Artificial Intelligence

Fission2 · Posted: Mon Aug 13, 2007 1:22 am Post subject: Artificial Intelligence

Can a machine ever be sentient and have intentionality? Can a machine understand things, have its own volition or is it forever condemned to follow it's programming? To what extent is a human different than a machine? What is a person?

This is a philosophical topic so there is no right or wrong; use as many examples of philosophers or philosophies as you please but try to be intelligent =P

~Have fun; This section needs some new blood~

red255 · Jikan Samurai Joined: Mar 13, 2007 Posts: 923

Gonna say ... can a human do those things too?
but you added that. My answer is I don't think most humans can do that.

Why people can be so easy to read by someone trained to look for stuff like that.

watched a show on Sci-Fi today about this guy 'reading people's minds' or something where he was just reading their tells and other stuff.

point is, people aren't sentient. they just think they are.

PsiKro · Heimin (Commoner) Joined: Dec 10, 2006 Posts: 86

I think its quite possible for machines to think and act like a human. Because in truth, humans too seem to follow a strict program. Our responses to certain things are predictable, such as being scared at certain things, being angry at certain things, wanting certain things, etc. We all have free will yes, but natural instinct and parental supervision has programmed us to act in certain limited ways, just as machines can only be programmed to act a certain way. What red said about 'reading minds' is a perfect example of how easily we humans can be taken advantage of due to predictable programming.

It is true that one of our greatest qualities is the ability to create. But even so, there will always be things that we can't even imagine. When it comes down to creativity, witting stories, poems, creating art, etc, most of it comes from things we have already thought of in the past. We simply build on them. It is difficult for us to truly think of a brand new concept. We are truly limited in this regard, which makes me think that we really aren't that smart. I strongly believe that a machine could be programmed to be just as sentient as a human in the future.

Fission2 · Posted: Tue Aug 14, 2007 12:16 am Post subject:

This is a thought experiment by the modern philosopher John Searle, widely known as the "Chinese Room" Experiment.

Imagine that you are in a closed 4 walled room, with only 2 openings on either side. You only understand English and don't know any Chinese. In the centre of the room there is a table and a book written in English. The book tells you how to match Chinese symbols to other Chinese symbols, but does not tell you what they mean.

Through one opening, a piece of paper comes through in Chinese symbols. You don't know what that message means, but you go through the book which tells you how to match these symbols to other Chinese symbols. You write down what the book tells you to write and you pass it through the other opening.

You become proficient at doing this task. The Chinese people outside are amazed that when they insert a message, a response comes out. Soon, the messages you receive are more complex then before, but your book tells you how to match those symbols all the same. Eventually, the messages you write become so advanced, the complexity of your responses appear to be that of a native Chinese speaker. The native Chinese outside then conclude that this room must be intelligent because it understands Chinese.

That is not the case. You, the keeper of the room, know nothing about the Chinese language. All you know is how to match Chinese symbols, without understanding of what they mean. This is an obvious analogy to how a computer works. The room is the computer, the book a program, the piece of paper coming in the input and your response the output.

If all machines are capable of doing is matching symbols and creating responses, then is the machine still sentient? Searle says no. Searle further says that from this experiment, he proves that machines can perform tasks, but not have understanding of the task they are performing. Do you agree?

Machines lack something called intentionality. When you eat, you eat because you are hungry. When you cry, you cry because you are sad. Machines do not. When they do something, they have no reason to do what they are doing (Guns don't kill people, people kill people).

xenocrisis0153 · Posted: Tue Aug 14, 2007 12:49 am Post subject:

PsiKro · Heimin (Commoner) Joined: Dec 10, 2006 Posts: 86

Well, in my opinion humans basically do the same thing. We start off as babies without really knowing anything. Our parents and teachers, everyday events become the book. We begin to match responses. The difference is we insert our own experiences into the book instead of having a book with all the answers. For example, we begin to associate loud noises as potential dangers. We begin to match bad smells as something to stay away from. We begin to associate the way certain people look as suspicious or welcoming.

This 'Chinese room' theory in my opinion, is only accurate to a certain extent. What about when computers gain the ability to learn instead of merely match responses? This can be done through experience. As a program experiences things and interacts with the real world, it can store in its memory banks, similar to humans, past experiences and act accordingly, just as humans do. Eventually an artificial being could being to realize reasons for doing these things as long as it is programmed the way humans are programmed: to survive. The moment it is damaged or threatened, it will at first, follow a set of instructions to do whatever it can to save itself. It will soon learn 'I stay away from loud noises because in the past I have been threatened by a falling object' or something similar to that. It no longer just does it. It knows why to do it. I personally don't think that a humans way of thinking is that much different. I don't have much back support for my argument right now, just my opinion.

I do agree it might not completely be possible for them to understand things as we do. But to a certain extent to which is is very very similar to a human, yes. Once something can learn, it can develop.

red255 · Jikan Samurai Joined: Mar 13, 2007 Posts: 923

I think this has gone into a free-will, pre-destiny debate.

I mean knowing all the factors of what a persons brain is doing (i.e. their history/programming) you will know how they will make a decision under certain situations.

frankly it depends how you define sentient, IMO. I mean you could say the WEATHER isn't sentinent.... but you can predict how people will act about as easily as you can predict the weather.

technomo12 · Posted: Wed Aug 15, 2007 11:42 am Post subject:

actually to tell most people the truth

THE MORE YOU KNOW THE MORE STRESSED YOU ARE AND IM VERY ENVIUS TO KIDS WHO ARE JUST FRESH FROM THE WORLD CUZ THEY DONT KNOW ANYTHING YET SO IT HAS TO BE THE SAME TO AS MACHINE CUZ THE LESS PROGRAM THEY HAVE THE MORE ABLE 4 THEM TO LEARN ON HOW TO MORE HUMAN LIKE well that my point but hey im just saying thing that really do seem very logical
_________________

07cbdj · Posted: Thu Aug 30, 2007 3:34 am Post subject:

mangaddict_reborn · Naginata Ashigaru Joined: Sep 06, 2006 Posts: 483

It is said that human minds can possibly be downloaded into a computer program to immortalize the person. If so, then computers can have intentionality.

07cbdj · Posted: Sat Sep 01, 2007 2:06 pm Post subject:

zyphelin · newbie! Joined: Oct 03, 2007 Posts: 10

as a computer engineer I do believe that In the not so near future(far)....
AI's would approach 50% of the human brain's functionalities...
It is quite difficult to make such AI's because materials such as semicons, chips, memories etc today are only enough to calculate logical functions
AI's aren't rational... They just choose either 1's or 0's or simply yes or no...

I think nano technology would be the best solution... so far I think only 5% of brain functionalities are immitated by AI's...