Internet Services and Filter Bubbles
A result of improved processing power is that it allowed Internet companies to analyze historic user behavior and filter a user’s future media based on complex analytic functions. These are used to investigate their previous browsing history and demographic properties. Web companies strive to tailor their services based on user tastes using machine-learning algorithms and have, in a sense, been successful in clustering users with content catered to these historic interests.
Internet activist Eli Parsier in this Ted talk questions this methodology of content personalization and its wide adoption across all Internet media centers from Facebook to New York Times. He believes this user-specific filtered content will very quickly lead to idiosyncratic content that the Internet believes a users want to see rather than what we may need to see. These tailoring algorithm stems from ideas similar to the political forum divides shown in class. Internet services seem to assume that user’s are more likely to stay within their spheres of interests and suggest links that cluster similar ideas. This is dangerous in that it leads to invisible autopropaganda due to the user’s inability to regulate these machine algorithms, amplifying our desire for familiarity and leaves us ignorant of what these algorithms determine to be irrelevant to us.