[MichaelMcQuirk]

As long as you are able to create and set goals for the entity (AI), goals that it must obey and cannot override, then this should never be a problem.

The three laws of robotics would be a good place to start. My greatest fear is it's code landing in the hands of someone with ill intentions. Imagine it, a machine with the combined knowledge of almost everyone who ever lived, capable of working 24/7, full speed, without any breaks. Given enough time it would be able to answer ANY question you could ask it, it could build anything you desired and it could cure every sickness known to man.

But it could also create unimaginably powerful weapons of mass destruction, it could create a pathogen capable of annihilating the entire human race and yes, if asked, it could create a robot army unlike any depicted by sci-fi before. Our fear should not be that of the AI itself, but rather the man behind the computer screen. For it was neither the maker nor the carrier of LittleBoy that was responsible for the annihilation of Nagasaki, but rather the President whom sent it on it's way.

I just pray that the creator of the first AI, is a kind one. For he shall poses both the greatest weapon known to mankind, And the greatest blessing ever to be bestowed upon us.