Social media platforms like Twitter and TikTok wield significant influence over the visibility and reach of content, raising critical questions: Are certain voices marginalized based on political leanings, or are platform interventions primarily aimed at mitigating the spread of disinformation? Recent discussions highlight allegations of political bias in account suspensions and “shadow banning,” a subtle mechanism where certain content is less visible to users, often without their knowledge.

Drawing from a field study on Twitter, we investigate these claims. Our findings suggest that while some account suspensions may appear biased, there is equally strong evidence indicating that these actions are strategic efforts to combat the spread of disinformation.

We then delve into the issue of shadow banning, demonstrating how this mechanism can be used to shape opinions into arbitrary distributions by solving a linear program. Through simulations on real social network topologies, we show that if shadow banning were driven by partisanship, its implementation could be so subtle that overt political biases remain undetected. This underscores the imperative for rigorous transparency and oversight in the actions of social media platforms.

This talk is based on the following papers:

1.  Analysis: Trade-offs between reducing misinformation and politically-balanced enforcement on social media: https://osf.io/preprints/psyarxiv/ay9q5
2.  Shaping Opinions in Social Networks with Shadow Banning: https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0299977