Re: virtual world
Hi Tom.
Your right about startrek, indeed I should've remembered the sentient Moriarty program because I am watching next gen at the moment as I've just bought deep space 9 and am about to start it. Actually I'm looking forward to watching ds9. I bought the series about 4 or 5 years ago but remember very little of it as I was extremely ill at the time, however I lent it to my brother who proceeded to lose several of the boxes meaning I ended up having to rebuy the hole thing.
In terms of Ai however, actually Data as a character (even ignoring his ability to walk, talk and have human senses), is pretty advanced in a way that no ai we have now is.
for example, I have taught Reever to find various objects in her guiding job such as doors and seats. These things she learns very quickly. She was able (as indeed a human would), to categorise the concept of "door" as applying to innumerable different things. it doesn't matter if we're inside or outside, if the door is wood or plastic or metal or glass, or has a window, or is revolving or automatic, or is closed or open. No matter what factors are involved just like a human Reever is able to identify "door" quite accurately, indeed accurately enough for me to risk getting lost on her identification.
Now if you were inventing some sort of visual scanning program to identify doors you'd need to program in a massive bunch of variables, you'd need to program it to recognize inumerable shapes, sizes, situations and positions, heck I'm not even certain that the unified characteristics of door are ones you could easily apply.
From the last I heard about android research, they were just at the level of getting androids as you said to recognize heat and cold and other vague sensations, indeed I particularly liked the idea of the android called graspy which was an android with a vocabulary of touch who could identify obj ects as soft or hard or rough or smooth, or use various other tactile descriptive terms. but I've not yet heard of any research that have been able to put these sensations together into forming the sorts of intuative categories of objects that humans and animals use is still dam difficult.
Another key problem I'm always struck by in the field of Ai is the question of consciousness, sinse one of the major issues in the philosophy of mind is whether it is possible to derive consciousness from function.
There is a very famous problem which illustrates this called the chinese room.
Imagine that you had a room with lots of draws labeled in chinese characters, which contained notes written in chinese. In this room is a person who doesn't themselves understand chinese.
Suppose you wrote a note in chinese and posted it into the room, and the person in the room took the note, looked at the chinese characters written on the note, matched them to a dr aw and posted the note back out.
So your note could say in chinese "Hello how are you" and when the person looks in the draww with that chinese message on there is a note saying "I am fine" which the person posts back to you.
You have hear got an intelligent response in chinese simply by virtue of pattern matching, and if there were enough draws and enough notes for this none chinese speaking person to post back you could imagine an intelligent conversation in chinese.
The problem however, is that obviously the person in the room has no idea what the heck is going on sinse they don't speak chinese themselves, and though you have what is apparently an intelligent conversation going on, there is no actual consciousness involved.
Now this example comes from the 1950's, when computer programming was very much in it's infancy, however you can see the problem for an Ai. With a human I can be relatively certain (at least as cert ain as cartesian doubt will let me be), that that human is a conscious being, that they are not just responding to certain inputs with given outputs, but that there is something else going on more fundamentally.
I can also make a rough intuative guess that at least some animals share this level of consciousness and do not simply respond mechanically, at least to some extent.
However, what about a computer program?
This is the area where I am not convinced on Ai research, and indeed the area where I tend to disagree with biological determinism as well, sinse consciousness does not seem to reduce down to any sort of explanation of inputs and outputs, especially as you said when you considder emotion, personality, and all the fun things conscisouness that being conscious involves.
It is fairly clear in Tng that data is intended to be a conscious being, indeed I recently watched the episode where he activated a part of his programming that produced random dream ing, which you actually got to see on screen. This was great as a story but I'm not exactly certain how possible it would be in reality to replicate, indeed the more that is learnt about the quantom structure of the brain the less likely it seems consciousness has a particularly easy answer.
Thus, I'm not exactly certain whether you could! write a program with that actual quality that makes people people.
Regarding the term contacts, well to an extent I agree, after all that is what E-mail programs call them and it is true in my contacts list in outlook express I have everyone from my tutor, to my parents, to game developers and even paypal. However I do think tit would be helpful to have some sort of extra terminology which could be used to distinguish people you merely stay in contact with for necessity such as paypal or game devs or formal arrangements, people you know in real life and are friends with or related to but use online means to stay in touch an d people you actively are friends with online but have not met in real life, ---- electronic pen pals as you put it.
The third group particularly is one that has only just arisen in the past few years thanks to the use of the net, sinse previously friendly contact between absolute strangers who had never met each other was far rarer and more difficult.
URL: http://forum.audiogames.net/viewtopic.php?pid=173740#p173740
_______________________________________________ Audiogames-reflector mailing list Audiogames-reflector@sabahattin-gucukoglu.com http://sabahattin-gucukoglu.com/cgi-bin/mailman/listinfo/audiogames-reflector
