Whether the risk score is discriminatory on protected classes can be easily measured. (After implementing a risk score system, is the fraction of bail jumpers equivalent across groups?) I assume that kind of measurement is what prompted the overhaul of cash bail.
But unless the system is either completely perfect (i.e., its predictions are never wrong), or completely worthless because it ignores all input data and always gives the same risk score, then there will always be some measurement that says it's discriminatory on protected classes: https://arxiv.org/abs/1707.01195
The paper you linked to describes three different definitions of fairness that don't agree. I gave a single definition of fairness that is self-consistent.
Yes, if everyone agreed that your definition of fairness were the right one (which I personally do) then there'd be no problem. But in practice there will be people who want to use other ones (in some cases because they want to be able to say it's discriminatory).
