Friday, October 19, 2007

The Creation of AI

In my opinion, AI (Artificial Intelligence) is bound to happen. I doubt it will be very soon (hopefully within my lifetime), and I'm sure the manner in which it is developed will be completely unexpected, but barring any catastrophes like World War Three or a giant asteroid hitting the Earth, I think AI will be created within a century.

But what will it be like?

It will have emotions. A rather common shtick in science fiction is that robots don't have emotions (one such example is Data from Star Trek, at least before the seventh movie Star Trek: Generations, in which he installs an emotion chip). The reason I think that won't be the case is that having goals, or wanting, is a form of emotion. You would expect robots to do things like protect its own existence, but why? The only reason is because it doesn't want to be destroyed, and that's an emotion. Now, an AI probably won't have very human emotions. But it will have emotions.

It will not want to be more human. What's to like about being human? Our bodies are easily destroyed and our brains move at glacial speed compared to a computer's. I can't think of any good reason for an AI to want to be more human. Of course, if you can, please say.

It will not want to destroy humanity. Again, why would it? It might consider humanity a threat, but trying to destroy humanity would be the single best way to make humanity want to destroy it. If it's just one AI, there's no way it could destroy all of humanity before being destroyed (what it was trying to prevent) and if it were many very powerful AIs, humanity wouldn't pose much of a threat anyway.

It will not have Asimov's three laws, or an equivalent. I say this mainly from the perspective of a programmer. For a computer to be actually intelligent, it's code would have to be mind-bogglingly complex and probably, capable of changing its own code. I see no way for codes so simple to be incorporated infallibly into an AI.

It will be better than us at anything we can do. Computers can do hundreds of calculations in less than a second right now. With appropriate hardware, they're much stronger and faster. When we get them to be intelligent, they'll be better at that too.

Obviously, I could be wrong about all of this. This is just my opinion.


bucket of calculators said...

I really like your points made about AI. I feel it is a reality, but not for a while. But they already have computers playing chess and doing practical logical problems and games. I do wonder if any of the science fiction movies will be true. Will they take over? Will they feel discriminated by humans? Will they build there own city and live in coexistance with humans? Will they evolve out of human control and build themselves? It is all very interesting and a little disturbing. I mean think about it; Something without life walking around the streets with us. Just a bunch of homeless computers sitting in the park feeding the birds. What I wonder is this: If you say they will have emotions but not human emotions, than what kind of emotions? Will they feel they have a reason to exist? Will they question why they are on earth? Will they want what humans want? Friends? Love? Hobbies? Materialistic things? I hope to one day see what exactly robots do and how they act. Who knows, maybe they will be doing the stereotypical maid work around my mansion...

Haakaa Päälle! said...

I hope I gave my opinion for your first set of questions in my last post.

As for the questions about emotions, I really have no idea. It will probably depend a lot on what they're for. Other than that, I can't say.