Which begs the question: how do you identify unfair data? If the data is not false or inaccurate, you shouldn't tamper with it. On the other hand, if you alrea

Which begs the question: how do you identify unfair data? If the data is not false or inaccurate, you shouldn't tamper with it. On the other hand, if you already know at the outset the result that you would like your algorithm to produce, then why bother with machine learning at all? Just hard-code your output.