I found this article particularly relevant after Sean Taylor’s guest lecture on Tuesday, and I think someone mentioned in a question as well. Taylor gave an overview of the ideas and roles of data scientists at Facebook using comment rank to enhance user experience. Their current strategies to best suit consumers are now being accused of being politically biased, particularly in the area of the upcoming presidential election.
Allegations of purposefully reducing conservative news sources popularity on Facebook was denied by the company. When asked about these accusations in class, Taylor backed Facebook’s focus on neutrality, saying that embedded political bias is against their company culture as well as their mantra of staying open minded and connected. However, he did add that the company is comprised of mainly younger, well-educated, and liberal employees.
I think Taylor’s efforts as a Data Scientist at Facebook illuminates the difficulties of the modern web. There seems to be a difficulty in compromising between collective maintenance and the efforts to personalize the Web. Taylor’s initial topic was simply demonstrating how difficult it is to determine a “quality” comment. Yet, even in their statistical, constantly evaluated formula, they face criticism from public news sources.
I found this particularly interesting in relation to what he talked about with surveys – that they want to know how they’re doing but it gets annoying to ask users, so they try to avoid it as much as possible. How much are users aware of this possible suppression of sources? How much is this accusation possible just through users actions as well? For me, I really don’t use Facebook to hear people’s political opinions, so – for example – I could unfriend someone for posting too much about their political views, and then my page could look unbalanced politically. So, how much is user activity behind Facebook’s criticism? Or, how much do the employees implicit biases affect the way they set up trending articles? I often assume that the web, computers, systems, etc, are so automated that they are unbiased, but when you’re reminded that they must be set up and implemented by individuals, it can be a refreshing perspective.