Now there might be things to verify beyond errors - points of view, opinions, approaches, etc. But the idea that the job of humans should be checking AI work for mistakes is likely not true today, and good riddance.