We are predisposed, as a species, to worrying. It's probably served us pretty well in the past. It is far better for us to mistake a rock for a bear and run away from it, than to mistake a bear for a rock and stay put.
With regards to artificial intelligence, this is no different. In many people's minds, artificial intelligence is a mysterious force vaguely threatening and inherently unknowable. And because it is potentially threatening and unknowable, people tend to be afraid of it.
But our day-today experience with artificial intelligence--guiding us through traffic, filtering our email for spam--would not give rise to any of those kinds of worries. Now, to be sure, there are all kinds of worrisome applications of narrow AI--warfare and to invade privacy. But the core technology somehow becoming sentient and sapient and acting against us is a worry that stems more from a lack of understanding the technology than from any real facts.