group 1.3

Algorithms: A contemporary tool of polarization

by Ayham Almohamad, Chang Ho Park, Janella Dujale, Karel Leertouwer, Kevin Ferdinand Aipassa

 Amsterdam, Mar 20 1019 - The algorithms of Facebook has long been a topic of discussion with regards to its filter bubble effect on Facebook users. Looking at the data we have acquired through Fbtrex along with the help of descriptive statistics, we have gathered results that demonstrate the impact of the Facebook algorithm on its users’ online experience.

Our observation shows that the algorithm filters the content provided by facebook on individual users’ news feed based on the data that Facebook has gathered from them—through their browsing history, location, interactions and online behaviour on the platform, as well as on other platforms that Facebook has built integrations with (such as Instagram), and this causes echo chambers where the content a user encounters consist of only those that align to his existing views, isolating one’s perception from alternative ideas, beliefs, and opinions.

This notion was manifested on the data we have gathered on our fake persona which comprised mainly of content that coincided with the characteristics that we wanted it to possess, including being politically inclined (extreme left), an activist for environmental and socially related matters, a festival goer, a vegan, and a cat lover.

 

Figure 1: Our fake persona’s data representation

Figure 1 shows that the largest part of the data is content about entertainment (41%), followed by news, anomalies (which comprise merely of random facebook profiles), hobbies, and then politics. We categorised these further in Figures 2 and 3 below, which showed that the largest part of the entertainment data were cat videos (64%), and protests (39%) for the data about hobbies.

Figure 2: Specific categorisation of data about Entertainment (Anne)

Figure 3: Specific categorisation of data about Hobbies (Anne)

In comparing both profiles, figure 4 and 5 below  represent the amount of posts, groups, events, photos, and videos of each facebook account. Although the data of our fake persona shows interaction with more groups, clearly, our real account’s data relatively consist of a wider variety of groups in comparison to the fake persona, which is due to the fact that the real account has been exposed to a wider variety of groups and content. in the time being that the person has been using Facebook.

Figure 2: Specific categorisation of data about Entertainment (Anne)

Figure 5: Statistical data of online behaviour on Facebook (Karel)

Looking at the Figures 6, below, compared to Figure 1, Figure 7 to Figure 2, and Figure 8 to Figure 3, both profiles differ largely on the content that the data has collected and this is naturally because of the divergence of their interests. Again, the real facebook profile contains a wider variety of data compared to the fake persona.

 

Figure 6: Our real facebook account’s data representation

Figure 7: Specific categorisation of data about Entertainment (Karel)

Figure 8: Specific categorisation of data about Hobbies (Karel)

Clearly the data collected was in line with the characteristics of Anne Van Dijk, or in our case, with the posts, groups, events, and pages that we, as a group, had interacted with during the duration of the fbtrex exploration. There was no observation of content that was in contrast to our fake persona’s views which further proves that the algorithm of Facebook stimulates content personalisation as well as create a filter bubble.

However, this poses a critical limitation that arise from the use of fbtrex and the creation of a fake persona because we argue that the comparison of a fake persona and a real account would not lead to an accurate analysis due to limited time constraints as well as limited accessibility to posts and pages for our fake persona’s account. The real facebook data would show a wider range of content. Moreover, the occurrence of the data of the fake persona would vary to our preferences: we liked and shared cat videos more in the beginning which resulted to the proliferation of more cat videos on our feed.

Moving forward, the algorithm is a tool used by Facebook, a larger corporation, in order to retain user attention on their platform. Through the information that users present as well as their online behaviour, data is created and further commodified into monetary value and proceed toward Facebook’s revenue. Although it eases users’ online experience as it essentially spoon-feeds the content that the users would like to see, it is a barrier to the free flow of information, which according to many Internet activists, is the essence of the Internet.

How might algorithms influence public and private life? Are we underestimating the impact of algorithms on our “real” lives, away from the Internet?  Perhaps the isolation of individual views would eventually lead to socio-political issues that involve the proliferation of chaos and the eradication of being open-minded to alternative views, which poses a larger problem in the society. Yet what are we, as intellectually progressing civilians, doing about it?