While thinking about the debate in class earlier today I realized that I agree and disagree with the Turing side of the Computer argument. It appears that he believes that computers will one day be able to think. The definition of think is to have a conscious mind, to some extent of reasoning, remembering experiences, making rational decisions, etc. By this definition a computer may be able to think one day based on it’s programmers ability to “teach” it to reason and choose something. For example for the chess playing computer, the computer rifles through its memory until it finds the code for the choice that has the best chance of winning. Based on this i agree that computers may be fully capable of thinking for themselves in the future.
However, the argument was basically about the ability of a computer to understand. The definition of understand is to perceive the meaning of; grasp the idea of; comprehend. Computers, i dont care how advanced, will forever be incapable of understanding. I mean we humans created understanding, and sometimes we dont even understand the complexities of certain things and/or situations, so how could a computer understand something? For example, my robot could see me and my dog everyday and it could watch my dog bark and then me feed it. Now one day when I am not at home my dog may go to my robot and bark. My robot because of it’s past experience may feed my dog, but if my dog doesnt eat, what would my robot do? It wouldnt understand the concept that my dog may wish to go outside or just play because all it can do is duplicate what it “thinks” is correct based on what it’s seen done in the past.
We also discussed the ability of a computer to care. The definition of care is to have an inclination, liking, fondness, or affection. Computers do not care about the things that they do. They do whatever we humans tell them to do without regard for if the things we are doing are ok. And in the future I don’t expect that they will be able to care either. What programmer would program their computer to care about what actions they perform? That programmer could face the chances that the computer may not wish to complete a command because it doesn’t care to do so. For the class example, the Turing believers stated that computers (robots) through past experience may see someone that they interpret as sad and give them a hug. But that isnt an explanation of the human form of caring. Many people are taught that if someone is sad, it’s nice to try to do something to cheer them up such as show them affection; however that doesnt mean that everytime i see someone that is sad i will give them a hug. And the computer that gives the sad person a hug isnt giving the person a hug out of caring, they are just performing a remote action that they think is correct, based on their programming and past experience. Also the computers interpretation of the person’s body language could be incorrect. They interpret the person’s mood based on their programmer’s understanding of human body language and the programmer could be wrong.
I also felt like the Turing believers placed a lot of human qualities in the computers based on the fact that we (humans and computers) are designed. i didnt really agree with this because even though computers are designed by people, they are not alive, they are not human, they are objects and should be seen as such, regardless of their intelligence. Besides our ability to think, I think humans are separated from animals (and computers) by having the ability to feel and express our fellings verbally and computers will never be able to do that.
In other words, I dont believe a computer will ever care or understand anything, no matter how advanced science gets, because we as humans don’t even completely scientifically understand what it is to care or understand so how would a computer science person program a computer to have those feelings? Also is it really caring/understanding if someone else gave you that understanding or feeling?
Just for something randomly extra, click the link
Lindsey Schwartz | 12-Oct-06 at 9:44 pm | Permalink
This was a really good post. You said a lot of the things that I believe and you brought up some new ideas that are really interesting. I liked what you said about how Turing tries to give computers human qualities - sure, human programmers put more and more capabilities into these machines, but will a computer ever have a heartbeat, feel pain, or have compassion? I think the closest they will ever come is being forced to demonstrate these things.
Randi Burns | 13-Oct-06 at 12:01 pm | Permalink
Yea i agree. Computers will never be able to feel because though scientists understand that we feel, i dont think there is a way for them to make robots/computers able to feel. I mean computers wont have the capabilities of having chemicals stream through their bodies (for fear of messing up their hardware) and is there really any other way to duplicate the sensation of feeling or the emotion of feeling?
Jessica Podeschi | 19-Oct-06 at 6:05 pm | Permalink
I agree with alot of what has been said already in this post. I was on the undecided side of the debate last week and I am still not sure which side I agree more with. I think Searle is right in his belief that machines can’t think, but I agree with Turing that they are capable of intelligence, because to me, the ability to learn from our experiences is a sign of intelligence.
Take, for example, someone who does not feel pain. It is merely a mutated strand of DNA, but the person is still intelligent. So I don’t see how things like the ability to feel pain and compassion are necessary in the definition of intelligence.
On the flip side, however, I think that it is impossible for a machine to understand, for many of the same reasons Randi has already mentioned.
I would just like to add another aspect that I thought of while reading this post. Isn’t it kind of comforting that we, as humans, are unique? Why would we want to create something equal to us? I don’t know about you, but I fear that possibility.
Christopher Lauschke | 19-Oct-06 at 6:23 pm | Permalink
This is in response to Jessica’s last thought. I thought of a couple of reasons why people might want to make computers equal or better than humans. Some designers/inventors are just looking for a challenge. They will puch the envelope jsut to see if they can, maybe without looking at the consequences. I could also see laziness being a big reason for designing equal capable computers/robots. The more equal to us that thay are, the more abilities they will be able to mimic. The more abilities that they can mimic, the less real work humans have to do.
Randi Burns | 19-Oct-06 at 10:07 pm | Permalink
I agree with Jessica. I am fascinated by the fact that people take science to unbelievable levels and are doing so many things with technology, however i am comforted knowing that we as humans are still special, that we are still, as of yet, unable to be completely duplicated. I think christopher is correct about designers wanting a challenge and humans being lazy wanting robots as much like us as possible. But i still cant fathom coming home one day and maybe talking to my robot about my relationship and my robot completely understanding and “feeling my pain”. I would want a robot that will cook and clean before i get home but i dont need it to do anymore than that.
Daniel Estrada | 06-Nov-06 at 1:15 pm | Permalink
I think you argument against machine understanding is too simple. You argument simply points out a particular situation in which a machine responds with X, when in fact Y and Z are also appropriate responses. The way around this argument is simply to design the machine to respond with Y and Z in some circumstances.
The interesting question is Turing’s question, which I don’t think you adequately addressed here: What happens when the machine responds in all the ways humans can respond, and in the situations in which the human would consider that response appropriate?
For instance: Say we have a machine that can cook food. It can use a variety of ingredients, it can dynamically adjust the spices it uses or the temperature it is cooking at to respond to the cooking situation, and so on. The analog of your objection is something like this: “But sometimes people are vegetarians, and the machine wouldn’t know what to do! Humans can make those decisions.”
But if that’s the case, then the machine is just badly designed, and it wouldn’t take much to make the machine sensitive to exactly those kinds of concerns that occupy a human in the same situation. Even if we don’t understand exactly what makes humans sensitive to certain things, we can still mimic the behavior of humans in order to capture the range of sensitivities.
So then we have a machine that can do everything a human can do, in exactly the same circumstances in which the human would do it.
Now why should we say that this machine doesn’t ‘grasp the idea of’ cooking? What’s left for the machine to understand?
Randi Burns | 07-Nov-06 at 4:10 pm | Permalink
Your point is well taken Daniel….so well taken that I may not be able to answer adequately. I don’t know I am fascinated by technology but I may also be short-sighted and simply unable to fathom a machine being able to choose correctly from a “pot of choices” all the time.
Humans have the ability to respond to different situations because we have been bred with sensitivity and compassion, complete honesty no matter the situation, etc. I can’t really “wrap my mind” around the ability of a machine to be programmed with various choices and then being able to pick the correct one all the time. Even some humans don’t make the correct choices all the time so I can’t really see how a machine could adapt to different situations all the time simply because its programmed that way.
A machine can be programmed with different options but not every situation can be calculated with statistical probabilities so I dont think a machine would pick the “right” choice all of the time.