Online Advertising and Machine Learning
Posted on March 17th, 2021 by Emilio Miles
In Chapter 4 of Weapons of Math Destruction, Cathy O’Neil talks about for-profit colleges as well as online advertisers and their methods of targeting underprivileged individuals to make money. She goes in depth about how for-profit colleges operate and uses different examples, such as Corinthian College or Vatterott College, to outline how their “scams” work.
Corinthian College, she mentions, was a for-profit university who targeted low-income individuals who had “low self-esteem” or were “impatient” or “stuck” in life. They charged incredibly high prices for tuition and offered loans to people to help pay for it in order to make the prospect of enrolling more attractive. Once they were looked into by the Obama administration, they put a hold on the company’s access to federal funding and they inevitably went bankrupt. Vatterott College was another for-profit university with questionable methods. Their recruiting manual directed recruiters to target “Welfare Mom w/Kids. Pregnant Ladies. Recent Divorce. Low Self-Esteem…,” etc.
She uses these colleges to describe how online advertisers act in a similar manner. They sell ads that pinpoint people in great need and sell them false or overpriced promises. They find inequality and feast on it. She also describes the use of machine learning to exploit these vulnerabilities. Machine Learning has improved vastly since the 1960s. Data scientists have been amassing data since then and have created models that can predict what a consumer is likely to be interested in. She describes how for-profit colleges also use machine learning in this manner for their benefit. Some of the revenue that for-profit universities accumulate ends up going to sites such as Google and Facebook which contain massive amounts of user data that can be used to help these companies tailor their ads to their targets.
I believe the biggest takeaway from this chapter is that Machine Learning can be used to hurt people in ways that one might not even think were happening. It appears to be the same theme that spans the course of the book. Unchecked, these algorithms can hurt a lot of people. It is important to be aware of how these tools can be used for bad not only to avoid falling victim to them, but also in hopes to reform the current environment.