The debates in social networks are increasingly characterized by polarization and fragmentation. A significant role in this phenomenon is played by the recommendation algorithms of major online platforms, which promote the spread of sensational and divisive content. We demonstrate how algorithms can contribute to a better digital discourse in an interactive web publication.

It has become evident, especially since the takeover of X (formerly Twitter) by Elon Musk, that the digital discourse is increasingly shaped by fragmentation and polarization. The recommendation systems of online platforms play a crucial role in this trend. These systems are based on algorithmic decision-making and are currently designed to maximize interaction with content. The critical factor here is not the quality of the content, but rather that the piece of content is clicked on, liked, shared, or commented on as much as possible. Therefore, sensational and divisive content is often preferred. The reason for these recommendation systems is simple: increased platform engagement leads to higher advertising revenue.

However, there is an alternative approach. Online platforms can align their recommendation systems in a way that goes beyond the sole maximization of interaction. To achieve this, they would need to base their recommendation systems on other criteria, such as the likelihood that different opinion groups would agree with a piece of content. Bridging algorithms work according to this principle, promoting mutual understanding and productive debate.

Bridging algorithms not only function in theory but also in practice

There are already successful examples of such recommendation systems, such as the real-time polling system “Polis,” developed in 2012 and operated by the U.S. based non-governmental organization “The Computational Project.” This open-source tool aims to visualize opinion groups in discussions to better identify major areas of agreement and disagreement across different groups. “Polis” has been successfully utilized in various contexts, including in Taiwan in 2015 when the government sought public opinion on the introduction of UberX.

Another example is the “Community Notes” feature for content moderation on Twitter/X. It allows users to collectively write comments on potentially misleading posts. Only those comments that have been rated as helpful by individuals with different opinions are published.

Leveraging Bridging Algorithms for Enhanced Digital Discourse

These two examples demonstrate that Bridging Algorithms operate successfully in various contexts. However, what is currently lacking is a more pronounced commitment from major online platforms to fulfill their societal responsibility in fostering improved digital discourses. This could involve the integration of Bridging Algorithms, among other strategies. Achieving this goal wouldn’t necessarily require a complete overhaul but rather an augmentation of existing recommendation systems with Bridging criteria. Presently, platforms are not taking sufficient measures to counteract the trends of fragmentation and polarization. Consequently, there is a need for increased public and political pressure, and, if deemed necessary, regulatory interventions.

Through our web publication, we aim to raise awareness and understanding of the topic of Bridging Algorithms. Our objective is to contribute to a landscape where recommendation algorithms oriented towards fostering productive discussions become the standard on platforms, rather than remaining the exception.

This text is licensed under a Creative Commons Attribution 4.0 International License