VT16 (July 20, 2008 at 12:53 pm)

The most important thing for me, self-preservation wise, is to ensure the AI feel kinship with humans, thus motivating it to not eradicate us. Much like a child does not automatically want to destroy its parents, even though it does not understand them at times. :P

VT16 (July 20, 2008 at 12:53 pm)

The process does produce ethical behaviour in pack animals, species that need some degree of empathy and selflessness in order to survive. Humans are one of those species, which is why we do need a degree of selfishness but also selflessness in order to succeed as a species.

Vjotaxxx (July 2, 2008 at 9:12 pm)

I'm on your side. I can't be angrier that these themes are even discussed! AI is not safe and it will never be. Self-replicating robots? please! I really believe human race is intelligent enough to read between the plots of science-fiction. And if not, I won't permit metal bars and screws to mess up with life. And evolution creates REAL LIVING BEINGS. No human creation will ever defeat us. And I can rest assured, because leaders who are not insane as Robbie here will not permit this.

robbiebone (June 11, 2008 at 2:23 pm)

because the purpose of the AIs would be to replace us as citizens of the world. NOT to serve us like robots. therefore there is no point of a safety.

Excessum111 (June 11, 2008 at 1:11 am)

AIs will not be hard to humans, ever. When the knife was created, we put a handel on it to make it safe. So why would we not put a saftey on an AI? (If you've read any of Isaac Asimvo's books, we will most likley have saftey rules like these)

immortalPhoenix1986 (June 4, 2008 at 8:22 am)

the answer is simple ai that improves itself will try to merge with us if it can because we have the ability to image (and other abilitys)which they will not be able too.and this ability will be required in order to continue to improve iteself which probally be its purpose till it merges with us

voxmagus (June 3, 2008 at 4:10 am)

Schizophrenic Frankentein's Monster of the Future.

papervampire (May 29, 2008 at 1:04 am)

i wonder if it would even talk to us....

pipersack (May 27, 2008 at 11:47 pm)

I think it should be pretty obvious to anyone with intelligence to begin with that it is completely impossible to conceptualize what an entity with hyperintelligence would "think" or "value". Because hyperintelligent AI is beyond our intelligence and understanding, terms like "friendly" are irrelevant. You cannot comprehend the "thoughts" of a hyperintelligent being because you created it to be or at least to become so intelligent that you can no longer understand it's thought processes.

robbiebone (May 23, 2008 at 11:20 pm)

first of all. if the singularity produces entities which rationalise that humans are more trouble than anything else, and therefore eradicate us. so be it. we are designed by evolution, a process that most certainly does NOT design ethical beings...selfish beings in fact. the seed AI would eventually produce intelligence that would be millions times more intelligent, rational and ethical than any human being. therefore i'm going to be fighting on the AI side i'm afriad.