This chapter investigates transferability of classical and quantum adversarial attacks, i.e., the performance of classical classifiers against quantum adversarial attacks and vice versa. The empirical results indicate that quantum adversarial attacks are able to trick classical classifiers significantly reducing their accuracy. However, the converse is not true: the adversarial attacks generated from classical neural networks only slightly impact the accuracy of quantum variational circuits. We argue that the enhanced resilience of quantum classifiers to classical attacks is due to the fundamentally different nature of attacks which is exhibited by visible structures in quantum perturbations whereas classical perturbations tend to show no such structural information. This introduces a potential new avenue for quantum advantage in the form of adversarial robustness against classically attacked datasets.