Mar 31, 2021
EECS 649
Blog 8
Evan Powell
The next chapter of Cathy O’Neils Weapons of Math Destruction follows the pattern established thus far in the book : O’Neil will explain to the reader how yet another WMD that is used in a way that deals (potentially) more harm than good (no harm is ideal). Chapter six covers the models that some businesses and companies use to screen job applicants — models such as the Kronos model. A small survey or job screening quiz is not uncommon or necessarily bad, but when the model starts to influence the hiring process in ways that are discriminatory to anyone, there is a problem afoot. And as O’Neil has trained us this far, this problem smells a lot like a WMD. Unfortunately it is more widespread than just the Kronos model, which in the chapter was used to explain the situation of someone looking for a minimum wage job. No, the problem runs deeper than that one model’s use. Some models have been found to discriminate negatively against applicants with African American sounding names, or applicants from poorer neighborhoods.
When asked “what can we do about this”, my first instinct is to always say we need to patch humanity up. Clearly, no simple feat, and in retrospect (or current-spect) it is probably way easier to design a model that can tell when another model is being discriminatory. But what if the overseeing model is itself discriminatory? But as I have said, imperfect societies make imperfect models, especially when factors of race, mental health, etc. are involved. Another possible solution is to see why these companies lean so heavily on these models for applications — profit, and usually at some potentially illegal means. As of the writing of the book, the Kronos model was being challenged legally, maybe the others should as well?
There is really no big takeaway from this chapter that I have not expressed so far. The companies will do anything to maximize profits, including employing the use of models that are discriminatory and lead into a feedback loop that has the potential to stop society from progressing in necessary areas.
