Data mining used to be a sort of dirty word in math modeling. It meant that when you oversample a (large) data set, you'll find whatever it is you want to find, rather than something that really could be supported by the data.
Thus, data mining can be used to provide pseudo-scientific justification for racism. That bias would be automated in software algorithms, such as are used for loan processing and sentencing applications.
Algorithms, data mining, racism and other miscarriages of justice Show more
While probably not broadly known outside of the fields of math and statistics, within those closely related disciplines how to produce and how to avoid producing biased results are well understood.
So whenever you see an article suggesting that an algorithm is somehow innocently being used to enforce systemic racism, realize this is bullshit.