Can artificial intelligences suffer from mental illness?

Discussion in 'Intelligence & Machines' started by Plazma Inferno!, Aug 2, 2016.

  1. cluelusshusbund + Public Dilemma + Valued Senior Member

    Messages:
    6,294
    I wont speek for Wegs... but i thank spiritual people like Wegs thank that AI pain is programed an biological pain is not programed... an what it boils down to is... a "soul" can only come from a higher power an can not be duplicated otherwise... so AI pain amounts to no more than an illusion of pain.!!!
     
  2. Google AdSense Guest Advertisement



    to hide all adverts.
  3. Write4U Valued Senior Member

    Messages:
    3,347
     
    Last edited: Dec 11, 2016
  4. Google AdSense Guest Advertisement



    to hide all adverts.
  5. Write4U Valued Senior Member

    Messages:
    3,347
    a belated p.s.
    If there is any kind of dynamical information sharing it can only be of a mathematical nature for it to be comprehensible and consistent to all.
    Isn't that what the Table of Elements is all about? It explains that fundamental physical particles can only be observed through mathematics of their atomic structure and their potentials which might become reality as long as they do not conflict with mathematical laws that are naturally imposed upon objects of certain mathematical values and potentials.
     
  6. Google AdSense Guest Advertisement



    to hide all adverts.
  7. wegs Matter & Pixie Dust Valued Senior Member

    Messages:
    3,343
    We'll have to agree to disagree, I suppose.
     
  8. billvon Valued Senior Member

    Messages:
    12,055
    But you could say the same thing about people.

    People come pre-programmed with the basic structure of their brain set beforehand. Then experiences/traumas/learning can change that. We might think that those changes are uncertain, but if you have a deep enough understanding of the underlying biology, you can say "that experience will change that person's brain by X" - and know exactly how they will react in the future. AI's are little different; they are just easier to understand because they are (usually) simpler.
     
    cluelusshusbund likes this.
  9. billvon Valued Senior Member

    Messages:
    12,055
    No.
    They were never "artificial." Their lives were aided through artificial means, just as yours is today (clothing, antibiotics etc.)
     
  10. iceaura Valued Senior Member

    Messages:
    22,332
    Define "basic structure". We know that the physical size and relative sizes and all manner of physical connections, as well as some influential details of the chemical composition, are plastic and malleable via common aspects of life and learning.
    Nobody has a deep enough understanding of the biology to know if that's even true.

    It's possible - easily possible - that you would have to know quite a bit about their history, and also the detailed particulars (including their own perceptions) about that future, and even then would be faced with a chaotic situation unpredictable in principle.
     
  11. Michael 345 Valued Senior Member

    Messages:
    1,402
    We don't 'think' the changes are uncertain, we know they are.

    We will never have a 'deep enough understanding of the underlying biology'.

    "that experience will change that person's brain by X" <<<<<< poppycock, unadulterated 100%. Not just the X, all of it.

    'know exactly how they will react in the future'. <<<<<<< poppycock piling up.

    Hard nosed Humpty
    Go easy Poe

    Please Register or Log in to view the hidden image!

     
  12. wegs Matter & Pixie Dust Valued Senior Member

    Messages:
    3,343
    Maybe it's the term ''suffer'' that I seem to only be able to attach to animals, humans or ''life'' in a general sense, but have a hard time seeing past that and applying it to machines, if that makes sense?
     
  13. billvon Valued Senior Member

    Messages:
    12,055
    What connects to where. Where the vision centers are; how the structure of the eye maps to portions of the visual cortex. Where the sensorium and motor control areas are, and how they are mapped.
    Exactly. And those same connections are just as malleable in self-training neural networks implemented on computers.
    Of course it's true; brains are deterministic. We don't have a deep enough understanding of the brain (yet) to be able to predict everything that will change within one, of course.
     
  14. billvon Valued Senior Member

    Messages:
    12,055
    We once said the same thing about the basic forces of nature.
    Glad to see that you, amongst everyone on this board, knows exactly how the future will turn out!
     
  15. billvon Valued Senior Member

    Messages:
    12,055
    That's understandable. We are hardwired to feel empathy for other humans - and to a lesser extent to other animals who are similar to us. We are less able to feel empathy for animals dissimilar to us (like octopi) even though they are quite intelligent. It's _very_ hard to feel empathy for a computer, since computer awareness is very far outside our experience.
     
  16. Michael 345 Valued Senior Member

    Messages:
    1,402
    And still don't have a deep understanding. We seem to be getting closer which I put down laws of physics being fixed and knowledge about fixed laws being built up.

    Raw biology obeys the fixed laws of physics but the gazillion variations of individuals operate allows individuals a gazillion ways to respond to events not open to prediction.

    ??? not sure how you conclude I have any predictive powers from my posts.

    Humpty
     
  17. iceaura Valued Senior Member

    Messages:
    22,332
    There are "deterministic" states that cannot be predicted (chaotic regimes), and the nature of what "determines" a brain is not a simple question - substrate does not determine pattern. To some extent a brain's low level behavior in certain respects may be most accurately "determined" by its higher level behavior - the larger patterns of thought over time and in feedback connection with the outer world having more say in the distribution of - say - electrical potentials and the like, than vice versa.

    Do you agree that a "deterministic" brain future that cannot be predicted without a complete whole-brain record of the first twentyfive years of REM sleep behavior is not what the determinist actually has in mind?
     
  18. Michael 345 Valued Senior Member

    Messages:
    1,402
    Makes good sense. I don't know of any machine which has a nervous system capable of experiencing pain.

    We might build one in the future (stand by for numerous moral discussions).

    Having built a nervous system capable of sensing pain and suffering we would need to inflict pain and suffering for the machine to learn

    Of course numerous other senses need to be added to the AI to provide contrasts.

    Dumpty
     
  19. billvon Valued Senior Member

    Messages:
    12,055
    Agreed with all the above - and all the above applies to AI as well.
    You'd need a lot more than the record of the first twenty-five years of REM sleep to do anything like that.
     
  20. billvon Valued Senior Member

    Messages:
    12,055
    If you understand the raw biology you understand the reaction. There's no supernatural "spirit" separate from our bodies that gives us intention.
    You have stated that learning how experience will change a person's brain in a specific way is "poppycock." That's quite the prediction - that we will never be able to do that. Not sure why you are so sure.
     
  21. Write4U Valued Senior Member

    Messages:
    3,347
    What causes a headache (pain) and what is a headache?
    Could this be a difference between chemical molecules and chemical bio-molecules? Sensitivity?
     
    Last edited: Dec 19, 2016
  22. Michael 345 Valued Senior Member

    Messages:
    1,402
    Understanding raw biology, which I take to be, we all have bodies which follow laws of physics, but overlaying that our individual temperament, experiences and the rest of the gazillion (I like the word) differences.

    I am not advocating any sort of 'spirit'.

    From my understanding of the brain and certain experiments done it appears decisions made by someone appear in the brain a fraction of a second before the person carries out the action.

    I think decision made in unconscious part of brain before being shipped up to conscious section is weird.

    I puzzle over who chats to the unconscious part, and....all the regressions down to bedrock.

    As to understanding how experience will change a brain in a specific way we do have a gross idea of changes which occur.

    Specific no.

    Again the gazillion possibilities come into play.

    Like the transporter beam in Star Trek, which has been estimated to require computing power impossible to achieve (I know I know never say impossible, look at what was thought impossible a few years ago).

    I would say it would require a computer of comparable size size to compute specific changes.

    And let's not forget as soon as the person is released from the machine all that data becomes obsolete.

    That's Humpty Dumpty prediction along with Poe's

    Please Register or Log in to view the hidden image!

     
  23. billvon Valued Senior Member

    Messages:
    12,055
    Well, not so much "overlaying" as "made up of." The gazillion differences come from the gazillion differences in anatomy and potentials.
    Yes; that's mainly propagation delay. (Our brains, while amazingly parallel and able to do many things at once, isn't all that fast when it comes to turning an intention into action.)
    It may also be that our perception of the decision is skewed. We "filter out" a lot of the things we perceive (like saccades) and we alter other things (like the perception of sensation coming from our feet.) We may well be just perceiving that the decision was made far after it really was.
    There are more fundamental problems with it than computing power.
    However, there may come a point where all that is required to achieve something like that is an improvement in computing. That would be great, and would mean that it's now barely possible. Because of all the things that man has accomplished, exponential improvements in computing power has been one of the more reliable achievements.
    I'd guess that the size of the computation would depend primarily on the fidelity you wanted.
    Want to simulate it to 100% accuracy? Might take a computer the size of the Earth.
    Want to simulate it to 99.9% accuracy? Might require something that can fit on your desk; that's a order of magnitudes easier problem to solve.
     

Share This Page