Thursday, September 9, 2010

Stephen Hawking 2.0

I'm not a fan of films where aliens or robots attempt to exterminate the human race. Too predictable. $200 million and 2 1/2 hours later, we've won through a series of improbable events. Like Independence Day, when Jeff Goldblum takes out the entire fleet with a computer virus. This only works because the ships are running on Windows. Blue Screen of Death.

Personally I'm all for space travel. Not enough room on this rock. But the idea of artificial intelligence is... a little worrying. Though the "in" term now is Artilect. By the end of the century is it predicted we will be able to construct artificial intelligences millions, if not billions of times more intelligent than humans. The stuff these intellects could do is literally inconceivable. And therein lies the problem: if we make machines more intelligent than us, machines capable of churning through millenia of human thought in seconds (if that), what happens to us?

Computers now are basically number crunchers: we put in the algorithms and out comes the physics. Or it plays ping pong. Mmm, progress! My laptop is excellent at maths but its not intelligent. It doesn't think. Human endeavour still matters because we're still the ones punching in the numbers. But what would be the point of anything if we had a machine that could run literally trillions of scenarios and give us all the answers to every question put to it?
For example, there are 88 notes on a piano and for arguments sake, most music works within this parameter. There are only a certain number of notes. And if you add words, of which there is also only a finite number, there is still only a certain number of songs possible. An artificial intelligence could write every single possible piece of music, without the personal and artistic processes. Such a machine would be beyond humanity in any conceivable form.

So far I'm still working within mathematical sequences. Personally I think it is important that we create artificial intelligence, but even more important that we don't bite off more than we can chew. The first digital person would be a landmark achievement in human history. It would help us understand ourselves. I cite "Bicentennial Man" as an attempt to grasp this idea, at least in popular culture. Andrew Kennedy in his book "Who is human anyway?" presents the idea that artificial intelligences will have severe personality disorders and will fall into four groups: the autistic, the collector, the ecstatic, and the victim. This makes sense. Just because a machine would be extremely intelligent, it would exist within an artificial environment and wouldn't be immersed in a social context like humans.

Just so you know, I'm confused by all of this and I'm trying to work out something. The difference between my laptop and an AI is that the former comes preprogrammed but the latter... doesn't. We shouldn't anticipate SkyNet coming to destroy us all. If you think about it, that entire franchise could have been avoided if they didn't give SkyNet access to the internet. Or set up an 'Off' button. A computer without internet is effectively useless.

I think what people are scared of is the complete redundancy of the human race. Hell I know I am. A future of indolent humans lying around while our robots do all the work? I think not. AI will be most useful in science I think. Leave the Arts to the meat bags. And don't forget the 'Off' button.

Just so you know I'm aware this isn't the most elegant grasp of the AI debate.

1 comment:

  1. Elegant doesn't matter - it's a really very insightful / thought-provoking post: thanks.

    ReplyDelete

Note: Only a member of this blog may post a comment.