Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Are you self aware?

  1. Nov 26, 2003 #1
    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 cant 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.
  2. jcsd
  3. Nov 26, 2003 #2
    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.
    Last edited: Nov 26, 2003
  4. Nov 26, 2003 #3


    User Avatar
    Science Advisor

    You'll have to define self-aware. Yes, this arguement 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.

  5. Nov 26, 2003 #4
    yeah cuz 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?
  6. Nov 26, 2003 #5


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    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.
  7. Nov 26, 2003 #6
    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.
    Last edited: Nov 26, 2003
  8. Nov 26, 2003 #7


    User Avatar
    Staff Emeritus
    Gold Member
    Dearly Missed

    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.
  9. Nov 26, 2003 #8
    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?
  10. Nov 26, 2003 #9
    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, ...)
  11. Nov 26, 2003 #10
    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.
  12. Nov 26, 2003 #11


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    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.
  13. Nov 26, 2003 #12
    Speaking of self-awareness, did you know that a cat has three names? ... First is the name his master calls him, for example lets 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? :wink:
  14. Nov 26, 2003 #13
    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?
  15. Nov 26, 2003 #14


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    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.

    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.
  16. Nov 26, 2003 #15
    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?
    Last edited: Nov 27, 2003
  17. Nov 27, 2003 #16
    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.
    Last edited: Nov 27, 2003
  18. Nov 27, 2003 #17
    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.
    Last edited: Nov 27, 2003
  19. Nov 27, 2003 #18
    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.


  20. Nov 27, 2003 #19
    Never say never.
    Some work on bio-computers.
  21. Nov 27, 2003 #20
    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.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook