The article focuses on an experimental tool that was able to measurably reduce partisan political hostility by changing the sorting of posts in the feed. The study shows that even small algorithmic adjustments – without deleting or blocking content – have a significant effect on users’ emotional attitudes. Posts with anti-democratic or highly polarizing content were simply moved down in the feed. As a result, participants’ negative attitudes toward those with different political views decreased, across party lines.
In her assessment, Dr. Josephine Schmitt emphasizes that the results show robust and, in some cases, surprisingly strong effects on affective polarization. The study illustrates that feed sorting is by no means neutral, but directly affects emotions and thus influences political attitudes. This finding is central to the social debate on transparency, responsibility, and the design of algorithmic systems.
At the same time, Josephine Schmitt points out the limits of transferability. The results must be interpreted against the backdrop of the US two-party system and can only be applied to a limited extent to countries such as Germany, where social media such as X play a lesser role in everyday political life.
The media response underscores the relevance of CAIS research and expertise in the field of digital public sphere and political communication. Josephine Schmitt’s analysis highlights how even minimal algorithmic decisions can strongly influence democratic discourse – and how important it is for society to engage in an informed debate about the power of platform algorithms.