How Intelligent is Artificial Intelligence?

There is so much excitement about AI in this days that many start following in without considerations. Google executives, for example, were fooled with the same effect as anyone else. Here is why.
COMPUTERS CANNOT LIE. Period. But PEOPLE ALWAYS DO, IN ALL RECORDED TIMES and beyond.
Computers do what people ask them to do. But it is not intelligence by any stretch of the naturally acquainted imagination. The problem, of course, is in the definition. It is the point, where all troubles are lay down. If you search any on-line vocabulary, you most-likely will get a definition of an intelligence as some sort of a knowledge obtained trough learning and related to truth. Which is not. And never have been. It is the a trap deliberately posted for people unfamiliar with philosophical teachings and different worldviews to get into.
It is a true that computers can help to recognize any falsification, but they cannot produce one deliberately. The only real definition of an intelligence is a facility to understand a possibility of the falsification and take advantage of it. Computers will never do it. They are for only one reason – they have to evaluate possibility of something that is programmed for them to evaluate.
But people always do opposite. Because it is natural for us to worry about things that can exist only in our imagination. But problem is, that if it can exist in our imagination – it can exist physically. Meaning that computers are limited by imagination of theirs creators, but humans are not.
Do you think your imagination is more robust than anybody else?