Filtering Information on the Internet Fosters Extreme Opinions
The article “The Daily We: Is the Internet really a blessing for democracy?” argues that the emergence of the Internet allows people to filter what content they access, therefore limiting their scope of knowledge. Before the age of the Internet people would be forced to read newspapers, converge with others, or watch the news on television if they wanted access to news. These sources did not filter stories based on individual preferences or interests. As a result, people would be exposed to an assortment of information that they would not necessarily pursue on their own. For example, a left-winged liberal that is opposed to the war could be exposed to a right-winged conservative article about how the United States should continue its enforcement in the Middle East. The left-winged person is being exposed to the opposite viewpoint and therefore can become more sympathetic to the right-winged person’s viewpoint.
Unlike older ways of obtaining news, the Internet allows people to customize their preferences to what they want to see. Once people click their preferences or rate articles they read, other similar articles are recommended to the user. These articles have most likely been rated by other users with similar preferences to the original user. Thus, the user will read more articles that reinforce their own beliefs and opinions as opposed to reading opposing opinions and viewpoints. In this sense the left-winged liberal will have firmer beliefs about their original opinions and become more left-winged. The same argument can be made for Internet users that are right-winged. This mechanism creates a larger split between the two parties because each party is only seeing their own viewpoint reinforced.
This example on the Internet can be seen as a cascade effect because the recommendation system reinforces the ideas the user already embodies. Take for example, a user who has no preference for either political party. The user just clicks on articles at random that he sees on the front page and then ranks them according to how he likes them. At first the recommendation system is unbiased because it does not know the user’s preference. However, once he begins to highly rate articles that have to do with one particular party (say the Republican party), then more right-winged articles will be suggested to the user. Eventually, after more rating, most of the articles will be right-winged and the user will have a strong positive biased for Republican type issues over time.