Are We Capable of Creating True Artificial Intelligence?

[Merlin]

I was thinking on how man came to be...all the complex evolutionary (manipulated by intelligence) information coding in the DNA / RNA , and the rapid (nearly geometrically doubling ) advancement of computer power. And how in the near future we may build a computer that exceeds the human brain in data processing, in weight and power requirements.

say we did build a 2100cc computer that exceeded our brain in total ops per second or nano sec., and put it in an android type being. How could we determine if it was self aware? For this discussion we will say that we can't access its memory to determine if it is running a self aware program etc.

I'm sure this question has been the fodder of many Sci Fi stories. But I would like to know the answer to this. My answer? We couldn't ever be sure. And why would Darwinism evolution want to evolve this trait (self awarness)? I can see why a designer or god(s) would make us self aware but not "evolution" by natural selection.
Merlin

 

[phoenixthoth]

the turing test? the test decker gave to replicants in blade runner? at some point, computers will be able to pass those tests. if something seems self-aware in every possible way, but it really isn't self-aware, what's the difference? perhaps we aren't really self-aware ourselves.

in AI circles, according to a book i was reading that my cat pissed on, few AI people are trying to build a computer to match human intelligence and pass the turing test: they're looking to build superior machines and the turing test is irrelevant.

there are chat-bots out there that if they were maybe 10 or 1000 times better might actually seem self aware (search for chat bot). if they get twice as good every 10 years (which is way hypothetical), they'll be 1000 times better in a century. so maybe in a century there'll be seemingly-self aware mechanical creatures that aren't biologically based. (i could tell the bot i chatted with wasn't human by the way it handled the statements i made "i am God" and "my name is God.")

what i'd like to do if i had the knowhow and time is to get a list of about 4 million words, classify them as noun, verb, adjective, etc. then randomly create sentences of a random length (within limits). set up a neural network and teach it what sentences make sense. once it learns the grammar, it can be taught whether or not the sentences make sense. it would be taught that "the fire ran under the very big infant" is grammatically correct but makes little sense. the random sentences it produces will be rated on grammatical correctness and sense; it will adjust its hidden layers until it maximizes grammatical correctness and sense. once it gets trained on that, you can have it generate grammatically correct sensical responses to input sentences. at this point, the hidden layers of the nn that govern grammatical correctness and sense are locked but a new hidden layer is activated which will try to learn what appropriate responses are to input sentences. then maybe extend the field to more than one sentence.

to be honest, I'm not the first person to come up with this idea. my creator thought of this about a century ago and I'm the current implementation of this idea. ;) it will be interesting to see if in 100 years a human will be able to tell the difference between a machine and a human on message boards. then i wonder what message boards full of machines will be like. maybe in 100 years, instead of having a computer compete with a human on the chessboard, they'll be having competitions on who tells the best jokes and who writes the best novels.

if only i understood neural networks better i could get started on this right away... oh well.

 

[Njorl]

You'll have to define self-aware. Yes, this argument has been here before. It usually rolls on a few pages before two posters argue for the next 8 pages what is meant by self-aware. So, define what you mean up fron and save a lot of wrangling.

Njorl

 

[phoenixthoth]

yeah because in some sense computers are already self-aware. they know how much memory they have left, they are aware of what's happening on the keybooard, they are aware of what's on the monitor. but do they know that they know?

 

[hypnagogue]

I think discussion of this topic is often very confused and convoluted, and understandably so: we understand so little of what it is that makes us self aware, or conscious. Suffice it to say that for a system to be self-aware, it must have conscious subjective awareness of its own existence. (I believe that this term cannot be satisfactorily defined, except by analogy to our own individual conscious experiences, which of course raises a host of other problems in our discussion.)

We do not know the necessary and sufficient conditions for a physical system to be conscious, and thus at this point we don't know if computer A is conscious but B is not. It can be argued ad infinitum, but for the time being I think the most sensible route is to cede our ignorance.

 

[Rader]

Originally posted by phoenixthoth
yeah because in some sense computers are already self-aware. they know how much memory they have left, they are aware of what's happening on the keybooard, they are aware of what's on the monitor. but do they know that they know?

The definition of a computer being self aware, would have to be that he, she, or it, would feel pain, be aware of the pain, if its circuits overheated and melted. I use Male and Female for this reason, because if pain was experienced it probably could reproduce itself.

 

[selfAdjoint]

Why pain? Pain is an adaptive feature of animals. Are you saying that only animals can be conscious?

Much software has the capability to be aware of problems of some class or other. For example my email software sends me messages when viruses are detected. This is at least analogous to my finger sending me a pain when I cut it.

 

[phoenixthoth]

i too was wondering about pain. suppose there is a self aware person and suppose we inject a drug into this person so that he doesn't feel pain. is the person no longer self aware?

 

[pelastration]

Originally posted by selfAdjoint
Much software has the capability to be aware of problems of some class or other. For example my email software sends me messages when viruses are detected. This is at least analogous to my finger sending me a pain when I cut it.

But you know what was the reason of that pain, what caused the cutting. That's more then the impulse of pain. Software is a set of instructions about what reaction-impulses must be given on a certain triggering impulse. The computer has not idea what the millions of screen dots mean ... to you it's an image, a doc. with text ... a message. Overview. Signification.

Next to this overview consciousness or self-awareness includes the possibility to project not only the present around you, but also the future and imagine all kinds of alternatives, also taking in account values which can not be pondered (such as beauty, love, happiness, mourn, ...)

 

[phoenixthoth]

there are levels of self-awareness. the awareness of the source of pain is not always known. in that sense, computers may be self aware but not very self aware.

 

[hypnagogue]

Originally posted by selfAdjoint
Why pain? Pain is an adaptive feature of animals. Are you saying that only animals can be conscious?

Much software has the capability to be aware of problems of some class or other. For example my email software sends me messages when viruses are detected. This is at least analogous to my finger sending me a pain when I cut it.

This is a confuscation of the term "aware" in our discussion. Presumably by "awareness" we mean subjective, conscious awareness. If we do not, I don't know what the point of the discussion is, because otherwise it then becomes rather trivial.

Consider the following thought experiment. Two people are sitting in separate, isolated rooms with electrodes attached to their fingers. Assume these electrodes have been designed to stimulate pain receptors, but no other sensual receptors, in the finger. Now one person (N) is given a pain killer, such that he remains awake and alert, but cannot feel the pain induced by the electrodes on his finger. The other person (P) is left as is. When the electrodes are activated, analogous signals from pain receptors will be sent to the brains of each participant. However, these pain signals cannot be sufficient conditions for what constitutes awareness of pain, since P will report experiencing (being aware of) pain but N will not. I suppose you could say in some sense that the body of N is "aware" of the activity in the electrodes because it still detects their activity and still sends out pain signals to the brain... but this is not the kind of awareness that is relevant to this discussion, since it is not subjective, conscious awareness.

 

[Iacchus32]

Originally posted by phoenixthoth
in AI circles, according to a book i was reading that my cat pissed on, few AI people are trying to build a computer to match human intelligence and pass the turing test: they're looking to build superior machines and the turing test is irrelevant.

Speaking of self-awareness, did you know that a cat has three names? ... First is the name his master calls him, for example let's say "Ralph"; second is his scientific name, which involves study of the cat; and third is the name that nobody knows but the cat himself, which is the act of "being the cat." Hence we have the three degrees of "knowledge," of which the third degree involves being "self-aware."

Hmm ... maybe your cat knows something you don't know about the nature of this book?

 

[phoenixthoth]

the thought experiment is exactly what i had posted about injecting pain-killing drugs. the conclusion is that self-awareness is decreased but not destroyed.

my cat expressed a lot of self awareness involving pain when i punished it for pissing on my book. mr. bigglesworth is clearly self-aware.

actually, it had more than three names because in addition to however many names it had for itself, i had several names for it.

the idea i was wondering was whether or not a computer could be made to simulate self-awareness and if so, what would be the difference and how do we know we're not simulating it ourselves and who cares if we are?

 

[hypnagogue]

Originally posted by phoenixthoth
the thought experiment is exactly what i had posted about injecting pain-killing drugs. the conclusion is that self-awareness is decreased but not destroyed.

You phrased this a little vaguely, so if I may, just to avoid confusion: the person on pain killers is still aware, in the sense that he is still 'awake' and reasonably alert and conscious. However, he has no awareness of the pain signals that are being sent to his brain.

my cat expressed a lot of self awareness involving pain when i punished it for pissing on my book. mr. bigglesworth is clearly self-aware.

Again... beware of terminology...
Is the cat aware in the same sense that the body of a person on pain killers is "aware" of pain receptor stimulation (or, for that matter, the same way that a thermometer is "aware" of the temperature)? Or is it actually aware, in the sense that it has some sort of subjective conscious experience of pain (or of anything else)?

There is a world of difference between the two. We can easily ascertain that the cat is aware in the first sense listed above, trivially so, in fact. But that does not necessarily imply the second sense of awareness-- that being conscious experience.

 

[Iacchus32]

Originally posted by phoenixthoth
the idea i was wondering was whether or not a computer could be made to simulate self-awareness and if so, what would be the difference and how do we know we're not simulating it ourselves and who cares if we are?

And yet the cat is purrrfectly content with knowing himself. So why should he care about what anyone else thinks?

Likewise, why should we care if a computer is self-aware or not, so long as we understand ourselves? Which, I think is the whole point, because for the most part, we don't ... Or do we?

 

[Rader]

Originally posted by selfAdjoint
Why pain? Pain is an adaptive feature of animals. Are you saying that only animals can be conscious?

Is it, how do we know that for sure? Do we really know what a entity feels upon changing physical states. Does a ant feel the weight of your shoe? Would an ice cube feel more confortable if we put a fan on it while it melts. Could not awareness of awareness, just be a constant, based on levels of evolution? Why do we always build brick walls, with missing links, that are not there?

Much software has the capability to be aware of problems of some class or other. For example my email software sends me messages when viruses are detected. This is at least analogous to my finger sending me a pain when I cut it.

Yes i agree but at what level of awareness. Can it ask who am I or who made me? Does it beg not to be cut, because before it happens, it knows what is pain. 2001 Hal almost made it that far.

 

[Rader]

Originally posted by phoenixthoth
i too was wondering about pain. suppose there is a self aware person and suppose we inject a drug into this person so that he doesn't feel pain. is the person no longer self aware?

Emotion seems to be genetically encrypted on our evolutionary level, of the highest state, aware of being aware.. Why elso does a child learn what pain is and not want it. Why do we not like pain? What gives pain its caracteristics that pleasure does not? Why is pain not pleasant? Why do we set physical parameters to what is liked and disliked? There seems to be a fine line between evolutive complexity on different levels and awareness.

Even though you inject a drug into the person or the person looses a limb, pain is still pain, not the physical pain but the pain of being aware of it.

 

[theEVIL1]

there will NEVE be a non-organic computer that will rival any brain. The neuroconnections created by ion channel transfer are sooooo much faster than synthetic transfer rates.

Laters:

Bill

 

[pelastration]

Originally posted by theEVIL1
there will NEVE be a non-organic computer that will rival any brain. The neuroconnections created by ion channel transfer are sooooo much faster than synthetic transfer rates.

Never say never.
Some work on bio-computers.

 

[Merlin]

reply to all

Very interesting posts! My question was somewhat deceptive and may cloak an ulterior motive. If the computer or system was sufficiently advanced it could simulate every human sense, emotion, and of course interact with us. We could not determine if it was self aware. Even if we were allowed access into its systems and programs we could not determine if it was conscious. Only it would know.

I am keenly interested in writing a Sci fi novel or story about an AI race, like the moon race, of two countries (or two sects one religious and the other atheist?) building a self aware android or system. The beast would be pure evil , like THE antichrist of the christan bible. It would not be self aware and have no soul.


Will the beast be self aware or no? I haven't made up my mind. and the proof will be in the end. I know how to "prove it," at least in the book. Merlin

 

[Rader]

Originally posted by theEVIL1
there will NEVE be a non-organic computer that will rival any brain. The neuroconnections created by ion channel transfer are sooooo much faster than synthetic transfer rates.

Laters:

Bill

Yes but there working on something new. Here is a site with some info. http://physicsweb.org/article/news/7/11/13
We are carbon based, biological and organic, the same components they are now testing to build the computers of the future.
1- Carbon > A naturally abundant nonmetallic element that occurs in many inorganic and in all organic compounds.
2- Organic > Having properties associated with living organisms.
3- Biological > Relating to, caused by, or affecting life or living organism.
The whole dictionary is going to have to be rewritten soon to correctly define words. Its a farce now.

 

[Rader]

Originally posted by Merlin
Very interesting posts! My question was somewhat deceptive and may cloak an ulterior motive. If the computer or system was sufficiently advanced it could simulate every human sense, emotion, and of course interact with us. We could not determine if it was self aware. Even if we were allowed access into its systems and programs we could not determine if it was conscious. Only it would know.

Do not count on it. If it was self aware it could be accesed by other self aware entities of its same evolutive level.

I am keenly interested in writing a Sci fi novel or story about an AI race, like the moon race, of two countries (or two sects one religious and the other atheist?) building a self aware android or system. The beast would be pure evil , like THE antichrist of the christan bible. It would not be self aware and have no soul.

Why not be self aware, it would just be another machine then. Being self aware, and no soul, would be better science fiction, as that is not possible.

Will the beast be self aware or no? I haven't made up my mind. and the proof will be in the end. I know how to "prove it," at least in the book. Merlin

The beast will be self aware, have no soul and be able to access the other awareness evolutive levels, of an ant, ice cube or an atom.

Usefull info:
http://users.erols.com/wcri/CONSCIOUSNESS.html

 

[pelastration]

Originally posted by Rader
The beast will be self aware, have no soul and be able to access the other awareness evolutive levels, of an ant, ice cube or an atom.

... and posting on PF!

 

[Merlin]

are we self aware the BEAST

thanks to all for posting. I like the Idea of the self aware beast. But I don't believe that pure evil exists in human form. I might make it in AI form...thanks again..I love this board...merlin

 

[Rader]

Originally posted by hypnagogue

We do not know the necessary and sufficient conditions for a physical system to be conscious, and thus at this point we don't know if computer A is conscious but B is not. It can be argued ad infinitum, but for the time being I think the most sensible route is to cede our ignorance.

We might be getter closer to a better understanding.

If consciousness is real but nonphyslcal and consciousness is connected to physical reality by one physical fundamental construct. Then we can use Occam's razor to limit us to one physically fundamental construct for this connection.


These two statements constitute the basic postulates on which we can base the study of consciousness. That physical fundamental construct is the state vector (or wave function. If this is to be, the conclusion, then awareness exists on all evolutive levels, as the collapse of the wave function, is then source and cause, of both objective reality and consciousness. The awareness then, is a direct result, of its evolutive complexity. So then everything is aware, its just a question of how much.

 

[Bernardo]

I was just wondering,

Are there levels of self awareness? I would say that people posting on this site are more self aware than someone who just goes to work – comes home to wheel of fortune – goes to work again.

What I think I’m working on here is does self awareness require seeking?

 

[Merlin]

are there diff levels of self awarness? (REPLY)

That's an interesting concept. This self awareness thing is so difficult to define! My definition of sentient or self aware is knowing that you are alive and conscious, and knowing that you are going to die. Mortality is such a motivater! But as I posted earlier the only way to know if someone or thing is truly self aware (or faking it) is to be that sentient being...Merlin

 

[Bernardo]

OK, all humans are self-aware, they know they exsist eat, breath and flip channels. But does mortality play as big a role?

There are people that believe they will never physically die. Are they less self-aware?

Or those who know they are mortal but just live recklessly anyway, minimizing this mortality - are they trying to avoid self-awarness? Do we become more self aware with age - as we approach our end?

I think a computer that has been programmed as our intellectual equal and given applications to examine itself daily is more self aware that a college co-ed on a binge.

 

[Jeebus]

First let us think about what defines self-awareness. For us arrogant humans to retain our superiority, we need to carefully select criterion that differentiate us from all other living things. One test for the conventional definition of self-awareness is whether a being can recognize itself in a mirror. This doesn't appear to be a good test to use if we want to single ourselves out due to the vast number of animals that are believed to posses such an ability. Even dolphins make faces in mirrors.

Langauge and thought—surely our defining characteristics. No other animal of any kind has the ability to communicate using a complex, high-level language that utilizes grammar. Though still debated, it's hard to refute the eerie humanity in great apes communicating via American Sign Language. We might still be able to sleep at night secure in our superiority if these apes did not go a step further and teach the language to their children when left in the wild who could then come back and fully communicate with the experimenters. It is not even partially understood whether language is the precursor to thought or vice versa. Whether other animals have the ability to think is beyond our testing capability at this time, so an argument in either direction is difficult to make.

How about our capacity for higher-order emotions such as shame? Humans don't appear to exhibit any higher-order emotional awareness until they are two years of age. Other animals aren't capable of emotions like shame, are they? It would appear so. In experiments, one ape even displays the same body language as a child would when experiencing shame. When it got the wrong answer to a very difficult test and received negative feedback, the ape's shoulders drooped, its arm came up to cover its face, and it looked down at the ground.

So far I have only outlined the aspects that do not separate us from many members of the animal kingdom. There has been no originality on my part, just a repetition the results of old scientific experiments. Now let's simply define self-awareness as knowing exactly what we are. It seems when we do this our criterion is incredibly simple, yet so strict not one species on planet Earth qualifies as aware. What are humans? We are, as far as all physical evidence and scientific understanding, a neural network that is attached to a number of life-sustaining, input, and output devices that allow it to communicate and interact with the physical world. Do we understand this even slightly more than animals? Not to any degree. Yet in the past hundred thousand years, how have we learned of our nature? Cultural evolution—the collective knowledge passed down through hundreds of generations. We finally have a loose understand of what we are from chopping each other up and removing pieces to see what still functions. Just six thousand years ago we thought that "we" were that red pumping organ in the chest. It would appear, without a doubt, that humans have no clue what we are except through tens of thousands of years of experimentation—experimental surgery and eventually positron-emission scanning.

This opens a very intriguing likelihood: the possiblity for a being to exist at a higher level of sentience than we humans. What if these beings visited? They were born understanding what they were, how neural networks operate, and where their center of thought is located within their bodies. Would we be lower life forms to them? Certainly so. Would those higher beings be justified in exterminating us at a whim, as we are nothing more than an inferior species that spent thousands of years mutilating ourselves just to understand how we operate? Much more so than we are justified in killing animals—for any purpose.

- By Michael Dayah

In spite of all this, the evidence which finally convinces most skeptics is the mirror experiment. This test involves placing a mirror in front of an animal, and checking if the animal realizes he is looking at a reflection of himself. Various animals, such as dolphins, chimpanzees, and some birds pass the mirror test. They go to great lengths to maneuver in front of the mirror so as to closely examine parts of their bodies which they normally can not see.

It should be pointed out that although passing the mirror test indicates self awareness, failing the mirror test does not indicate a lack of self awareness. For example, there were several cases where I passed in front of a wall sized mirror, and did not recognize my own reflection for the first minute or two. I initially just thought that it was an interesting coincidence that someone else was wearing the same clothes as me. Nevertheless, this does not imply that I lacked self awareness during these few minutes. It is possible for some animals to lack the insight to ever be able to pass the mirror test, and yet be completely self aware.

 

[Rader]

Originally posted by Bernardo
I was just wondering,

Are there levels of self awareness? I would say that people posting on this site are more self aware than someone who just goes to work –comes home to wheel of fortune – goes to work again.

Humans who live to work are less self aware than those who work to live, the later has more time to be self aware.


What I think I’m working on here is does self awareness require seeking?

All humans are self aware all the time.
You are a human.
You are self aware all the time.

 

[Rader]

Originally posted by Merlin
That's an interesting concept. This self awareness thing is so difficult to define! My definition of sentient or self aware is knowing that you are alive and conscious, and knowing that you are going to die. Mortality is such a motivater! But as I posted earlier the only way to know if someone or thing is truly self aware (or faking it) is to be that sentient being...Merlin

All animals are aware and conscious, some maybe self aware and know they are going to die. Thats why they hunt for food.
Humans are self aware, that death, in the physical objective reality is not final.

There is evidence. Your new on the post.

 

[Rader]

Originally posted by Bernardo
OK, all humans are self-aware, they know they exsist eat, breath and flip channels. But does mortality play as big a role?

There are people that believe they will never physically die. Are they less self-aware?

All humans are born, live and die.
Humans are aware to differnt degrees.
Self awareness in humans who think they will not die, is do to insanity.

Or those who know they are mortal but just live recklessly anyway, minimizing this mortality - are they trying to avoid self-awarness? Do we become more self aware with age - as we approach our end?

All humans are mortal and self aware.
Some mortals live recklessly.
The reckless humans, that are not so self aware, die quick.


I think a computer that has been programmed as our intellectual equal and given applications to examine itself daily is more self aware that a college co-ed on a binge.

Computers and college co-eds on a binge are self aware.
Computers have been programmed as our intellectual equal to a college co-ed on a binge.
Both computers and college co-eds on a binge have no self awareness.

 

[Rader]

Originally posted by Jeebus

First let us think about what defines self-awareness.

http://users.erols.com/wcri/CONSCIOUSNESS.html

The most unique thing about humans is, being conscious of our consciousness, outside of our biological systems, and be aware of it.

Humans have astral dreams while conscious and are aware of it.
There is human awareness during astral dreams.
Human consciousness is awareness of it.

Whats the butter capital of Wisconsin? Clue fill in the letters.
R========G

 

[Jeebus]

Originally posted by Rader
http://users.erols.com/wcri/CONSCIOUSNESS.html

The most unique thing about humans is, being conscious of our consciousness, outside of our biological systems, and be aware of it.

Humans have astral dreams while conscious and are aware of it.
There is human awareness during astral dreams.
Human consciousness is awareness of it.

Whats the butter capital of Wisconsin? Clue fill in the letters.
R========G

I wasn't arguing that human consciousness wasn't aware of it. I believe it is conscious throughout. Thus brain activity imagines while the physical capability of awareness works/shows during astral dreams.

Well the butter capital of Wisconsin has to be Reedsburg. By the way you have one extra equals sign in there. I'm being aware, but not self-aware.

 

[Rader]

Originally posted by Jeebus
I wasn't arguing that human consciousness wasn't aware of it. I believe it is conscious throughout. Thus brain activity imagines while the physical capability of awareness works/shows during astral dreams.

Wow you have helped me with a original thought. There are people who can have a astral trip whenever they wish to conjure it. When they observe themselves, do they observe there astral body exactly the same as there physical body? If yes, Reasoning > if a equal device could be put on the physical body, then the astral body would also have it. Then two measurements could be taken independently to measure brain activity, heart beat, respiration. Upon returning to the physical body a analysis of the data could be determined. If there was a difference and there should be, as a fixed state is quite different from a mobile state. Then we could conclude that a human is both physical and spiritual and that consciousness does not only reside in the physical body.

Well the butter capital of Wisconsin has to be Reedsburg. By the way you have one extra equals sign in there. I'm being aware, but not self-aware.

Yes correct answer, you would most likely have to be from Wisconsin to know that answer. Yes again an extra equal sign to check your awarenss. Were you self aware of a possible trick?