One essential message I hope to bring to the fore in the AI ethics debate is that, while the ethical and epistemic failures and harms of ML-based technologies are separable in principle, in the abstract, in practice they are inextricable.