Doporučovací systémy streamovacích platforem a jejich vliv na homogenizaci filmových preferencí

This thesis examines how users of streaming platforms perceive recommender systems and whether their experiences reflect mechanisms that the academic literature associates with the homogenization of recommended content. The theoretical section outlines the main approaches to recommendation, the factors shaping the form and variability of recommended content, and mechanisms such as popularity bias, overspecialization, filter bubbles, low serendipity, and feedback effects. It also introduces the concept of algorithmic culture and discusses the role of streaming services in the contemporary film environment. The empirical part is based on nine semi-structured interviews with users of Netflix, HBO Max, and Disney+. The analysis shows that respondents across platforms perceive repeated recommendations, thematic narrowing of the offer, and a more limited diversity of recommendations, which correspond to selected mechanisms described in the literature. At the same time, differences emerge between individual services, largely related to the nature of their catalogues and the way platforms prioritize and promote content. The findings suggest that the perceived manifestations of recommender systems may contribute to a certain degree of content homogenization, and that users often notice these patterns even if they do not explicitly interpret them as outcomes of algorithmic mechanisms.