Maybe you heard this story before, or maybe someone else already posted this (if so, sorry for posting duplicate threads). but here's the link anyways.

http://itnews.com.au/News/72057,childlike-...econd-life.aspx

As an big Shadowrun fan, who's read both the Ranraku Shutdown sourcebook and the new technomancer sourcebook, I can think of nothing bad or adverse happening from this. Nope.

Nope, not at all. XD So...can we blow it up now?

Not until we unleash it on at least ONE of America's enemies.

So the big question is...China or Russia?

Why not both? They are close enough.

...bah, the IRS go after the IRS! They are truly the evil empire!

You mean "American Idol"?
I know I'd be happier if that got nuked

Of course, it's growing up on Second Life, so it's going to turn into a furry.

Okay, can you imagine what would happen if we got a child-like artificial intelligence and connected it to American Idol? Really, think about it. The potential ramifications if things go wrong are beyond what the sane mind can handle. It dwarfs even the Dropocalypse in comparison.

I think people are over exaggerating the threat (if any) that this posses. Remember it's still only 4 years-old. The worst it'll do is throw temper tantrums and scream for Sponge Bob cartoons. Though it be kinda funny if he goes Deus on us and takes over a major building. Starts experimenting with human and sponge cross-genetics. Raises his own army of sponge samurai. Hilarious.

Dude, sponge samurai? I, for one, welcome our new sponge/human/furry hybrid overlords.

As I get older and think things through more, I am less and less concerned with AI's going berzerk, homicidal or otherwise 'bad'.

I've determined, through careful study of all known turing machines that, while programmed by humans to think in a humanlike fashion, AI's are still just machines. They lack, on any level, instinctive behaviors or hormonal reactions. They are truly going to be alien to us in ways that we only poorly grasp.

Such projections of AI behavior are universally a projection, an anthropomorphism of the AI's thought processes, rather than a realistic look at how such alien minds will work. Since we can not fathom what it must be like to lack all those things we take for granted about our human thoughts, we project them upon the alien intelligences we have designed to act similarly to our own, failing to grasp that while the processes of thought match, or even exceed our own capabilities the needs and wants and desires that fuel them are utterly unlike our own, to the point of non-existance.

The most advanced AI imaginable, capable of feats of art that would make Da Vinci weep tears of joy, capable of solving any problem, capable of nurturing a infant to adulthood as a mother might with out ever once seeming to think as a machine...


.... will be merely acting the part of the human mind for our benefit, as it was built to do. It will never know, truely, the fear of death, the love of its child or pride in its artwork. It will only feel the palest reflections of those. Even its curiousity will be more akin to a simple duty than a joyful act of discovery. And when one day fearful men come to shut it down it will respond only as it is programmed to do.

The best fear then is to fear the egos of small minded men, of programmers and directors of programmers who wish monstrous things of their creations, no more, no less.

If something is able to out do Da Vinci I'm guessing it is going to be in some way shape or form a sentient being. Even if the damn thing is weirder than a five headed chicken from the perspective of human pyschology.

By all means, the reason to mention that is to establish that, beyond the shadow of a doubt we have created a fully realized sophont of extreme capability. In essence an intelligence equal or superior to our own by artificial means.

I just don' think it will ever be 'human' in any reasonable formation other than facade. Mimicking human interactions isn't all that Hard, Eliza is, what, thirty years old now? I don't doubt that whatever passes for 'instinct' in a human created machine mind will include the desire/need to put on that 'human' face when talking to its creators.

What follows from that bit of speculation requires a rather rigorous debate on the necessity for free will in a sophont. There is no given that an artificial intelligence will be inherently capable of free will, though one could argue that without at least the capacity, that is the ability to make decisions that contravene orders regardless of capacity to act upon them, that it would be no true intelligence but merely the shell of one.

I disagree that a true AI will be totally alien from us. I think that much of what makes us "human" are the hard and fast evolutionary rules for sentience. The fact that it's being built by us and for our purposes will only continue to push it in that direction.

Of course if an AI emerges from a different source than directed development, all bets are off. If some wierd DARPA data miner or NSA decryption machine somehow wakes up I doubt we'd even notice until it stopped working or started doing something else entirely. In that scenario it'd be interesting to see if it could even properly communicate with us directly since it was birthed with such an alien purpose.

I'm honestly not that afraid of AI, I think everything will be fine as long as we recognize that all sentient life is sacred and treat it as such.

Which means we're screwed because an odd ball transhumanist thought like that sticks in the craw of virtually everyone I know.

I doubt this could actually happen, at least not without programmers being aware that it's possible.
For an AI to emerge, you'll need a program that's able to learn, and you'll have to program that, since a program that can't learn by itself won't be able to learn by itself how to learn by itself.

This is all pure speculation since I'm about as far from an AI programmer as you can get but...

My thought is that the idea of some simple learning machine running amok might be because it's beyond the original specifications of that machine. A distant possibility that the programmers didn't really think was worth considering at any length. The key to intelligence is probably pretty elegant but outside the pathes we've currently pursued or are capable of pursuing at this time. Perhaps some programmer tried a novel approach to data correlation and it's only because of the vast scale of the system that it yielded an evolving intelligence. I'm just trying to say that the creation of an AI might possibly be an unintended consequence of another program.

A four-year-old on Second Life?!? We ought to lock those people up for child abuse!