How could someone possibly label posts correctly if they don't actually know what accounts are bots. Seems like the training data was selected to aid the answe

How could someone possibly label posts correctly if they don't actually know what accounts are bots. Seems like the training data was selected to aid the answer they were looking for. You have to remember people that believe their point if view is correct tend to share more and comment more so other people know they are misinformed.