When we get actual AI, that question may make sense.

can an AI make the right decision 100% of the time?

Can humans make the right decision at a rate better or worse than AI? If AI makes the wrong decision, is a human empowered to step in and overrule that decision?