Abstract
YouTube has been accused of radicalizing users on their platform, through the unforeseen
consequence of the concept of the so-called filter bubbles. Through a series of experiments and case
studies, based on general implications of machine learning, black boxes and surveillance capitalism,
this research team has explored the validity of filter bubbles as a theory. By developing a model for
measuring nuttiness based on elements of structured observation, the team is able to determine the
general nut-number of any given YouTube video and measure the influence of clicks on the
recommender system
