Stanford Seminar - Interest-based & Ideological Biases in Online Content Consumption

Magdalena Wojcieszak discusses the dangers of the current communication environment, arguing that the focus on radicalization and echo chambers online is too narrow and overshadows the more fundamental problem of interest bias in recommender systems. She presents a study using sock puppets on YouTube, which found that while ideological biases exist, they are not as prevalent as the interest-based biases that lead to low consumption of quality news. Wojcieszak shares findings from a large-scale study analyzing online browsing data, revealing that news and political content comprise a small fraction of online consumption. She then presents two interventions aimed at incentivizing greater exposure to verified news, including a field experiment on Twitter and a browser extension for YouTube, highlighting the potential for algorithmic nudges to increase news consumption and diversify exposure, while also acknowledging the challenges and ethical considerations of such interventions.