Are SEOs Responsible For Google Search Bias?
Are SEOs Responsible For Google Search Bias?
In the article “Are SEOs Responsible for Google Search Bias” by Helen Jelenc, she discusses the bias present in search results today. Internet access is considered a human right by the United Nations because it produces access to adorable information and expression. The internet has become such a crucial part of our social infrastructure and biased algorithms and search results can have a large impact on our lives.
There have been many tests on the bias in search results. Jelenc discussed when a researcher looked up professional photographers, they found that the majority of those on the first page were male even if the field is half-male. This means that the search results show a bias towards professional male photographers which can have a huge impact on female-run businesses.
Next, she discusses problems with PageRank, the algorithm Google uses to rank articles and websites on the web. Because of PageRank’s algorithm, top-ranking websites stay at the top and other websites can’t compete. They wanted to test homophily which is defined as the “tendency to connect to similar others” by using a model with 20 nodes, websites, that each had a PageRank score and were classified as part of the minority or majority. They used the Gini coefficient to see how these websites compared to an equal distribution. Their findings discovered that “PageRank can reduce, replicate or amplify biases” (Jelenc, 2022). Models with high homophily found that majority voices were dominated while models with heterophilic majority had minority domination.
Next, another study where countries’ social gender inequality was compared to search results of the word “person ” in each country’s respective languages using VPNs. They found that the search result“person” resulted in more male entries in countries with higher societal gender inequalities. They concluded that there’s a link between societal norms and the algorithm’s output.
Jelenc states the responsible parties for these algorithms are those creating the algorithm, society, culture, and environment we live in. Jelenc’s recommendations were: don’t replicate biased content, audit AI content, create algorithm audits, support dei education, provide multilingual resources to encourage diverse voices, create less biased algorithms & AI— stop search gentrification where more companies are in the search landscape.
Personally, I think there have been so many examples of search result bias that I’ve heard of. Although I haven’t heard of most of these studies. I agree with Jelenc that these results are because of our flawed social systems and the developers of the algorithms. But I beg the question, is it the developer’s responsibility to create an algorithm that is reflective of the society we strive to be or the society we actually are? If we expect developers to create search algorithms that actively fight against inequalities and bias, this requires more experts in social behaviors before implementation as software developers do not have the expertise to make these conclusions.
This relates to course concepts because we’ve discussed both the benefits and problems with PageRank. PageRank uses the direct relationships between authorities and hubs to rank web pages. As seen in class and homework, I saw that once a web page becomes more powerful and these pages link to them, they become more powerful which shows both the stability of the algorithm and how the majority is bound to win online.
Jelenc, H. (2022, September 15). Are SEOs Responsible For Google Search Bias? Search Engine Journal. https://www.searchenginejournal.com/google-search-bias/463549/