That would not be a better solution because it would entail either “solid” rules that are actually always fluctuating and/or people having prior assessments overturned after the fact.
As opposed to… humans working through a specific situation for a specific solution?