Vocal, Visible and Vaulting ? (Dis)connected niche audiences in the age of artificial intelligence


Societies are polarizing and misinformation is spreading. These developments are often traced to so-called filter bubbles arising from the use of algorithmic filter systems on social media platforms. However, extant research reports seemingly contradictory findings. On the one hand, there is clear evidence of groups of users exchanging attitude-confirming information within algorithmically curated echo chambers. On the other hand, a growing body of research finds scant evidence of filter bubbles existing at the aggregate level of the population. These findings, I argue, are not contradictory but a consequence of studying filter bubbles in the wrong place. Artificial intelligence-afforded communication bubbles are thriving in the form of fringe bubbles—among niche audiences made up of smaller, specific segments of the population. Yet, due to the affordances of algorithmic filter systems, fringe bubbles can grow and gain the power to set public agendas.

With a starting point in media effects theory, I posit that the biased and filtered nature of information within fringe bubbles affects the perceptions of those who inhabit them. More specifically, members of online communities come to inaccurately perceive what the majority of the population thinks, often leading to heightened political mobilization and participation with consequences for society at large. Through innovative multi-method research designs including automated content analysis, survey research, experiments, and analysis of digital trace data, the proposed research project will: (1) map the landscape of fringe bubbles in the Netherlands; (2) study their effects on perceptions of public opinion among people who inhabit these bubbles; and (3) examine the relationship between people inside and outside of communication bubbles. In doing so, this research furthers our understanding of the implications of artificial intelligence on societal divides and polarization.





Dr. J.E. Möller

Verbonden aan

Universiteit van Amsterdam, Faculteit der Maatschappij- en Gedragswetenschappen, Communicatiewetenschap


01/09/2019 tot 31/08/2022