Cross-platform impact of social media algorithmic adjustments on public discourse

TL;DR

This study examines how a transparent change to Twitter's recommender system in 2023 affected public discourse, revealing impacts on content visibility, engagement, and community reactions across platforms.

Contribution

It provides a rare analysis of the cross-platform effects of a real-time, transparent recommender system adjustment on public discourse and community dynamics.

Findings

Algorithm changes increased content visibility and engagement.

Higher visibility influenced social dialogue and community reactions.

Recommender system adjustments can both positively and negatively impact discourse.

Abstract

In the hypertrophic and uncharted information world, recommender systems are gatekeepers of knowledge. The evolution of these algorithms is usually an opaque process, but in February 2023, the recommender system of X, formerly Twitter, was altered by its chairman (Elon Musk) transparently, offering a unique study opportunity. This paper analyses the cross-platform impact of adjusting the platform's recommender system on public discourse. We focus on the account of Elon Musk and, for comparison, the account of the President of the United States (Joe Biden). Our results highlight how algorithm adjustments can boost content visibility, user engagement, and community involvement without increasing the engagement or involvement probabilities. We find that higher visibility can increase the influence on social dialogue but also backfire, triggering negative community reactions. Finally, our analysis offers insights to detect future less transparent changes to recommender systems.