In Cathy O’Neil’s chapter 6 of Weapons of Math Destruction, the discussion turns to personality tests and how they can determine one’s job. There is a lot of discussion in this chapter about how it is questionable and potentially illegal to discriminate based on something like the five personality test. This is in line with the topics in the book so far as it is another example in which a model is applied to something in the hopes of increasing some positive aspect, but instead comes with consequences that would not have been seen with solely human judgement. On one hand I agree that such trivial tests should not determine one’s being hired by a company, as I would surely be upset if I was turned away from any job on account of this. However, I do think there is something that is not being said in this chapter. I don’t really understand the huge problem with using these tests, even when someone would’ve been a good hire without them, the models would shift accordingly if good hires are being lost OR if too many people are being rejected (a store still needs to hire people). The big 5 test is not trivial like what kind of ice cream you like, it can actually hold vital and relevant information. This all being said, I think there is a real problem with employers tracking and trying to optimize employees after they have been hired. This is something that really would clash against a model, and would be much better being evaluated by humans. Thankfully, it seems that tech companies, at least, have realized this and allow for much more relaxed work environments (there are some that are worse, but I suspect this will change over time). I am much more of the belief that the output of one’s work is much more important than the time they spend on it (in an office setting, no less). In this way, I still agree with the broad point Cathy is making, but I think the specifics miss the mark in this chapter.
