Are search engines biased?
Siri does not know how to respond to the question “I have been sexually assaulted.” When searching for abortion clinics on Apple Maps, one is instead shown adoption clinics. Finding a clinic or pharmacy that has Viagra is apparently easier than finding one with birth control.
Many view the Internet as an unbiased, objective, tool. But in actuality, the Internet is influenced by gender bias. This is especially true when it comes to search engines — our guides to exploring the vast Internet, which shape what we see and act as gatekeepers to often vital information. Many see search engines as reflective of an “objective” reality of facts (if there even is such a reality), but they are actually composed of structures and codes that often overlook, ignore or even dismiss the experiences and perspectives of women. Search engines are less neatly organized libraries than they are chaotic marketplaces of ideas and facts all fighting to become the most relevant.
To understand how this bias emerges, it’s important to understand how search engines work at all. The “spider,” or web crawler, is most vital aspect of search engines. Simply stated, it is a bot that endlessly performs the repetitive task of searching. In the background, this bot first gathers information, keeps copies of that information and then builds an index. In the foreground, it understands a query, determines the relevance of each result, determines how to rank relevant results and then presents them to the viewers.
It is in this foreground where the bias becomes most evident. While query understanding tech and natural language queries are improving accuracy, there are still serious problems — like confusing abortion and adoption and overlooking the importance of women’s reproductive needs. The foreground of search engines don’t act as truth-tellers or librarians trying to help you find the most accurate information, therefore, but rather as brokers, which creates a tyranny — a bias of majority interest.
There are likely many reasons this bias exists, but perhaps the biggest is the serious lack of women programmers in Silicone Valley. As NPR reported, many big tech companies recently revealed how few of their programming and technical employees are female. Google, which had one of the highest rates, still only has a technical staff that is 17 percent female. When the vast majority of programmers are men, the issues most important to women are not actively considered and are often ignored altogether. This is not necessarily an act of intentional malice, but one of casual indifference.
It seems this problem could be easily solved by simply hiring more woman to work on query understanding tech. But the reality of the problem is much more convoluted. Silicon Valley is often hailed as the new “boys club” that perpetuates a culture that hardly welcomes women. For example, in the article “The Loneliness Of The Female Coder” Ciara, the only female software developer on her team, states that, “Although I was like them in many respects, I looked at things from a different angle, one which my colleagues often didn’t recognize or adapt to. Why should they? I felt like the lone voice in the wilderness.”
This issue is made even more complicated by looking at the limited possibilities of how search engines can be funded. Users could pay for search engines, the government or non-profits could pay for search engines, or advertisements could pay for search engines. Each option has drawbacks and biases, but it seems public search engines funded by advertisements have won.
Take the huge success of Google, for example. Google, a service that’s free for its users, receives a large amount of traffic and a large number of “clicks” on any given link, which incentivizes advertisers. In order to keep commercial sponsorship, Google has to be an effective tool and deliver quality results for its users. To accomplish this, the FTC required search engines to clarify what was paid in 2001 and advertisements were given distinguishing new names such as, “sponsored,” “paid advertisement,” “feature listing” or “partner results.” But while these advertisers may be algorithmic and ostensibly regulated, they are still not unbiased. Not all search engines are created equal, which is frequently documented by obvious kinds of distortion and search engine optimization manipulation.
Google is a private company, and does not need to disclose its code. As WIRED aptly put it, Google “guards its search platform like the crown jewels.” Search engines are a part of the free market, and Google does not want competition, so the way their algorithms work is not public knowledge and cannot be checked for accuracy or fairness. This means when viewing the world online, one cannot be sure exactly through whose lens they are viewing it. It’s highly likely, however, that this lack of regulation as well as the“boys clubs” of both big business and computer engineering both played significant roles in what they see.
At its core, the patriarchy seeks to control women’s options and choices. When the male-dominated businesses of Silicon Valley control women’s perceptions click by click, it seems that the Internet and, by proxy, search engines, accomplish just that.