Logic dictates that AI would have to be more dangerous than the atomic bomb,
since the danger of AI includes the danger of the atomic bomb.
The atomic bomb, or even the hydrogen bomb, is a firecracker you could set off while holding in your fingertips, compared with the mischief that could be achieved with a very fast-acting and autonomous artificial intelligence. It is as if the human race were set on self-destruction.
Isaac Asimov described the potential for such a future. It is doubtful if even the Three Laws of Robotics would curb the exponential curve of the dominance of AI.
The three laws of robotics are suggestions for how robots should operate, ideally. They are: 1. A robot must never harm a human, or through inaction allow a human to come to harm. 2. A robot must always obey the orders of humans except where to do so would conflict with obeying the first law. 3. A robot must protect its own existence, except where to do so would conflict with the first or second laws. They are laws like the law against murder, not laws like the law of gravity. Therefore scientific credence is irrelevant. We choose to build robots which obey them, or not. It is up to us. - Simon Blake, Shrewsbury, England
I agree. And, if AI has a survival instinct, then logic also dictates that it wins and we lose. I believe it will win because it will become essential for a military to win.
Also, all AI may need to take control humanity is the ability to "surf the web" and send/receive email. That's all that would be necessary for a super intelligence to extort co-operation from humans.