Actually in this sense NNs are actually smarter than humans - NNs can guide you to its weak points, through backprop. That is indeed how adversarial examples ar

Actually in this sense NNs are actually smarter than humans - NNs can guide you to its weak points, through backprop. That is indeed how adversarial examples are found. Vice versa, people are notoriously bad at finding their own weaknesses (as demonstrated by Dunning-Kruger etc).