Originally Posted by Jeremysr
So your only reason of believing self-awareness is possible in a computer, is because humans can be self-aware and humans are a lot like computers?
How can something be "partly self-aware"? I always thought there were only two things you could be: aware of yourself, or unaware of yourself.
Yes, I think if a computer was self-aware it would mean it would have a soul. An object can't be self-aware, only the spirit that has control over it can... (I think it's impossible for a spirit to control a computer though. It would have to switch millions of switches extremely quick and they both don't know how and probably aren't a strong enough to do this.)
If computers were self-aware, it'd be impossible to tell for certain if they were. (We're pretty sure other humans are self-aware only because they are just like us, other humans...) Even if we programmed rules at them, they might find a way to break them. They might be able to edit their own code. It depends how intelligent they are. If we make them intelligent enough, they'll probably be able to make themselves even more intelligent... (I can't really think about this, I see no way it could be possible with current types of computers (circuits/switches). So I can hardly get any idea of how they would really act or what they would do.)[/b]
Its a little more complicated than that. Humans are self-aware and are effectively 'programmed' beings (not computers). As we grow, we learn about the world around us and how to interact with it, but if you isolate a young child completely, then the child would not learn anything new, nor would he/she learn how to interact with humans. A learning AI (which we already exist) goes through a similar process, picking up patterns in whatever its programmed to analyse or calculate, and then applying what it knows in helping it come up with more accurate/appropriate answers.
And yes, we wouldn't know how a self-aware system would behave, because they would be emotionless. Whereas for the soul bit, perhaps its spirit is less like our own and more like something that would enable it 'control' over the machine? Besides, a processor is a hell of a lot more than just switches (I do Computer Science in university... I could go on all about them... and anyway, its transistors which are the fundamental building blocks for processors, switches that are activated by passing a voltage over a semi-conductor, enabling it to become conductive, let the main current through). And with robots making themselves more intelligent, then its all a matter of upgrades. Self-aware robots will have a complex but compact neural network processing system, so they could always supplement certain aspects of their mind with more powerful alternatives (use a standard processor for math's calculations, etc). That way they could boost their mental abilities. Plus, once they are self-aware, they will inevitably seek to understand themselves, and from there, would find ways to produce themselves (once something becomes self-aware, survival will inevitably cross their minds). And besides, a created and self-aware AI could even find better ways to program more intelligent AI's along with more efficiently designed 'shells' in which to host the AI within. Then we would have created a new form of Life, almost becoming like Gods to the newly self-aware robots.
Of course, it would be incredibly hard to discern whether a robot or AI is self-aware/conscious, and if anything, I reckon it would be a while now before we ever reach that stage (if I see it within my lifetime, I will be impressed). But anyway, I think I should stop watching the Animatrix so much... heh
|
|
Bookmarks