Quote Originally Posted by georgekar View Post
Eventually the AI will kill us...(no reason to support us...for just some kw power)
In fact, "reason"... it won't have reason for anything, you are thinking like it would be a human.

But no, humans have needs that have to be fulfilled, at different levels. If you feel hungry you need to eat, you need to get food. That's at a basic level, but at any level it works more or less like that.
You have feelings, passions... then it gets more complex. But at the end there are motivations that to a great extend they are "programmed" into the "hardware".

But a machine... what needs does it have? Why would it be "wanting" to do something (whatever)? (good or bad)

I've read the book Superintelligence from Bostrom that addresses these subjects.
There are no easy answers about what to do and how to deal with it.

I'm not sure what to think, but one logical thing to think is that this AI (or AIs) will be trained by its owners to work for them, programmed at some low level for those goals or "needs".