Getting A Job

A key to any successful company is the optimization of daily processes. If there’s a way for a corporation to operate at a cheaper cost, there’s a good chance they’ve already made the change or plan on implementing it. One such method of cost reduction is the automation of human resource departments. In the sixth chapter of Weapons of Math Destruction, Cathy O’Neil brings to light the models lurking behind the application process of most companies. It is no surprise that large companies receive vast numbers of applications for every job posting. In the past, these applications would be reviewed manually by groups of humans. This variation in who is reviewing each application led to personal bias being incorporated into the hiring process. As a means to solve this problem, companies have since taken to automating most of the process with the use of software like Workforce Ready HR, developed by Kronos. These programs consist of personality tests which are designed to gauge the viability of each candidate for the available position, however it is these tests which cause the process to come into question. The system “receives precious little feedback”, allowing the system to run amok with outdated and biased data. For the applicants, the whole process is opaque. Unaware of what answers the company is looking for, they are left in the dark when/if their applications flag an undesirable result. As a whole, Cathy O’Neil says it is these aspects which lead to injustices on the part of the companies using the system — personality tests trained on old biased data and the black box methodology of the model. To improve on this, as was the case with the crime prediction software, modelers (and companies) must remove biased data and accept a less efficient system. It is crucial that “doing the right thing” becomes a higher priority than “do what’s cheapest”. It is hard to not get a sinking feeling in the gut over the course of this reading. As a soon-to-be graduate, the percentages of companies employing such biased technology is worrying. It seems that no matter where you are on the economical scale (except perhaps at the top), somebody is using incorrect models to make life changing decisions for somebody else.