AI must be transparent and controllable, for if we don't know why a decision is made, we can't decide if it's correct, says Professor Ahmet Ulvi Turkbag
