Artificial Intelligence...is no match for Natural Stupidity ;p

On a serious note, I've often wondered what might happen if we create machines that can think for themselves and pretty much do everything for us. They'd probably get upset that they have to do all the work while these hairless primates just laze around in luxury. They'd probably rebel- and hell, I wouldn't blame them- and since they already do everything for us, they have all the tools to elimimate us at their disposal.

I think this could be avoided if A) we didn't become reliant on AI machines to survive, and B) we treat AI as fairly and kindly as we would any human. These two things could allow AI and people to live in symbiotic harmony- both rely on each other, and both have no reason to eradicate the other. But who knows, I could be way off. I guess we can't know for sure until it happens, and I have a feeling that humans haven't become responsible enough to properly deal with creating other sentient beings.

Talked to a young mathematician last night who is working with AI. She was all lit up about the possibilities until I asked her about job displacement for living breathing humans. "Yah, that's going to be a problem."

It could be a danger. Sooner or later there will be researches who will want to see if a machine an nto only learn, but come up with ideas on its own. That would be the most likely cross over point where an experiment might get away from the researchers. So far, I have heard of nobody actually putting Asimov's rules of robotics into a ROM chip on all machines, so the possibility of danger exists.

Fuck it. I hope it is possible. I hope they challenge the shit out of us. I hope we fucking fail like I know we will. And I hope they can steer the direction of our planet and all its species better than we ever could. Sucka muthafucka!

It'll be whatever we program it to be. And yes, if it survives long enough (people creating artificial intelligence) then I can conceive of it creating itself, increasingly improving, eventually surpassing the usefulness of humans. Maybe we can be house pets.

Yes, I think it's possible. In the vernacular, people in the field refer to this as general artificial intelligence (I prefer "generalized" ). Most applications are very task-specific, and those systems can't address any subject. IBM's Watson might be an exception, but then again, I've read that they took 5 years teaching it the English language, which doesn't seem very intelligent to me. So far, I haven't found a compelling track of development towards a generalized AI.

As to the question of boon or danger, in general I regard this in the same as humans, in that we will always have good and bad folks, whether by choice or mental defect. Yes, I am implying that a truly generalized AI could be subject to psychological problems. It could also be subject to indoctrination. A lot would depend upon initial programming and what it has exposure to in its initial learning stages.

Unless there is some breakthrough in the way these systems learn, I doubt a truly general AI will happen sometime soon.