Page 8 of 8 FirstFirst ... 678
Results 71 to 76 of 76

Thread: Artificial intelligence would robots, androids and cyborgs have civil rights?

  1. Top | #71
    Contributor DrZoidberg's Avatar
    Join Date
    Nov 2007
    Location
    Copenhagen
    Posts
    7,686
    Archived
    5,746
    Total Posts
    13,432
    Rep Power
    51
    Quote Originally Posted by fromderinside View Post
    Quote Originally Posted by DrZoidberg View Post


    I suggest reading up on the Chinese Room thought experiment by Searle.

    https://en.wikipedia.org/wiki/Chinese_room
    Good suggestion. Seems an AI guy and I were put to the test on a very similar problem back in 1988. Our manager asked us to devise arguments for list based or AI based solution for handling degrading systems in commercial A/C flight as aids for pilots. Both of us were familiar with graceful degradation solutions extant. The idea is that a simple error in systems operation leads to increasingly more serious errors as time passes and attempts to normalize operations continues. One error corrected leads to other existing but non-operative errors to become relevant in an ever increasing chain of events ultimately leading to total system failure and plane destruction.

    Originally this was the problem confronted by IBM with their system 370 which was a very large bit of software, by measures at that time, which had passed testing leaving many issues undetected. Those errors weren't relevant untiol fixes were made nor accounting for them which brought them forward leading to more fixes. Obviously this leads to the gradual degradation of an OS which lead to the development of structured design and exhaustive testing as remedies.

    The child was worse than the parent. Now large became the issue. There is a limit to which humans in organizations can manage complexity exceeds the capacity of the organization to manage groups and individuals.

    When I read the evolution of the chines room problem I had a de ja vous moment. The Chines language is too complex for one to provide solutions satisfying every condition one might encounter in communicating. The problem in a non-problem.

    Going back to our cascading degeneration of operability is is best to keep the problem within the scope of those designated to handle it. So a simple hierarchy of lists acting as the current do succeeds where the design of an intelligent system requires understanding of the operators which soon exceeds the ability of designers to develop solutions to the next cascading problem. That is the AI method is too complex for resolving the problem in less than infinite time ow with infinite money. The list guy won the day.
    ...if you are programming the AI using decision trees (with LISP). Which is how they thought the human brain worked from the 1950'ies to 1990'ies and we just couldn't figure out what we were doing wrong. Until we learned that it wasn't at all how the human brain worked, and that worked was all scrapped.

    Quote Originally Posted by fromderinside View Post
    So if I had to respond to Searle's problem I'd suggest an intuitive tool like a list of lists which can be searched for terminations by competent crew.
    The Chinese room is about consciousness. Does the guy in the Chinese room know how to speak Chinese? He's the engine of room, so he's the one talking. But he's just passing notes (he can't understand) back and forth. Translated to what we're talking about. When a stack of English notes is slipped under the door (your brain receives stimuli) you look into your catalogue and stick the corresponding Chinese word notes under the door back (you feel pleasure/dopamine release).

    Or to make it more gruesome. The difference between an elevator and you is that you have a cat in a box constantly being tortured and when you reach the desired floor we stop torturing the cat temporarily. But soon the torture resumes. Which is the mechanism with which life "encourages" us to explore the world.

    Searle has just removed the tortured cat from the room. If we're torturing the cat in the room then Searle thinks we can truly speak Chinese. Because nailing it makes us happy. That's my interpretation of where Searle goes wrong.

    Quote Originally Posted by fromderinside View Post
    When you spoke of the sea squirt you included something with a nervous system along a chord parallel to the digestive tract and the means of motion available to the squirt. That's a damn sight more intelligent than broccoli. By intelligent I mean more problems were solved by the design, like changing locations of nutrients, getting to them, processing them etc.
    Bah.. plants can photosynthesise. They can kill entire limbs if attacked. They communicate with each other. They can turn toxin production on and off. Plants are intelligent in their own way. I don't see how you can compare the types of intelligences between these two beings. Both have the "brains" they need to excel in their environment.

    Don't get me wrong. I think humans are intelligent. The most intelligent creature on this planet. Due to the flexibility of it and our ability to effortlessly switch between the symbolic and real (allowing us to plan ahead and fantasise). It's not perfect. Which is why so many believe in God. Anyhoo... it's a really cool system. But there's still no magic involved. If nature programmed us to have feelings, than we can certainly programme computers to have feelings.

    Like I said earlier. Humans are programmed using chemical engines with proteins are their base. Proteins are extremely tightly wound and efficient units for storing information in many different states. Because it's chemical putting it into stable states is simple. Computers today are programmed using silicone chips with simple circuits. They require constant power to be in stable states, or rely on inefficient relays. Computers send signals much much faster than what's going on in a human brain. The human brain communicates at a glacial pace by comparison. The human brains signalling system sends much less information at a go.

    Bottom line, the basic architecture for a human brain and a computer is probably too different for us to be able to manage to copy it in a computer. The only animal's brain that we've managed to simulate accurately in a computer is the tiny nematode "C Elegans". Perhaps that's the best we can manage with our current computing power. We tried to build a mouse brain, but we didn't have the hardware necessary. Aka "Blue mouse project"

  2. Top | #72
    Mazzie Daius fromderinside's Avatar
    Join Date
    Oct 2008
    Location
    Oregon's westernmost
    Posts
    10,551
    Archived
    18,213
    Total Posts
    28,764
    Rep Power
    52
    Quote Originally Posted by DrZoidberg View Post

    ...if you are programming the AI using decision trees (with LISP). Which is how they thought the human brain worked from the 1950'ies to 1990'ies and we just couldn't figure out what we were doing wrong. Until we learned that it wasn't at all how the human brain worked, and that worked was all scrapped.
    Good point except my contention is not that Searle took the wrong tack - he did - he presumed the solution was soluble by technical input - it isn't - so his contention is one of question begging as you are so fondly and frequently point out are the arguments of others.

    Quote Originally Posted by fromderinside View Post
    So if I had to respond to Searle's problem I'd suggest an intuitive tool like a list of lists which can be searched for terminations by competent crew.
    Quote Originally Posted by DrZoidberg View Post
    The Chinese room is about consciousness. Does the guy in the Chinese room know how to speak Chinese? He's the engine of room, so he's the one talking. But he's just passing notes (he can't understand) back and forth. Translated to what we're talking about. When a stack of English notes is slipped under the door (your brain receives stimuli) you look into your catalogue and stick the corresponding Chinese word notes under the door back (you feel pleasure/dopamine release).

    Or to make it more gruesome. The difference between an elevator and you is that you have a cat in a box constantly being tortured and when you reach the desired floor we stop torturing the cat temporarily. But soon the torture resumes. Which is the mechanism with which life "encourages" us to explore the world.

    Searle has just removed the tortured cat from the room. If we're torturing the cat in the room then Searle thinks we can truly speak Chinese. Because nailing it makes us happy. That's my interpretation of where Searle goes wrong.
    He not only removed the tortured cat from the room he presumed sequences of exchanges which could be predicted. He can't. His problem is one which reaching an infinity of possibilities very quickly. IOW referring to my models one can't gracefully degrade if one doesn't know the outcomes a priori.

    Quote Originally Posted by fromderinside View Post
    When you spoke of the sea squirt you included something with a nervous system along a chord parallel to the digestive tract and the means of motion available to the squirt. That's a damn sight more intelligent than broccoli. By intelligent I mean more problems were solved by the design, like changing locations of nutrients, getting to them, processing them etc.
    Quote Originally Posted by DrZoidberg View Post
    Bah.. plants can photosynthesise. They can kill entire limbs if attacked. They communicate with each other. They can turn toxin production on and off. Plants are intelligent in their own way. I don't see how you can compare the types of intelligences between these two beings. Both have the "brains" they need to excel in their environment.

    Don't get me wrong. I think humans are intelligent. The most intelligent creature on this planet. Due to the flexibility of it and our ability to effortlessly switch between the symbolic and real (allowing us to plan ahead and fantasise). It's not perfect. Which is why so many believe in God. Anyhoo... it's a really cool system. But there's still no magic involved. If nature programmed us to have feelings, than we can certainly programme computers to have feelings.

    Like I said earlier. Humans are programmed using chemical engines with proteins are their base. Proteins are extremely tightly wound and efficient units for storing information in many different states. Because it's chemical putting it into stable states is simple. Computers today are programmed using silicone chips with simple circuits. They require constant power to be in stable states, or rely on inefficient relays. Computers send signals much much faster than what's going on in a human brain. The human brain communicates at a glacial pace by comparison. The human brains signalling system sends much less information at a go.

    Bottom line, the basic architecture for a human brain and a computer is probably too different for us to be able to manage to copy it in a computer. The only animal's brain that we've managed to simulate accurately in a computer is the tiny nematode "C Elegans". Perhaps that's the best we can manage with our current computing power. We tried to build a mouse brain, but we didn't have the hardware necessary. Aka "Blue mouse project"
    Although you give plants creds for this type of problem we disagree on what constitutes proper analysis of the problem presented in which we use both plants and animals. All I assertd was the marker of intelligence in Sea Squirts vis a vis plants was the genetic outcome reached by each in generating behavior. Plants resolve the problem of appropriate behavior using genetics before the realization of the plant whereas animals use genetics to provide systems capable of plastic behavior in the presence of plastic conditions. Huge difference in meaning and implications. A tree might resolve the problem of periodic infestations by beetles by designing in genetic responses for each of the beetle infestations according to other factors like moisture and temperature.

    If physical conditions change unexpectedly and change the relation between physical conditions and pest infestations the tree's response becomes worthless and extinction becomes a threat.

    The animal, on the other hand comes in to the situation with equipment designed to adjust to conditions that have been survived in the past. While it may be challenged by extreme changes, the human bottleneck caused by extreme cold, the likelihood that there would be a change in genotype required is lessened resulting in a higher probability of genetic continuity over the environmental event.

    It's that the tree species would cease to survive because using a priori genetic specification modifications, epigenetic transformation, is more likely to fail than is on the scene adaptability. MY contention.

    Obviously having the capability to build structures and change existing conditions is much more reliable than is methylation of one sort or another.

    As for computer models we can model ants and fruit flies and several bacterial species. Computers are in flux just as program languages are in flux. Given the cloud, infinite memory, one can anticipate humans to exactly model the human brain in the foreseeable future. that won't solve the problem of too large to develop all solutions, but, it will resolve the problem of how brain works.

    thanks for staying with it.

    We're both blowin much out the arse. Isn't it fun. We might actually be doing something serendipitous.

  3. Top | #73
    Contributor DrZoidberg's Avatar
    Join Date
    Nov 2007
    Location
    Copenhagen
    Posts
    7,686
    Archived
    5,746
    Total Posts
    13,432
    Rep Power
    51
    Quote Originally Posted by fromderinside View Post
    The animal, on the other hand comes in to the situation with equipment designed to adjust to conditions that have been survived in the past. While it may be challenged by extreme changes, the human bottleneck caused by extreme cold, the likelihood that there would be a change in genotype required is lessened resulting in a higher probability of genetic continuity over the environmental event.
    A sea squirt floats around until it finds purchase and then stays there for the rest of it's life. It's sucks in water through it's mouth and shoots it out through it's anus. It's lifecycle is for practical purposes identical to a plant. Which is why I used it as an example. It's brain is still very similar in architecture to the human brain. But I understand it's not the point you're making. Just thought I'd point that out.

    Quote Originally Posted by fromderinside View Post
    It's that the tree species would cease to survive because using a priori genetic specification modifications, epigenetic transformation, is more likely to fail than is on the scene adaptability. MY contention.
    But trees do remember and do plan ahead, and do think. Just last week I listened to a podcast where a scientists had dedicated their lives to figuring out just how trees remember. Because it adapts when it starts flowering depending on the temperature this time of year last year.

    Both plant and animal intelligence is adaptive. They both "learn". What sets higher mammalian intelligence apart (and octopi) from humans is more the complexity of thought, and how we process symbolic knowledge. An octopus who has never seen a jar before can very quickly figure out how it works. It has the ability to move around symbolic and abstract models in its mind which it then tries out in the real world. That's remarkable and something different than what you're talking about.

    Quote Originally Posted by fromderinside View Post
    As for computer models we can model ants and fruit flies and several bacterial species. Computers are in flux just as program languages are in flux. Given the cloud, infinite memory, one can anticipate humans to exactly model the human brain in the foreseeable future. that won't solve the problem of too large to develop all solutions, but, it will resolve the problem of how brain works.
    I don't think we will. Not because we can't. It's because it's a worthless enterprise without practical application. If we do it it will only be out of pure blue sky research curiosity. But that won't happen because modelling it with the current tools is prohibitively expensive. And once we have something that seems to work just like a human brain, how would you test that it is? We don't know how the human brain works yet. Intelligence tests don't really measure intelligence. Humans given the same problem will solve it in many different ways. We are creative. That's what's great about human brains. But it also makes us a greased up pig to catch and measure. And that will make the people who paid for the human brain simulation unhappy. If we know in advance they'll be unhappy, we know in advance they won't do it.

    Quote Originally Posted by fromderinside View Post
    We're both blowin much out the arse. Isn't it fun. We might actually be doing something serendipitous.
    Oh, no. What if we create Skynet by mistake?

  4. Top | #74
    Mazzie Daius fromderinside's Avatar
    Join Date
    Oct 2008
    Location
    Oregon's westernmost
    Posts
    10,551
    Archived
    18,213
    Total Posts
    28,764
    Rep Power
    52
    Its Lippy the Lion and Har d'har Har .... or Oggie Doggie and Doggie Daddy .... or Tom and Jerry, Mickey and Minnie, Donald and Daisy .... or Scrooge McDuck and The Beagle Boys .... of Rowan and Martin or ......Dean and Jerry ..... or Ed, Johnnie and Doc ... or the Katzenjammer Kids ....

    enough. It's just Nancy and Sluggo.and Mutt and Jeff along with Brenda Starr, Reporter. Oh look. there's Dondi.

  5. Top | #75
    Veteran Member
    Join Date
    Nov 2017
    Location
    seattle
    Posts
    4,602
    Rep Power
    11
    Will robots have artificial stupefy and artificial character flaws? Will they be prone to anger and emotional outburst? Will they have individual personalities?

    Will they need robot psychologists?

    Current AI mimics mechanicall aspects of humans, like image recognitions and voice recognition.

    If we emulate the total human what happens? Can a robot be jailed?

  6. Top | #76
    Mazzie Daius fromderinside's Avatar
    Join Date
    Oct 2008
    Location
    Oregon's westernmost
    Posts
    10,551
    Archived
    18,213
    Total Posts
    28,764
    Rep Power
    52
    What you ask would be pure conceit of not much use at all. Show me the dollars to be made?

Similar Threads

  1. Classifying artificial-intelligence technologies
    By lpetrich in forum Computers and Technology
    Replies: 3
    Last Post: 01-12-2019, 12:42 AM
  2. Artificial intelligence: more difficult than one might expect?
    By lpetrich in forum Computers and Technology
    Replies: 34
    Last Post: 11-10-2018, 04:53 AM
  3. Replies: 13
    Last Post: 10-05-2018, 09:46 PM
  4. Office of Civil Rights to Suppress Rights to Legal Medical Care
    By Copernicus in forum Separation of Church & State
    Replies: 2
    Last Post: 01-22-2018, 10:21 PM
  5. Google buys two more UK artificial intelligence startups
    By NobleSavage in forum Natural Science
    Replies: 34
    Last Post: 11-04-2014, 07:26 AM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •