Humans are able to recognize these as optical illusions. (otherwise we wouldn't be calling them "illusions")
Further, none of the optical illusions shown are so wildly "off" and "crazy" as the algorithmic goofs demonstrated in the paper referenced above.