Individualizing Without Excluding: Ethical And Technical Challenges : Filter Bubbles and their Effects on Society

Sammanfattning: In December 2009 Google had a press release, in which they stated that they would start individualize their users' search results. This is the year individualization began, according to Eli Pariser, who also coined the expression filter bubble. A filter bubble is said to occur when a user only receives individualized feeds, causing them to only consume information that align with their beliefs and excluding them from contradicting information. The main goal of this thesis is to evaluate the existence of filter bubbles and the effects of them. In order to reach the goal, an experiment on a large social media platform is conducted, as well as a qualitative literature study. The experiment consists of fourteen bots that are assigned different behaviors which are performed on a social media platform, which expresses that they use individualization of users' information feeds. The presented results from the experiment are based on data crawled from the bots' feeds. The main finding of the experiment is that no significant filter bubbles are observed and that users experience less exclusion the more of the feed they consume. The literature study consists of 24 papers that both do and do not support the theory of filter bubbles. The main effects found are that filter bubbles fosters polarization of opinions and confirmation bias. It is also said that this could lead to undermining deliberative democracy.