I don't know if an intelligence created by humans is going to jeopardise humans; while its information processing power is on a different level, won't the facts that it's mimicking human intelligence and that it is made by humans condition it to its makers' premises? The objective purpose of any healthy living being is to perpetuate its own species. I doubt any kind of mega fast manmade information processing machine running for a billion years could justify the denial of the most primal biological impulse, as long as it respects logic. Logic is essentially consistency, and induction is inconsistent unlike deduction, so unless humans find a way to purposely tweak the AI's decision giving it indeterministic aspects, I think any machine that thinks like a human is not going to disrespect a human's most important premise.
-----
This is of course assuming that intelligence is created in the image of our own. Not that I know if we even know how to create the intelligence of other living beings.