WALTER FRICK: And there’s an example in the book about this that essentially– I think it’s a hospital– uses a ton of data on previously essentially what have hiring managers decided, thinking, well, we’ve got all this data, let’s put it to use. So that’s another thing that happens, is that once we assign scores to people, we often put them on the hook for bad scores and give them credit for good scores, even if one of the real reasons that they’re getting a bad score or a good score is just by dint of their demography. CATHY O’NEILL: I don’t think you can rely on the individual ethics of data scientists just simply because they work within a company where their boss basically tells them what to do and how to optimize their algorithm, which is usually on money.