Algorithm bias and why we’re all angry about it


Kevin Slavin: How algorithms shape our world

It’s not just #OccupyWallStreet-ers who are angry about algorithms. In any business that has been online for any amount of time there exists at least one person who frets about algorithms, surmises new ways to make those algorithms work in their company’s favor, and inevitably gives up knowing that jumping feet first into the world of algorithms is like entering a black hole. And, with no end in sight, we give up, angry and frustrated.

It’s just that the Occupy-ers are the only group of folks to complain about it and receive news coverage for having done so. The NPR story linked here does a solid job of uncovering what most of us already knew about algorithms (that they favor topics trending locally, at a given moment in time) and that anyone actually trying to get their name known through a chance appearance on Twitter’s most famous list is really just wasting their time and money on a frivolous pursuit. Like the article says, it’s more likely that #thingsthirstypeopledo will find its way to the top ten long before any advertising campaign messages do. And, they’ll inspire greater interaction.

The more intriguing part of this story, as is often the case with most NPR pieces, is the final section, titled “Getting Used to an Algorithmic Editor.” Quoting Cornell University communication professor Tarleton Gillespie, the article paints an algorithm as nothing more than your local newspaper editor.

Or, better yet, think of an algorithm as yourself while you edit that term paper you’ve been working on for months. “This doesn’t belong there,” you say, and you scratch it out. “This would work perfectly over here,” and you add it. You are the filter, the gatekeeper over your work. What you let through is made public. What you don’t is not. And, the reason you let some things through while others are left behind is because you’ve likely received years of training in this particular discipline. You know intuitively what ought to belong, what may merit the highest grade, what sounds best, how many sources you should have, how many block quotes you shouldn’t, and so on and so on.

But, your intuitive knowledge is loaded with subjectivity as well. Though you were taught how to write essays in junior high school, everything from your style and voice to your research techniques to the number of times you’ve written anything before this point all serve to form in you a particular way of writing thatĀ  is all together different from your professor or fellow classmates. And you still have to consider your particular ideological positions that inform the way you think about your topic.

Algorithms had to be written and edited. They are still being edited. The NPR article indicates that even just a few years ago, Amazon had an algorithm snaffoo that allowed adult-themed titles onto their best-seller list – something they decided should not happen. There are people trained in the science of algorithm writing and they, just like you writing your paper, are biased.

“The important point is that one can never generalize beyond known data without making at least some assumptions,” Martin Sewell wrote for Futures Magazine. The point of the algorithm is to represent information that has not happened based on a mixture of the known data and the assumptions that can be drawn from that data.

Our problem as consumers is that we have been trained to think about anything electronic as inherently unbiased, objective. You can buy something on Amazon without ever interacting with a human being. You do this because there is little fuss, no salesmanship and it can easily be done from in front of the television or as a distraction from work. You do it without thinking. Without engaging. What you do not think about is the host of work happening just on the other side of your computer screen as you clicked the “purchase” button. Every search term, click, and scroll you made while on that site has been recorded in a way that can be marketed back to you. Every second you stayed on oneĀ  page longer than another and every source you used to get to that particular place are now part of your online DNA.

But, we ignore this. We get angry when the algorithms don’t work in our favor and we fight about online privacy when we’re the ones exposed. It certainly seems like until we purposely get a hold of our habits online, this is one fight mankind will not win.