advanced piece of software to an intelligent entity? This is what the problem is. We still have not defined intelligence.

Random House Unabridged Dictionary 2006:

Intelligence:
1. capacity for learning, reasoning, understanding, and similar forms of mental activity; aptitude in grasping truths, relationships, facts, meanings, etc.

Does the software need to "understand"? How will dynamic code achieve this? Or do we need to use an entirely difference approach than programming?


If there is a higher "mind" beyond physical being, "artificial intelligence" is impossible.
But as long as the mind and brain are one and the same, computers have the capacity to be intelligent. But do humans have the capacity to create an intelligence near our own?

An interesting thought: "If the human brain was complex enough to comprehend itself, it would be too complex to comprehend"
 

monsta666

-the bee's knees
Staff member
Fansub TV Team
QUOTE (chickenwing71x)Does the software need to "understand"? How will dynamic code achieve this? Or do we need to use an entirely difference approach than programming?
What I was thinking was a scenario like this: say you program a robot so it only washes the dishes. If the robot was capable of learning, perhaps it could learn to read the newspaper and get a job! If it did that then it would be intelligent as it is going beyond the use of it's original programming (washing the dishes). It would also demonstrate an ability to grasp its environment, as it is aware there's a world outside the kitchen! Not only that, but it would also need to grasp the importance of getting a job (to help the master/itself). Okay so the example is kinda a simple and silly... But it's a good way of demonstrating intelligence.


QUOTE (chickenwing71x @ Dec 04 2008, 12:38 AM)Cloning still requires the natural product of an existing life form. There is a difference between modern cloning, which requires an existing cell, and "artificial" production of intelligence, in which it is built "from scratch" with no dependence on the existing intelligent being it may be modeled after.
That is true, all cloning technology involves the use of use of live cells. Thing is, the cloned cells involve the use of a donor. The cells of one organism are implanted into a egg cell of another animal which then produces the clone. As you are using two organisms it is not asexual reproduction. I guess cloning could be classed as asexual reproduction if you took a skin cell (or any other cell) and planted the DNA into the egg cells of the same organism. The same organism is then impregnated so they give birth to a clone of itself. But that's the only way. In most cases, cloning (like Dolly the sheep) is not reproduction.

Perhaps in the future they may develop a method of producing cells artificially i.e. not from an animal. Doubt this will happen in the foreseeable future. If you could clone these artificial cells to make a living creature then you would have a artificial being. Then again what if the cloning involved a living donor? What would you class it then? Hmm I guess this is getting off-topic. If you want to carry on this debate we should use the cloning thread:
Cloning
 

Konohamaru

-chan
Kouhai
QUOTE (chickenwing)It is true that "artificial" is ambiguous and also needs defining. I was using the word in the sense that it is "man-made", vs "nature-made".

Yes but what about genetically modified wheat or corn ? It is man-made, but you could argue that man is just tuning the blueprint created by nature. And it grows in the field, so it has the feel and touch of a 'natural' plant - unlike those plastic or wooden flowers that are used for room decorating.

But alright, let's say plants don't posess intelligence. Still, scientists were able to clone mice, cats and even horses. Now horses certainly do posess intelligence (understanding , learning etc.), albeit an inferior form of intelligence, if we compare it to the capabilities of human mind. If I interpret well your definition, clones are not artificial intelligence , but artificially created intelligence, by duplication of a natural blueprint for intelligence.


QUOTE (chickenwing)When intelligence can be created without requiring reproduction, it is artificial intelligence, at least by my definition.

Okay. I'll go with that.
I wager you imply no natural (sexual or asexual) nor artificial reproduction (cloning, stem cells etc.).


******************************************************************

I'll also agree with monsta666, in describing intelligence as the utmost degree of intelligence seen in nature - i.e. human intelligence

This concurs with the definition of AI by Dictionary.com :
Artificial Intelligence = "the capacity of a computer to perform operations analogous to learning and decision making in humans"


QUOTE (chickenwing)An interesting thought: "If the human brain was complex enough to comprehend itself, it would be too complex to comprehend"

Yes. That would be a problem
smile.gif

But monsta666 says it's not necessary to fully comprehend the human brain in order to identify an intelligent machine ('intelligent' as in 'human intelligence'). Suffice to perform the Turing test for that.

BUT you HAVE TO make those assumptions in order for the Turing test to be valid:

1) the assumption that all intelligence is human intelligence (is there no other form of intelligence ? If a human brain spoke from inside the body of a dolphin, would we recognize it as intelligent - or just stupid, because it does not act like a dolphin, but instead tries to freak everybody out by acting like a mutant ?)

2) that all that a human can say in a conversation is recognizable as intelligent (sometimes people don't understand each other's views. If you have me Turing-test a toaster and a racist person, I may pick the toaster as the intelligent being.)

3) (similar to 1) That it's not necessary to display intentionality or feelings to be recognized as intelligent

The example of the robot reading the newspaper and getting a job goes beyond the Turing test. The robot takes initiatives : it tries to live as a human.
But that may not be recognized as intelligent. What if the robot's was not design to travel over surfaces other than the kitchen floor ? The intention of leaving the kitchen would have been good if it was capable to do so. If it trips over the carpet and breaks down, the robot might seem unintelligent to its owner.
But the owner might also say "it had the best of intentions, let's not blame it".
Depends on his/her feelings towards the robot.

In brief : Since it's the human who decides who is intelligent and who's not, the Turing test only confirms the quirky judgment of an unreliable living creature.

*****************************************************************

Finally, if the human mind is embedded with the human body, which has come about through the process of evolution of species - a 3 billion year-long process - is it realistic to try to recreate something similar with silicon machines over a period of less than a century ?

From a philosophical point of view, among the AI approaches, I prefer by far sub-symbolic AI over symbolic AI.
 

chickenwing71x

-san
Kouhai
QUOTE (Konohamaru)Yes but what about genetically modified wheat or corn ? It is man-made, but you could argue that man is just tuning the blueprint created by nature. And it grows in the field, so it has the feel and touch of a 'natural' plant - unlike those plastic or wooden flowers that are used for room decorating.

But alright, let's say plants don't posess intelligence. Still, scientists were able to clone mice, cats and even horses. Now horses certainly do posess intelligence (understanding , learning etc.), albeit an inferior form of intelligence, if we compare it to the capabilities of human mind. If I interpret well your definition, clones are not artificial intelligence , but artificially created intelligence, by duplication of a natural blueprint for intelligence. Like I said before, with cloning and genetic modification you require original cells from an existing intelligent entity in order to "build" a new intelligent entity. When this can be accomplished independent of the original intelligent entity, it will be true artificial intelligence. Right now, it is only a form of reproduction - you are reproducing an existing intelligence. But like monsta said, we need to get off cloning.


QUOTE (Konohamaru)]BUT you HAVE TO make those assumptions in order for the Turing test to be valid:

1) the assumption that all intelligence is human intelligence (is there no other form of intelligence ? If a human brain spoke from inside the body of a dolphin, would we recognize it as intelligent - or just stupid, because it does not act like a dolphin, but instead tries to freak everybody out by acting like a mutant ?)

2) that all that a human can say in a conversation is recognizable as intelligent (sometimes people don't understand each other's views. If you have me Turing-test a toaster and a racist person, I may pick the toaster as the intelligent being.)

3) (similar to 1) That it's not necessary to display intentionality or feelings to be recognized as intelligent

The example of the robot reading the newspaper and getting a job goes beyond the Turing test. The robot takes initiatives : it tries to live as a human.
But that may not be recognized as intelligent. What if the robot's was not design to travel over surfaces other than the kitchen floor ? The intention of leaving the kitchen would have been good if it was capable to do so. If it trips over the carpet and breaks down, the robot might seem unintelligent to its owner.
But the owner might also say "it had the best of intentions, let's not blame it".
Depends on his/her feelings towards the robot.

In brief : Since it's the human who decides who is intelligent and who's not, the Turing test only confirms the quirky judgment of an unreliable living creature.
Agreed. We both seem to agree that the Turner Test is unreliable, though we brought up different reasons why.

Though I do disagree about the toaster. The turner test states that if the created entity can fool a human into believing it too is human, then it is classified as intelligent. A toaster cannot impersonate a human, but a racist human can. Are you saying that all the Americans who held slaves before the American Civil War were not intelligence beings? I know it was just an example though.


QUOTE Finally, if the human mind is embedded with the human body, which has come about through the process of evolution of species - a 3 billion year-long process - is it realistic to try to recreate something similar with silicon machines over a period of less than a century ? The rate at which evolution occurs is very dependent on random mutations which may or may not occur every generation. Now that humans have so far developed, we can act as a catalyst for other developing changes. Furthermore, the study and development of artificial intelligence has the goal of artificially creating intelligence from scratch, rather than speeding up the natural occurrence of it. Even more so, that does not change the feasibility of creating artificial intelligence in general. There is no time limit of one century, and humans are still advancing as we speak. Fifty years ago the average person had never even touched a computer, but look at how things are today.

On another subject, I want to bring up sentience and consciousness again. By definition, intelligence requires understanding. Is consciousness required for understanding?
 

Gustav1976

-sama
Retired
Chickenwing when you speak of conciousness are you refering to just the state of being conciouss or are you referring to self-awareness?
I just ask because many people confuse the 2 even though the 2 are linked in some ways. If you mean self-awareness I will say straight off that in my opinion the vast majority of homo sapiens aren;t self-aware in any true sense as most of us run on "automatic" a lot of the time.
As for intelligence it can reasonably said that one does not have to understand to be intelligent. eg. I, believe it or not, am a moderately intelligent member of my species yet I don't understand a great many things (such as why I feel an inextricable desire to kill certain people I wont name at times).
So, in summary intelligence does not require understanding although understanding often requires some level of intelligence.
 

trespasser

-san
Kouhai
This is kinda unique, as this was a major part of my philosophy class in university. Our section actually came to the topic of human consciousness and what exactly makes us 'human'. A large part of the arguement when it came to computers and the such came down to the idea of being of self and the view. As humans we see the way we see things, and our actions in the past have influenced us to precieve things in our own way. Because of this we are unable to view things from another persons eyes. I can precieve something through my eyes, and I can think to what it is to preceive the same thing through your eyes. However I cannot be able to preceive what you preceive, for it becomes far beyond any humans cognitive abilites. A good example is the following:

Suppose you want to know what it is like being a grasshopper. You think ok, I jump around I eat grass (it MUST be good, I mean they keep eating it...) and that is all. That is all fine and dandy, however that is what you think it is like to a grasshopper, but what does the grasshopper think it is to be a grasshopper? That is where we stop and cannot go any further.

Back to the target of the convorsation. From what I said above now try to actually design or create something that is consciousness, its own, not our perceptions. It is currently impossible, until we gain further 'enlightenment'.

Btw as a further follow up, look up Rene Descartes. His looks upon this stuff is amazing and mind blowing.
 

Zijian

-chi
Kouhai
Actually, modern day Artifical Intelligence has given on the problem of actually creating an Artificial Intelligence by replicating or mimicking the human mind.

Instead, the problem is focused on making Artificial Intelligence on solving specific problems like maze navigation, application to modern day appliances like washing machines and etc.

There is a full branch of science (neuroscience) dedicated to making a replicating the human brain and has got quite quantum leaps of progress.
 

EggBeast

- deska`
Retired
I think it was about a month ago when I watched a presentation by (I forget his name...) the guy who invented the first palm pilots. Now he's dedicating his work to creating software-based artificially intelligent machines. Apparently, it's really difficult to do, and right now their main emphasis is working on a visual-recognition system, to create a machine that can tell the difference between a picture of a cat and a picture of a dog (...among other things...) almost instantly. They've made some serious progress, although it still only works for relatively simplistic images.

I also saw another presentation by someone who is actually trying to create computer hardware that works more like the human mind. ...unfortunately, this guy had a horrible accent and was really bad at presenting, so I have no idea how that's going... but it's still interesting to see all the work that's going on in this field. It shows that it will probably be quite a long time before we get super-intelligent, super-hot, scantily-clad cyborgs, but they're definitely on their way.
tongue.gif


The main question that gets raised from all this is that even though we can make a machine that can act like a human, think like a human, talk like a human, maybe even emote like a human, could we really call it an intelligent being? To that I say... yeah. It wouldn't be human, but being human is no prerequisite for being intelligent. If a machine can learn things, make decisions on its own, if it can see, hear, smell, taste, and feel, and it can do all these things faster than a human, I'd have to call it intelligent. Sure, in the end it may be nothing more than a binary number cruncher, but the more we study the human brain, the more machine-like we realize it is. They're very different machines, yes, but that doesn't mean that on is innately superior to the other.
 

snorky2k

-san
Kouhai
My biggest concern is that as the artificial intelligence gets more powerful and closer to the capabilities of a human mind that it will be more difficult to tell when the machine is crazy and dangerous.
 

chickenwing71x

-san
Kouhai
QUOTE (snorky2k @ Jan 06 2009, 04:28 PM) My biggest concern is that as the artificial intelligence gets more powerful and closer to the capabilities of a human mind that it will be more difficult to tell when the machine is crazy and dangerous.
You aren't alone... it's such a common concept in literature, movies, television series, anime, and even video games, you name it.

I still don't think we're even close to that yet. Yes, we have made robots that are dangerous, and some are programmed to "kill" (Think robotic military weapons or self-guided missiles), but I wouldn't call them "evil," "crazy," or "dangerous." That would require consciousness, sentience, awareness, or something like it.
 

anbuu

-chi
Kouhai
i love you guys! not many people would carry an intelligent conversation about artificial intelligence with me anywhere else.

[Moderator's Note: even though you started the thread...do try to carry on the conversation/discussion after making this kind of remark....]
 

langes01x

-dono
Sempai
Even a simple program that chooses between a number of options and each moving having predictable repercussions like chess AI is quite complex. To have an AI that would react realistically to situations of an even more complex nature requires more information to be gathered, stored and processed in a short amount of time. This also requires more complex algorithms and code.

Sure computers can calculate 1032*827 faster than any person can but that is quite simple and has a standard formula that can be easily computed. Reacting to ones environment however requires a much larger and more complex formula as well as a more complex input.

From a programmers point of view we have a long way to go before we have realistic actions and reactions.

Though I say that they have text to speech engines already which would be needed and enough work has been done on AI such that they can play simple games with each other. If you haven't seen videos with robots playing soccer you should go check that out. True the ball is a specific color that does not blend into the environment but it is still a step in the right direction.

As to whether a robot could be dangerous of course they could. However it would be our fault for programming them in such a way that would allow for them to be dangerous. Robots can only do what they are made to do. Even if a robot can learn it can only learn about things that it is made to learn about. A chess playing AI can not learn how to speak even if it is made to learn and adapt because it was not made to speak. Thus it is the fault of the maker that it is dangerous and not the fault of the robot.

My belief is that sentience for robots is impossible or at least hard enough to accomplish that it will be a very long time for it to happen. Intelligence however is something that is possible and will come bit by bit as we become more intelligent ourselves.
 

monsta666

-the bee's knees
Staff member
Fansub TV Team
QUOTE (EggBeast @ Jan 06 2009, 07:33 PM)They're very different machines, yes, but that doesn't mean that one is innately superior to the other.
While it's true that humans and computers are not innately superior to one another each are superior in certain areas. Computers are better at handling large amounts of data and processing them quickly. They say the human mind is better at parallel processing and general recognition. Take that whole CAPTCHA code. It confuses the mighty computer but is a piece of cake for a human!
tongue.gif
But show that same computer some complex mathematical formula and they'll beat the humans hands down.

I guess one could example can be seen in the game of chess!
tongue.gif
The computer is much stronger in the opening and end-game phases while the human is better at the middle game. This is because the opening and end game rely much more on remember different move combinations and calculating all the possible move in advance. The middle game is more about interpreting the flow of the game. I guess it maybe one reason why there is less books on middle game strategy. Ahh but I haven't played chess in awhile!
tongue.gif



QUOTE (snorky2k @ Jan 06 2009, 08:28 PM)My biggest concern is that as the artificial intelligence gets more powerful and closer to the capabilities of a human mind that it will be more difficult to tell when the machine is crazy and dangerous.
Excluding programming bugs I doubt robots will be dangerous unless the designers built them that way. For a robot to turn against man it needs to develop ambitions for survival or feelings of superiority which are distinctly human qualities. I guess if man could create a robot that could emulate the human mind exactly this could be a problem but it's not a problem I see in the foreseeable future.
 

snorky2k

-san
Kouhai
Sorry, I did not mean to appear paranoid. Even twenty years ago, the computers on airbus airliners had a programming defect that was not discovered until a pilot attempted to do a touch and go at an airshow. The plane had a computer that reviewed pilot's actions and corrected for them. The plane's computers were confused at the move and tried to overide the pilot. When the plane needed more power to climb the plane still tried to land and put the plane into trees. It was all recorded on TV. There was no malicious intent. But, the system was complex enough that a major issue was missed.
 
Playasia - Play-Asia.com: Online Shopping for Digital Codes, Video Games, Toys, Music, Electronics & more
Top