EECS 649

University Of Kansas

Blog 6

Evan Powell

Chapter Four of O’neils’ Weapons of Math Destruction discusses modern-day advertising, and more particularly, the weaponized version that is implemented and wielded by for-profit colleges.

Growing up in the blossoming of the age of information, I was used to seeing advertisements everywhere. Seemingly to me, it was the cost of engaging with entertainment: there were commercials on T.V., there were commercials on the radio; there were commercials and ads on the internet. I was, at the time, unaware of how calculated these advertisements could be. I now know that these advertisements can be made more efficient by an amazing feat of data modeling, they can be placed at the mathematically best times to maximize their returns on investment. In a vacuum, this is a feat — hard work was done and an incredible model was built that could save people and companies time; if I am starving and this model sees this and gives me an advertisement for food with a coupon, we all come out on top. But only in the vacuum.

These corporations and executives go farther, ignoring privacy concerns to have a finger on the pulse of the data, and sometimes, like these for-profit colleges, go farther to mix in psychology, to seek- to hunt- people who could be suffering from depression or anxiety. They hunt these people with ads because they know they are more likely to fall victim to their unethical business traps. O’neil describes this process as the way these machine learning data models are used to exploit vulnerable people. This is only more evidence that the powerful can harness technology and these models as ways to further their power and simultaneously deepen the divide of fiscal and social classes. As I learn more about computer science and perhaps more about AI learning, I wonder if I will see that there does exist a more morally intact way to operate with these types of advertisements. O’neil did not shy away from these for-profit degree mills, and these places and operators should be shunned, but I wonder if there are companies that are using this in a better way. And by better I truly mean less worse.