Biased Search Engine Results?
In his article “Google’s Algorithm Isn’t Biased, It’s Just Not Human,” Noam Cohen writes about the common political conservative critiques against Google’s search algorithms. According to Cohen, American conservatives feel that they are disproportionately represented in a poor light by Google’s search results. But with the widespread lack of understanding of technology on the Hill, it is not so surprising to see so many elected representatives cling to the belief that there is someone in Silicon Valley dedicated to tarnishing their names. In reality, there is no specific line of code within any of Google’s programs which dictates its search results to denigrate all of American conservative elected representatives.
As we learned in lecture, the technology which actually backs Google’s powerful search engine is an algorithm named PageRank. Though in lecture, we mainly focused on learning about how the algorithm actually operates, and how we can manually compute rankings in a variety of hypothetical situations–PageRank is related to Cohen’s article in a more high-level, conceptual way. In this article, PageRank is described as the breakthrough algorithm Google utilized to make sense of the chaos that the Web once was. Through its iterative update rule, PageRank helps sort the billions of web pages that currently exist to find the “best” authority. It is a seminal algorithm that has propelled Google to the forefront of the search engine industry, and helps facilitate the tool that nearly every internet user in the world uses on a daily basis.
However, Cohen goes further in his article to argue that Google made a misstep in applying the PageRank algorithm to the realm of politics. In his eyes, “it sowed chaos where there had been a modicum of order.” He suggests that when it comes to matters of politics and public policy, human-to-human interaction would still provide the best type of information. To support this claim, he brings up the case of Dylann Roof. In 2015, Dylann Roof killed nine African Americans in a church in Charleston, South Carolina. Roof writes in his manifesto that he specifically remembers searching “black on white crime” on Google, and that the first website he came to was the Council of Conservative Citizens. Cohen stipulates that had Roof gone to, for example, a library to learn more about “black on white crime,” he would have been met with a human being who could recognize impending trouble and find Roof proper assistance.
My issue with this argument is that it very well ignores the fact that human beings can be just as biased, and potentially even more persuasive than the algorithm which gave Dylann Roof the search results he saw. But past this, Cohen’s argument does raise an interesting question of whether big tech firms have the authority to more carefully filter what information flows where. Not to say that violent extremist propaganda should be prolifically circulated, but there is a very slippery slope when it comes to censoring free speech. I find that these are important considerations that we should pay close attention to as our society’s dependency on these technologies grows with each day.
Link to article: https://www.wired.com/story/google-algorithm-conservatives-biased-its-just-not-human/