[bo1024]

Great points. These are exactly the kinds of issues researchers are grappling with in trying to nicely define, and attempt to achieve, fairness in ML.