Incorrect comments on against “AI risk”