Rabble-Rousers in the New King's Court: Algorithmic Effects on Account Visibility in Pre-X Twitter

A team of researchers from the University of Washington, including Alexandros Efstratiou, Kayla Duskin, Kate Starbird, and Emma Spiro, have published a peer-reviewed study analyzing algorithmic bias on Twitter during a critical transition period. The paper, 'Rabble-Rousers in the New King's Court: Algorithmic Effects on Account Visibility in Pre-X Twitter,' was accepted at The 20th International AAAI Conference on Web and Social Media (ICWSM 2026). It examines data from user feeds collected after Elon Musk's ownership began but before the platform was rebranded to X, replicating and expanding on prior findings about political exposure.

The core finding challenges a simple political narrative. The research confirms that right-leaning accounts received significantly more exposure in algorithmically curated feeds compared to reverse-chronological ones. However, the analysis reveals this boost was correlated not with ideology itself, but with specific, rewardable behaviors. Accounts that posted more 'agitating' content and, crucially, those that received attention (replies, mentions) from Elon Musk—who was identified as the most central node in the network—gained disproportionate algorithmic reach. This suggests the platform's mechanics incentivized certain types of engagement-driven behavior.

A secondary major finding highlights a shift in verification status value. The study shows that legacy 'verified' accounts, typically belonging to journalists, government officials, and established institutions, received less exposure in the algorithmic feed compared to both non-verified accounts and new Twitter Blue subscribers. This indicates a fundamental recalibration of which voices the platform's central algorithm amplified during this period, with potential implications for the visibility of authoritative sources versus those optimized for platform engagement.