Using the past to inform the future. Pretty much sums up the problem. But the issue is that there is no solution. Tech is starting to run into fundamental philosophical limits. With Fake News, it can’t discern truth from lies. With data mining, do users have a right to their data? And with AI, it can’t predict how the future can differ from the past. No one can.

But what it might be able to do is aim for a goal. In this case, when crime risk is (roughly) equal for different demographics.