I was watching a special on the history channel about 7 ways that scientists think the world could to end. The first two were not unreasonable – ‘gamma ray bursts’ occur when a star collapses, and if this happened in our galaxy, apparently the burst would destroy our ozone layer. Without that ozone layer, we’d be killed by solar radiation. The second way we could potentially meet our doom was a roaming black hole. If that thing came too close to the earth, we’d be done for. The physics sounded about right to me, although a few things were exagerated. If you fell into the event horizon of the black hole, you’d be dead before you were able to tell what was going on. I don’t think it’d be particularly painful, and the gruesome images they gave, although they might have been accurate, would have occurred so quickly that they’d be irrelevant.
The third way we could meet our doom, however, was completely ludicrous. They had some AI researches (as well as steven hawking) claiming that computers would become more intelligent than human beings, and that for some reason they’d decide to wipe us out. I think the entire field of AI is completely misnamed. Most of those guys would, I think, make lousy theoreticians. They spend their time dreaming up ’solutions’ to NP Complete problems that run in exponential time and have absolutely no guarantees. When they’re not doing that, it appears that they’re busy making absurd predictions and prognostications about the nature of intelligence and what computers will do for us. They even had Steven Hawking claiming that within 100 years, computers would be more intelligent than human beings.
First off, people have been predicting human-level intelligence from algorithms for years. I’m not saying it won’t happen, but I think we’ve got no way of knowing how far off in the future it will be. Personally, I don’t think we’ll have human level intelligence from computers untill we can simulate an entire brain on a cellular (or maybe even molecular) level. When and if humans do produce computers with human level (or greater) intelligence, the AI that is produced will be used to accomplish all kinds of trivial crap that human beings don’t want to do. The thought of an AI that would decide to destory the world, and have the capability to do such a thing, is absurd – the AI would have to want to destory the world, and the only reason it would want such things is if it were designed to want them.
If you designed an AI capable of talking with human beings to do things like answer phone calls, you’d make it so it would be happy if it satisfied customers, and unhappy if it didn’t satisfy the customer. Human beings have all kinds of tendencies that we’d never have any decent reason to incorporate into AI – boredom, tempers, and greed. I could go on an on about this, but basically i think it’s completely absurd to be afraid that computers will take over the world. If we ever develop computers that pass turing tests and can think like humans do, it’ll be a great day for humanity because we could eliminate the need for all kinds of thankless, unpleasant tasks.
In any event, the thing that struck me the most was the difference between a small town fireman in Iowa talking about how he’d respond to one of these distaters, and the big name scientists they had talking about how the world was going to end. The scientists seemed so sure of themselves and their own intellgience. The fireman and police captain in the small town said that, in the event that one of these disasters were to occur, perhaps the only thing they could do would be to pray. I dig that. I dont like arrogant people. I’m afraid I am becoming one, however. It’s something I’ll have to watch out for.