Abaddon, I really don't know how creative A.I. will be either. I've read a lot of different perspectives but one thing that seems to be for sure is that if they do reach the level of conciousness and self awareness, they will not be doing so from a human perspective and may be very alien to us. I think this is more plausible than seeing the Holywood style themes of computers who want to become human. I do think they will want to become robotic and mobile, but our worldviews, sense of creativity, morals, and more comes from flesh and blood. They will have none of that instinctual or genetic influences or the history to fall back on. That is a sobering thought.
It could very well be that we judge them creative and insightful but at the same time we might be terrified of them. I do know this based on human nature though, is that we will build them anyway because we will hope they have answers to our problems. Perhaps one of the new problems then will be, do we like their answers?
Hey Jan, I know you are into computers a lot too. Yes, we don't really understand consciousness but hey, if it looks like a duck, it walks like a duck, and it quacks like a duck...well you get the picture. haha Maybe true AI won't come from human hands. It might come from the machines themselves using a combination of programing and evolutionary algorhytms to create something more sophisticated and capable of working in the world. Maybe we will be so terrified of true AI we will never allow it to develop on any large scale.
Fodeja, I'm very familiar with Eliza and the following improvements, like Ractor (actually this software was used to write a comical book once called "The Policeman's Beard Is Half Funny"} These programs had no real intelligence at all. In fact they were absolutely crazy. What made them seem intelligent at first glance is that they used the rules of language structure very well and incorporated bits and pieces of previous conversational input into the current topic. It would screw up royally though in short order.
These programs are light years away though from neural networks which not only model the way the brain truly works, they generate brain waves as well. Even so, neural networks are in their infancy and are primarily limited now by computational power. Neural networks need massive parallel computer processing...which is just now starting to be developed on a large scale. Once we reach that level where there is plenty of room for the software to "use" then lets see what the verdict is on how far they develop intelligence.
There is an old rule of thumb about what intelligence is. If a man is standing on a railroad track and sees a train rushing towards him, this is information or knowledge. Intelligence is using knowledge to recognize the consequences of what will happen if the train hits him. Wisdom is getting off the tracks.
Using this as a guide...it is a fairly safe bet to guess that computers will gain intelligence this way but will they gain wisdom?
Kind Regards,
Skipper
