Artificial intelligence will not be human, so we must always prevent terms like “reputable AI” that don't just humanize AI but will also indicate a standard of dependability that simply just will not exist. Epistemic principles constitute the conditions for an investigation of