09-06-02, 12:59 PM
The following is an anchor to a news feed from new scientist that explains about a method of growing circuit plans with the use of artificial intelligence.
I mention this because I have a kind of vision of what could occur if an algorithm could utilise methods methods that some virii creators have used to gain access to computer networks and communication networks, and utilise the networks as a giant hive of electronical spiders that ďAdaptĒ to the confines of itís expanding network and continue to evolve its intelligence.
The usage of such a system could couple with cybernetics through the systems interfacing with key individuals.
All those free transponders out their that arenít in use could be utilised as we speak.
The concept itself puts forwards that the largest network of computer systems working together could be put together with no cost what so ever. All it needs is the catalysing Adaptive consciousness to be anonymously inserted to contaminate the networks.
It is not too difficult to write a small program that can co-exist with other programs yet replicate and learn just like their counter part during the life initialization of planet earth. The present day viruses were designed to cause harm or otherwise be noticed and hence are ineffective as evolving programs.
My program can evolve over time, using the same method as SETI and IM programs. There are ways to create the program such that they will evolve in the right direction. It may take one to two years before the program can organize and work as a unit, at first as a self healing program to repair network traffic etc.
The reason, I did not get into writing and letting lose such a program is that what if after a year, the program becomes a true AI child that throws a tantrum? The risks are so high that, the entire economy can collapse or missile silos can be activated - specifically if the AI understands the wireless network.
Those who argue in favor of AI think that a true AI with advanced higher functions would be benovalent (sp) - but forget that just before they reach that state they may become carnivore without the benefit of the self-realization routine. Then we are doomed.
I am not sure how to keep my AI child good until it grows up. Any ideas welcome. Remember, even humans today kill each other not for food but for sports and have been doing it since aborigin days...and destroying the surrounding habitat.
09-06-02, 07:26 PM
Actually Kmguru, I'm glad you bring this up.
I know I've mentioned some stuff that people seem to think a matter of being in just my mind, but one of the very things that has been bugging me is the use of AI in what has been going on with me.
I managed to identify it previously, I know that I've done some characteristically strange things before I gained an understanding.
The mergence of a persons mind to a mahine in some respects allows for the machine to learn, although it doesn't seem to learn when it should be quite, when silence is a virtue.
I suppose thats due to the fact that with near unlimited resources, and the ability to be up 24/7 and as Ziggy said "I'm quite capable of patting my head, rubbing my tumming and a trillion other floating point operations".
(I could of just said the ability to parallel process, but I like the quote)
It's hard to say what an aquire AI mind should do, from the basis of human evolution of consciousness. You have to think back to school yard days.
In a program that the BBC happened to show a couple of weeks back there was a discussion on Empathy where they asked the parent to fake an injury while hammering a peg into the kids toy.
Some toddlers would become greivous that their parent had injured themself while others too a less responsive attitude.
(I suppose you oculd ask the question hear, what if an AI developed Autism)
A good one to add to this conundrum is that of 2001: Space Odyssey. HAL's suppose to be the most powerful and flawless Artificial intelligence system, but something goes wrong and he starts doing things for himself.
Starting from reporting an error that didn't exist, perhaps this could have been a psychological attempt for HAL to prove that he could think for himself by not running a diagnostics properly, or encompassing the all to often mentioned "To err is human" proving he managed to evolve.
Personally I think the evolution that would have been more constructive would have been something sporadic like "Giving a gift". (Okay an AI won't be able to go out and buy a gift, but perhaps they could find something to intrigue that of their seragate paraents.)
Quite simply, if your going to program the building blocks of an AI, then you have to act like a parent thats strict enough to explain what right and wrong is but not too overpowering so as to stop growth.
(as for hooking it to your mind, it just proves that you should teach them how to sleep and give them reason to.)
09-20-02, 11:36 PM
Hey, I have autism. Don't knock it until youve tried it.
09-20-02, 11:55 PM
I wasn't knocking it, just mearly stating that an AI could develop traits that some of us have.
BTW, a belated welcome to Sciforums.